Of course, every idea list has hits and misses. But without transparent accounting, it’s easy to blur the lines. In reality, whether any of these trades actually worked is a question of entry, exit, and allocation. Absent clear communication and fair ledger, there is no credible way to assess value — and as a reader it is harder to extract value.

Now, I have a solution: a transparent, real time, live model portfolio (Details below).

Introducing TLBS Model Portfolios

There is no shortage of smart folks talking about Google or NVIDIA and I’m not sure how much value I have to add. Instead, I’ve found that I have better luck uncovering lesser known ideas — especially ones that go against the grain.

Rather than a permabear, The Last Bear Standing is contrarian at heart. That can mean fading market darlings or finding diamonds in the rough. It could be the collapse of Chinese property developers or the rebound of the Hang Seng index.

Today, I’ve written up eight small and mid-size (SMID) stocks that will be the start of a dynamic TLBS Contrarian SMID model portfolio.

Here’s my selection process.

Criteria: I like off-beat companies that are on the cusp of inflection — either in fundamental performance or market sentiment. My timeframe is neither the next quarter nor 2050 — I’m looking for concrete developments that I expect to be realized (or recognized) within the next 1-2 years. These companies have largely idiosyncratic business drivers rather than hinging on macroeconomic trends. Given this horizon, the equity positions should not be expressed through weekly calls, even if some short-term bets might be sprinkled into the portfolio as opportunities arise.

Fundamentals: For better or worse, my analysis considers financial fundamentals, including both current valuation metrics and a reasonably underwritten medium-term outlook. On capital structure, I appreciate a “margin of safety”, but I’m also intrigued by leverage when it’s opportunistic and asymmetric, particularly in a positive turnaround story.

Technical: I like stocks that are showing reversals after long periods of one-way price action (i.e. coming out of multi-year troughs or coming off extreme peaks). My favorite sentiment is a stale consensus.

Allocation: Four core longs will anchor the portfolio with the balance split among four smaller allocations with shorter-term considerations. The portfolio is concentrated in fairly small and illiquid names, which prioritizes total return over volatility.

…

This model portfolio shouldn’t be taken as “investment advice” but rather a quantifiable ledger. At a minimum, it should provide a transparent expression of my evolving convictions, which I hope will provide significantly more clarity and value to the reader.

This portfolio will be dynamic. My intention is to introduce any new positions with relatively detailed posts outlining the idea. Any update to the portfolio will come with a note to readers explaining the rationale. Without further ado…

What stock do you secretly like but are too ashamed to tell anyone? Here is mine…

Read more

This is not a projection of what might go wrong, its a post-mortem on what already has.

Loans originated by the used-car dealer Carvana are going bad, fast. Loans that were packaged into asset-backed securitizations (ABS) and sold to investors over the past four years are realizing pronounced losses. This isn’t just a problem for residual speculators. Traded and rated notes, still valued above par today, are heading for losses.

Credit rating agencies, so far, have only affirmed their ratings while clinging to stale assumptions. Noteholders remain either blissfully unaware or unwilling to prompt a public markdown cascade.

None of this matters. Even if impairments haven’t been recognized, the underlying losses have already occurred. The market just has to catch up.

Today, I crawl into the depths of these subprime securitizations to determine which structures will break, and how deep the cracks go. Further, what does it mean for Carvana itself?

Read more

For years, the most popular children’s gaming platform, Roblox (RBLX), has been accused of providing a playground for predators. Roblox users, mainstream media, and financial analysts have exposed a consistent and disturbing reality of the platform — adults and children mingling freely, large groups seeking child pornography, and sexually explicit games available to users as young as five years old. The grooming of children that begins on Roblox has repeatedly led to a horrifying yet predictable pattern of digital exploitation, real-life abuse, abduction, and rape around the globe. Many perpetrators will go on to victimize dozens of children.

According to Roblox, one third of its users are under the age of 13 — the actual proportion is likely higher given the lack of any meaningful age verification.

These allegations against Roblox have been covered by Bloomberg, Business Insider, BBC, Daily Mail, PC Gamer and many other credible outlets. They’ve also been addressed in more detailed manner by the Edwin Dorsey’s The Bear Cave starting with Problems at Roblox back in 2022, and short-seller Hindenburg Research’s Inflated Key Metrics For Wall Street And A Pedophile Hellscape For Kids, published in October 2024. A single incident might be written off as a bad apple that slipped through the inevitable cracks of responsible moderation, but the sheer volume and consistency of these cases paint a remarkable pattern of gross negligence, if not something darker.

Since examples of the sexual conduct prevalent on Roblox has already been well documented, I won’t repeat them here. Curious readers can read the reports linked above or simply go to Roblox.com and type in keywords like “adult” or “XYZ” to see what the platform is serving to children right now.

But on Wall Street, Roblox is a darling.

In the last quarter, the company posted record numbers with accelerating revenue growth and user engagement. The massive growth in users was largely driven by the viral hit game “Grow a Garden”, a Farmville-style game which was first released in March 2025 and exploded in popularity throughout the summer. In June, Grow a Garden smashed Fortnight’s global record for concurrent online users in a single game with over 20 million players.

On the back of this success 2Q25 Roblox upped its fiscal year guidance, sending its share price to a new all-time high, implying a market capitalization over $95 billion. Since Hindenburg’s scathing report was published ten months ago, the share price has tripled. (The company has yet to produce a GAAP profit).

Among the sell-side analyst community, the RBLX enjoys wide support with 24 buy ratings, 8 holds just 4 sells. The average year-ahead price target stands at $141 per share.

Given that “safety issues” (a rather insincere obfuscation) are well known, the market simply assumes that whatever risk it brings must already be priced in. If it didn’t matter then, why should it now?

While scandalous allegations have been made for years, Roblox has yet to experience the true scandal. That is, despite the numerous reports of ongoing child endangerment, there has yet to be a true public or legal firestorm — a widespread and viral acknowledgment of the platform’s problems and patterns of neglect. That storm has arrived.

Over the past several weeks, a flurry of lawsuits have been filed against the company from across the country, describing horrifying stories of child abuse and laying the groundwork for a federal class action suit. Last Wednesday, a US Representative Ro Khanna (CA.D) demanded answers from Roblox. Last Thursday. Louisiana’s Attorney General filed a civil suit against the company calling it a “perfect place for pedophiles”. Last Friday, Chris Hansen of To Catch a Predator has confirmed he will soon be releasing a documentary about abuse on Roblox.

But the intense viral backlash — maybe the most significant factor in Roblox’s woes — is about a guy named “Schlep”. Over the past week, under intense public outrage and growing market scrutiny, Roblox has made a series of public relation missteps that only continue to compound.

Schlep and the Streisand Effect

Schlep is the digital persona of a 22 year-old young man named Michael from Texas. Schlep was an avid Roblox player growing up before being groomed on the platform by a well known developer beginning at the age of 13. At 15, Schlep was hospitalized after attempting to take his own life because of the experience. Despite his mother’s outreach to the company, Roblox refused to act and the abuser remained on the platform before eventually being banned years later.

Motivated by his own experience and the alleged refusal of Roblox to respond to detailed reports of abuse, Schlep began to fight the problem of pedophilia on the platform by himself, staging “sting operations”. Over the past year, Schlep’s investigations, in partnership with other internet “predator catchers” and law enforcement, has led to the arrest of six individuals. Each instance, including the Roblox interactions that laid the groundwork, are documented on his YouTube channel.

These videos were surely drawing some negative attention to the company, and one could understand why Roblox would want to stop this “sting” activity regardless of their actual moderation deficiencies. But the situation escalated dramatically on August 8th when Roblox terminated Schlep’s accounts and sent a cease and desist letter, threatening legal action if he continued.

But rather than squashing this pesky PR problem, Schlep and Roblox went viral — a textbook example of the Streisand Effect. By banning this “vigilante” Roblox only solidified the narrative; Schlep is a hero, Roblox is a villain.

Here is a quick timeline of Roblox’ rocky week:

• August 8 - 12th:

  • Schlep announces his termination from Roblox and cease and desist letter

  • Hashtags like #FREESCHLEP began trending on X, and large public accounts across the social media landscape began to amplify this story to a wide audience, including those previously unaware of the issues with Roblox. "Virtual protests” pop up on Roblox (ironically, they are quickly banned)

  • Social media momentum continues to grow by the day

  • Indonesia mulls banning Roblox pending transparency in moderation

• Wednesday August 13th:

  • In response to growing criticism, Roblox publishes a blog post “More on our Removal of Vigilantes From Roblox”, effectively doubling down on their position that its content moderation is effective while comparing the vigilante activity to the predatory behavior it targets

  • US House Representative Ro Khanna (CA. D) creates a petition asking for transparency from Roblox regarding their child safety issues

  • Qatar bans Roblox over safety concerns

• Thursday August 14th:

  • Louisiana AG announces it is suing Roblox, while calling the platform a “perfect place for pedophiles”

  • The largest Roblox streamer on YouTube, KreekCraft, quits the Roblox Star Program, along with several other major influencers

  • TD Cowen research note warns of legal risk, and reiterate sell rating $54 price target (after market hours)

• Friday August 15th (market hours):

  • Oppenheimer research notes “safety risk” but reiterates buy rating with $158 price target pre-market

  • Roblox responds to Louisiana suit denying claims

  • Citi says it’s “not concerned” about the Louisiana lawsuit because it is “only one state” and doesn’t see engagement declining. Maintains buy rating at $152 price target

  • RBLX opens down ~9%, and ends the day down 6.4%

• Post-Friday Close:

  • Chris Hansen confirms he is “deep” in the development of Roblox documentary and has already interviewed law enforcement and victims

  • Roblox releases two updates, “Clarifying Our Policy on Romantic and Sexual Content”, and “Update On Our Safety Initiatives (video)” promising further incremental changes

  • Petition to remove CEO David Baszucki reaches 100,000 signatures

  • Negative Roblox and awareness of the “Schlep” situation continues to escalate across social media, mainstream media, and YouTube

In each of its responses, Roblox has essentially denied there is any problem at all while begrudgingly announcing incremental safety improvements. It has been met with jeers. On X, posts from the Roblox account or David Baszucki are immediately labeled with damning “community notes” and the comments are flooded with negative responses. On YouTube, videos condemning the company’s general failure in moderation and specific response in this case are amassing millions of views.

Even Roblox’s new changes to its moderation policy opened up some uncomfortable questions. For example, in Clarifying Our Policy on Romantic and Sexual Content, Roblox states:

“We’re clarifying our policy that prohibits romantic and sexual content to also explicitly prohibit content, settings, or behavior that implies sexual activity.”

So prior to Friday, Roblox did not police “implied” sexual activity, which of course ran rampant on our site? It’s been a free-for-all so long as you maintain a modicum of plausible deniability or a strategic misspelling.

Or…

We’ve seen that Social Hangouts… can sometimes contain inappropriate user behavior when they include certain settings. We’re updating our policy to restrict social hangout experiences depicting private spaces (such as bedrooms and bathrooms) to users who are ID-verified and 17 or older. Additionally, experiences that primarily take place in these private spaces, and in settings intended for adults (such as bars and clubs), will also be restricted to only ID-verified users 17 or older.

Not only do they acknowledge that “private” bedrooms and bathrooms are being used to simulate sexual interaction between users, but up until now these games have been available to all ages. Even under this new policy, a 17-year old — still a minor — would be able to access these private bathrooms, bedrooms, bars and clubs. (After being presented with this point, Roblox has since clarified they are “considering” changing the cutoff to 18).

Finally, Roblox points to its National Center for Missing & Exploited Children (NCMEC) reports (which are required by law) as both evidence of its moderation, and the relatively low frequency of reports as proof of the platform’s safety. But clearly the numbers should be read in the opposite direction. Despite running the most popular platform for young children, the company produces only a tiny fraction of NCNEM reports submitted by social networks, even after adjusting for user base, suggesting incredibly weak reporting standards.

In all fairness, some problematic games have been removed — like this one titled “XYZ Public Bathroom Simulator Hangout” which was active on Roblox site since 2024 and visited over 7 million times.

Yet, as of Sunday evening, the first result in a query for “Public Bathroom Simulator” on Roblox’s discover page revealed plenty of new options, like this one below — created on Saturday afternoon (after Roblox’s alleged rule-change) which has already been visited 13,700 times, with 243 active users and no age restrictions. The purpose of this game, apparently, is to “use the toilet, wash your hands, and strike up some conversation!”

While these bathroom examples stand out as strikingly lewd, there are plethora of highly suggestive, completely unrestricted, games that are equally or even more disturbing; titles like “Boys and Girls Immersive Party Roleplay” — currently hosting 695 active players of all ages.

And as the internet firestorm continues to rage, some sleuths are digging up new dirt.

For example, Chief Safety Officer Matt Kaufman has previously identified his personal Roblox account as “RbxRocketMan” at official Roblox events in the past. RbxRocketMan’s user profile — still publicly accessible as of writing — shows “badges” earned from at least six different highly suggestive games (like “condo games” which are rooms used for sexual simulation) many of which have since been banned after backlash. Screenshots from these “games” are below.

The first "favorited” game on RbxRocketMan’s profile is “Meep City” a game notorious for its private “party room” features and frequent sexual content, that was eventually pulled from the platform in response to years of backlash. The game has since been reinstated, with private rooms disabled.

Assuming this is indeed Kaufman’s account, the best possible explanation is that he was exploring these games for moderation purposes (though moderating through a personal account seems like an unusual method). And even if the purpose was moderation — these rooms remained online long after RbxRocketMan’s visits, meaning that his explorations found them to be suitable for children. Only after the intense public backlash (and market cap losses) did Roblox begin to remove some of this content.

While maybe not outright damning, these breads crumbs mirror prior concerns about Roblox staff, like those outlined in Problems at Roblox which showed that the former Head of Moderation and key social media managers were engaging with or producing explicit sexual content with childish themes on their personal social media accounts. When combined with high-profile Roblox developers being convicted for grooming and abuse, and the mind-numbing refusal of the platform to introduce basic safety features (except under extreme market pressure), it becomes harder to assume that this deviant behavior is merely an external force of bad actors.

Legal Firestorm

This crescendo of public attention has been mirrored by crescendo of legal action. So far, I have identified thirteen separate civil lawsuits in 2025 alone naming Roblox as a defendant, spanning seven state courts and three federal US district courts. This list is probably not exhaustive. All of the cases outline strikingly similar patterns: including grooming, exploitation, abduction, and rape, with victims as young as eight. All of the interactions began on the Roblox platform.

Nine of these cases have been filed since July alone, with the pace rapidly increasing over the past two months. Most suits are being filed by a handful of mass-tort and class-action litigators (Anapol Weiss, Milberg, Dohman Law) with strikingly similar legal arguments and fact patterns across many state and federal jurisdictions.

This is just the tip of the iceberg. Milberg — a self-described pioneer of class-action lawsuits — indicates it already represents 400 victims of sexual exploitation on Roblox. In short, the “seeding” of a class-action lawsuit is well underway, with this flurry of independent lawsuits across the country meant to demonstrate a nationwide concern.

Beyond these direct suits from citizens, Roblox received a subpoena from the Attorney General of Florida concerning child endangerment in April 2025, long before the Louisiana AG’s lawsuit was announced last week.

Roblox has dodged or settled lawsuits in the past — including ones related to child endangerment — but it has never received an onslaught like what we are seeing today, including the involvement of state AGs (and maybe eventually the feds). Despite many of these lawsuits (and the Florida subpoena) occurring before Roblox released its 2Q25 10Q on July 30th, the company makes no mention of the cases in their legal contingencies or subsequent events, meaning the company continues to view them as “immaterial”.

While online platforms enjoy broad protection under Section 230, Roblox is far from bulletproof. Section 230 would not shield the company from design negligence claims, fraudulent misrepresentation in public statements, or sex-trafficking. Even in its own risk disclosures, Roblox acknowledges that it could be found criminally liable for child pornography or sexual exploitation of a child domestically and abroad. Per its 10K filing:

“non-compliance by us or allegations of non-compliance by us with respect to U.S. federal laws on child pornography or the sexual exploitation of children could significantly harm our reputation, create criminal liability, and could be costly and time consuming to address or defend. We may also be subject to additional criminal liability related to child pornography or child sexual exploitation under other domestic and international laws and regulations.”

Notably, I emphasize the role of Robux — the game’s virtual currency. Robux can be transferred between players for a 30% fee through in-game purchases and cashed-out by developers for 65% fee (the fees are taken by Roblox as profit). Many predator cases describe Robux payments being used to coerce children into providing sexual content, or simply being used to facilitate financial transfers between adult predators (even if the actual content is exchanged off the platform). In these exchanges Roblox is both laundering and directly profiting from the financial exchange of criminal material. (This currency connection is already the subject of illegal gambling accusations — a separate legal can of worms we will ignore for this piece).

Roblox faces numerous legal vulnerabilities, ranging from civil penalties, to mandated reforms, and even potentially criminal liability. Even if a settlement is the most likely outcome — largely to avoid discovery — this issue is not going away. And the more Roblox is forced to speak about it honestly, the worse it becomes.

Priced In

The bull case for Roblox is simple. The company is putting up great numbers with growing users and viral hits. “Safety” allegations have been made before and the “shorts” have gotten smoked. This too shall pass.

At 20x sales after a nearly 100% YTD run, the market has certainly believed the bull case. Only on Friday, after the Louisiana AG suit, did we see the first real admissions of investor concerns.

We don’t know if Roblox will be held accountable for its blatant moderation failures. We don’t know when it that will occur, and what the consequences would look like — undoubtedly there will be a massive legal battle on the way. I also recognize that there are many players who genuinely love the platform and have not been exposed to harmful content. But a serious reckoning and reform is long overdue — and might just be arriving.

At what point does an investor admit that a children’s platform crawling with pedophiles is a material risk? At what point does an analyst admit that endorsing such a platform is a career risk? At what point do you stop believing that “the predators are priced in”?

Let me tell you about an Ugly Duckling.

This company operates in a capital intensive and dangerous business. The common shares sit behind mountain of high-cost senior capital. It’s public share price plunged 90% since its debut via SPAC merger in 2023. Its 2023 annual filing came stamped with the scarlet letter — a “going concern” warning.

The company is not clean. An investment is not “safe”. But if you can look through the risks, a compelling narrative emerges.

The company saves lives and property from intensifying risks. Its revenues have grown at a 97% CAGR over the past seven years, and its earnings are benefiting from operating leverage. The wave of capital investment has created rare assets that generate strong cash flow. Last year, the company flipped cash flow positive and shed its going concern precaution.

Over the past year, the company crossed a critical point of inflection, boosting hopes for its shareholders. But that future just got much brighter. New legislation, new budgets, and new executive orders all announced in the past eight weeks stand to accelerate this industry and unlock a new phase of growth.

What happens when a company teetering on the edge of bankruptcy tips in the right direction? When new legislation brings a windfall? When a capex leads to cash flow? When leverage becomes a springboard?

When the ugly duckling sheds its feathers will the flock begin to notice?

The Duck

The Duck is Bridger Aerospace (BAER). Founded in 2014 by former Navy SEAL Tim Sheehy and based in Belgrade, Montana, the company operates a fleet of specialized planes used to fight wildfires. (If the name sounds familiar, its because Sheehy was elected to the US Senate in October 2024 — more on this later).

To scale from a small time operator to a serious national player, the company took growth investment from Blackstone Tactical Operations group, with funds used to expand its aerial fleet. Blackstone remains the largest equity holder through its common and preferred equity positions. In 2023, Bridger came public via a merger with SPAC Jack Creek Investment Corp, and today trades under the ticker BAER. Common equity warrants trade separately under BAERW.

Super Scoopers

Early in its life, Bridger operated a handful of small planes for reconnaissance, surveillance, and coordination — serving as a “quarterback” for large tankers dumping water or retardant. But since 2020, Bridger has expanded its fleet to include 14 planes, including its prized assets — six “Super Scoopers”.

Super Scoopers are a unique firefighting asset. Rather than loading up water or retardant on the tarmac for a single drop, these planes can repeatedly refill their tanks by skimming the surface of a nearby bodies of water. Smaller and more maneuverable than large tankers, Scoopers can fly lower (dropping water more accurately) and refill their tanks much quicker — making them a highly effective component of the aerial firefighting arsenal.

For more on these planes, see this Business Insider rundown with Bridger.

Scoopers Scoopers (CL-415s) are highly specialized, rare aircraft. These planes (and the predecessor CL-215s) were only produced by from 1993-2015 and less than 100 exist globally, almost entirely owned by governments. Bridger’s six aircraft is the largest private fleet and the only scoopers owned in the United States1.

Given the unique capabilities and rarity of the assets (especially in the US), demand for Scoopers significantly exceeds supply, with roughly a third of federal flight requests going unfulfilled.

Beyond its active US fleet, Bridger also won a bid from the Spanish government to return four older CL-215 Scoopers to service for use in Europe. Two of these aircraft should be active for the 2025 wildfire season in Europe this year, though the company has yet to reveal contracts or estimated financial uplift.

The global supply constraint, combined with growing concerns of wildfire destruction has prompted two OEMs to restart production of scoopers. First, a new CL-515 Super Scooper is under development by Canada’s De Havilland, but first deliveries are not expected until 2028 and nearly all of the order book is slated to head to Europe and Asia. Second, France’s Positive Aviation Inc is in the early stages of development of a FF72 aircraft and signed an MOU with Bridger as the exclusive North American customer, though the earliest planes aren’t expected until at least 2029. In other words, the US market will remain supply constrained for years to come.

While the Scoopers and smaller reconnaissance plane constitute the majority of Bridger’s business, the company is also developing software data platform — Ignis — and has acquired a specialty maintenance and repair organization FMS Aerospace. Given the level of bespoke work in upgrading and maintaining these specialized planes, the acquisition of FMS was a strategic play to continue to vertically integrate the operations, including installing specialized equipment (i.e. sensing technology on smaller planes) and reducing ongoing maintenance expenses. FMS also opens the door for more consistent year-round revenue and non-firefighting assignments.

Contracting & Competition

The vast majority of Bridger’s revenues are derived from government contracts, with most funding coming through a web of US federal agencies. The budget is provided through the Department of the Interior (primarily, the Bureau of Land Management) and the Department of Agriculture (the US Forest Service) which partners with state-level operations.

The company enters into both “Call When Needed” (on-demand) and “Exclusive Use” (dedicated) contract structures, which generate revenue for both standby and flight hours. For the peak wildfire season (July - October), most of the company’s assets are contracted by a state or federal government agency, while shoulder months see infrequent, sporadic utilization. Bridger has achieved a 100% federal contract renewal rate since its inception.

Bridger is one of several private contractors in this space. Its primary competitor is Aeroflite, the US subsidiary of Canada’s Conair, which is the largest private aerial firefighting company globally, with operations in Canada, US, Australia, and Europe. Aeroflite is also the only US-based competitor with access to Scoopers, while other companies only operate large tankers or helicopters. Even as Aero-flight competes with Bridger in the Scooper market, demand still significantly exceeds supply.

Inflection

Since 2020, Bridger has spent ~$200 million to bring the six-plane scooper fleet into operation. This wave of capex continues to burden the company’s balance sheet through its expensive senior debt and preferred equity, but the company appears to have turned.

Now online and operational, these Scoopers are highly cash-flow generative. The company has not required outside capital since the SPAC infusion, and it generated positive free cash flow and cash from operations for the first time in 2024. With strong demand for its assets, cash flow positivity, and limited capex, the company is beginning to reap a return on its years of investment.

In 2025, the company expects further growth in revenue, EBITDA and cash flow.

All of this bodes well for Bridger. But renewed federal focus on wildfire prevention is set to provide a massive new tailwind for the business.

Legislation

In the past several years, North America has experienced a series of highly destructive wildfires that have forced the attention of the public and lawmakers. In 2021, an extreme wildfire season blanketed much of the continental US in smoke. The 2023 Maui wildfires were the most deadly in a century, creating $5.5 billion in damages. Most recently, the January 2025 wildfires devastated regions of Los Angeles, with the economic toll estimated to reach $50 billion.

As peak wildfire season is extending later into the fall, and fires are more frequently impacting historically low-risk areas like the East Coast, there is a growing need for more comprehensive year-round solution.

While fingers point blame in different directions (climate change, forest management), the reality is the governments are usually reactive, not proactive, to crises. This case is no different. Recent wildfire devastation has forced an overhaul in federal wildfire fighting strategy.

The need to streamline federal contracting and prioritize aerial firefighting assets holds widespread support. A bipartisan Wildland Fire Mitigation and Management Commission, created by Biden’s Infrastructure Investment and Jobs Act (IIJA), found the need for enhanced federal contracting for aviation assets that prioritizes strategic value over cost.

“The Commission recommends that contracting process should meet operational demands, including the option of reliable longer term contracts for baseline capacity needs and every effort should be made to improve the effectiveness and efficiency of the contracting process. The consideration of longer-term contracts should include both multi-year contracts and contracts that extend for longer operational periods within the year. More generally, the type of contract used should meet the needs of a national strategy rather than allowing cost considerations and established procurement policies to override programmatic needs. Aviation operations should drive contracts and not vice versa.”

Now, concrete actions are being taken.

Federal Overhaul

On June 12, 2025, Trump’s executive order “Empowering Commonsense Wildfire Prevention and Response” fundamentally reshapes how the federal government will handle wildfire prevention.

The order consolidates federal wildfire operations across departments into a single newly created entity — US Wildland Fire Service — to be housed within the Department of the Interior. (Along with this consolidation, the Executive Order also calls for the Department of Defense to auction off excess aircraft for wildfire fighting purposes.)

This new agency consolidates firefighting contracting, preparation, and operations into a single streamlined agency. Rather than a patchwork of assets and agencies, this move paves the way for a single continental dispatch model. Ideally, this should lead to more efficient asset utilization and coordination across the nation, which currently suffers from the disjointed approach.

This new agency is also set to see a massive increase in budget for FY2026, which begins in October 2025. As shown below, the total federal wildfire budget is set to jump over 67% next year, from ~$4 billion across departments to $6.5 billion consolidated in a single agency.

This budget includes major increases across line-items, including a 3.6x year-over-year increase for “suppression operations” within in the DOI’s budget, and the establishment of a new intelligence and technology category.

As federal contracts account for the vast majority of Bridger’s revenue, this increase in budget appropriation is a major boon. As one of only several aerial contractors, Bridger should receive at least some part of this growing pie. A consolidated federal framework also stands to improve efficiency and performance.

• Streamlined Federal Contracting: Simplifies bidding process, likely leading to fewer but larger and more comprehensive federal contracts.

• Higher Utilization: A consolidated federal dispatch model should lead to higher asset utilization and flight hours, as opposed to planes being held on standby through Exclusive Use contracts to a single agency.

• Longer Contracts: As recommended by the commission, a larger budget allows for extended or year-round contracts with planes pre-positioned in high risk areas outside peak season.

• New Categories: New intelligence and data contract possibilities with the company’s Ignis platform and FMS.

• Expanded Fleet: Potential to acquire airframes from the Department of Defense, in mandated sales per the Executive Order.

Even in an above-average fire year in 2024, Bridger still maintained excess capacity on its Scooper fleet — operating 350 flight hours vs. 450/500 hour maximum before heavy maintenance is required. Incremental flight hours and extended contract terms would be highly accretive revenue.

The company’s 2025E guidance assumes no increase in federal budget (nor revenue from the Spanish Scoopers), and with contract terms locked for the current fire season, the new budget won’t have a meaningful impact on this year’s financial performance. But the company seems positioned for contract expansion in 2026.

The Insider Angle

While there is bipartisan support for improving wildfire prevention, we should acknowledge political factors at play.

Founder and former CEO Tim Sheehy stepped down from the company in June 2024 to run for the US Senate as a Republican, ultimately defeating three-term incumbent Democrat, Jon Tester. As one of three senate seats that flipped red, Sheehy’s Montana race was critical in delivering the Senate to the Republicans. Trump both endorsed and campaigned with Sheehy.

In just his first several months Senator, Sheehy has co-sponsored at least five bills relating to federal wildfire strategy, including the Aerial Firefighting Enhancement Act of 2025, which was passed into law on June 12, 2025 (with unanimous approval in the Senate). With the new US Wildland Fire Service set to be led by Trump appointees, its likely that Sheehy has already played some role in the structure, budget and priorities of the agency, if not directly then through his relationship with the administration.

Even before the executive order was announced, Bridger’s management suggested these changes were coming in the 1Q25 earnings call, implying foreknowledge if not influence.

“The good news is there has been significant movement recently in Congress where multiple bills have been introduced in the House and the Senate to improve the efficiency and regulatory structure impacting wildfire response. Additionally, we anticipate executive actions as well as additional agency changes that will support and streamline the ability of wildland firefighters to respond quickly and directly to wildfire crises that threaten the country […] Stay tuned as this develops further."

Notably, Sheehy has moved his assets — including his sizeable Bridger equity stake — into a blind trust just one week prior to the Executive Order, perhaps in anticipation of what’s to come.

Leverage

If all of the above sounds like a good story, you might be wondering why this company trades at just $1.84 per share.

Bridger’s common equity makes up only a small fraction of its capital. With a enterprise value of $665 million today, its $100 million market capitalization is subordinated to $200 million of debt, and $387 million of preferred equity. Relative to 2025E adjusted EBITDA guidance, this implies a relatively healthy 14.5x valuation.

Bridger’s senior debt includes $160 million of 11.5% Series 2022 bonds due in September 2027, along with a recently completed $46 million sale-leaseback financing arrangement that will likely be used to pay off existing equipment loans. This high-cost financing demands over $20 million in cash interest payments annually — a cash flow burden that seriously jeopardized the company’s financial standing and forced the going concern warning in 2023.

But with net leverage now <4x, the company should be able to refinance these bonds before their maturity in 2027, perhaps saving a bit of interest expense in the process.

Meanwhile the preferred equity, owned by Blackstone, accrues at 7% annually through 2028, with the rate stepping up to 9% and then 11% before coming due in 2032. This preferred equity doesn’t eat cash flow, but its ongoing accretion takes value from the common equity at a rate of >$25 million per year. By the time of its maturity in 2032, these preferred equity shares will have a liquidation preference of $729 million — exceeding the company’s entire valuation today.

The combination of the preferred equity accrual plus other dilution from things like stock-based compensation will have a meaningful drag on the common equity valuation going forward, if the company only grows modestly.

But leverage works in both directions. Only a small increase in total valuation has an outsize impact on the common equity. And if the company sees a near-term boost in revenue, the benefit could quickly outweigh the effect of the preferred equity. With significant headroom before the preferred equity / warrants conversions are in the money, the share price can increase 6x before the dilutive effects of conversion.

For example, every 10% increase in total valuation would increase the share price by 65% (at least in theory). Alternatively, assuming a constant 15x multiple, every $3.6 million of incremental EBITDA is good for $1.00 of share price appreciation (vs. $1.84 current price).

Near term catalysts for growth (Spanish Scoopers, increased federal budget), combined with operating leverage (through fixed costs) and capital structure leverage sets up an asymmetric risk / reward profile. It’s impossible to know the exact impact that the new budget will have on operations, but an illustrative example, below is an implied share price assuming a 15x multiple on an EBITDA range of $45 - $65 million (vs. the midpoint of 2025E guidance of $46 million).

In the worst case scenario, the common equity may go to zero. But improved performance and a rebound in sentiment could lead to quick appreciation. As the common stock traded over $4.00 as recently as January, a 2.0x return from here would constitute a more of a rebound than a breakout.

(And if the common equity already represents some sort of “option value”, its worth considering the publicly traded warrants — trading for $0.08 under the ticker BAERW — with a strike price of $11.50 and an expiry in January 2028).

Summary

I called Bridger an Ugly Duckling for a reason. The company is subject to year-to-year variability in weather and fire activity. Management has had a checkered history of underperformance vs. expectations in its limited public history, and has faced accusations of accounting deficiencies. The common equity sits behind significant senior debt and accreting preferred equity.

But I think there is an underappreciated equity story here that could play out over the next twelve months. The turnaround in cash flow in the past year has reduced existential risk and the market does not seem to be appreciating the budget and agency restructuring announcements of the past month. I expect that management will provide specific concrete color on the benefit of these changes in the 2Q25 earnings call on August 7th. Improved federal contracts serve as major catalyst leading into 2026.

With BAER shares already trading up 85% since bottoming at $1.02 in early April, it seems that some durable momentum is building for the first time in several years. In short, I believe Bridger is an interesting asymmetric market opportunity over the near term.

There is nothing new under the sun. Throughout time, the state has shaken hands with commerce for self-interested gains. Yet, we still expect a certain level of decorum — closed door deals, public denial, the projection of propriety. That’s what makes this current moment so unique.

In this Golden Age of Graft, pretense is checked at the door. We’ve moved from “10% for the big guy” hidden in private emails to the in-your-face, pay-for-play paradigm of the new Trump era. From self-serving backers, to the Trump kids’ expanding public interests, to eponymous cryptocurrency — a new paradigm of money, access and patronage has taken form.

It may seem like ages ago (two months), but there was a time when Wall Street feared Trump. But ever since the Tariff trade turned into a TACO (Trump Always Chickens Out), the vibes have swung in the opposite direction. With the administration’s pivot from stock puncher to stock pusher, a new investment theme has taken shape. A certain niche asset has been among the biggest winners in this rally — publicly traded entities with proximity to the President.

Take, for example, Colombier Acquisition Corp (CLBR) — a SPAC backed by Trump affiliate Omeed Malik that announced a merger with GrabAGun, an online firearm retailer which elected Donald Trump Jr to the Board of Directors on April 29.

Or, Kindly MD (NAKA) — now known as Nakamoto holdings after announcing a merger with Trump crypto advisor David Bailey’s firm on May 12 — transforming a medical marijuana prescription business into a crypto holding company.

Even Bailey can’t believe his fortunes. In his words from March:

“If a year ago you put me into hypnosis and said, ‘Describe to me your deepest dreams of what could happen,’ this would be straight-up fantasy,”

Or take bitcoin miner Gryphon Digital Mining (GRYP), which announced a deal on May 12 with HUT 8 (HUT) and Dominari Holdings (DOMH) to acquire and develop the American Bitcoin, a mining company supported by both Eric Trump and Donald Trump Jr.

This just scrapes the surface.

Commerce Secretary Howard Lutnick’s Cantor Fitzgerald has spawned multiple SPACs in the past several months which are trading significant premiums to NAV even before merger announcements. Prominent Trump supporters like Bill Ackman and Peter Theil seek to reap returns on their political investments as the administration contemplates the privatization of the Fannie Mae (FNMA) and Freddie Mac (FMCC), and Palantir (PLTR) is awarded wide-ranging government contracts. Top TRUMP coin holders gain presidential access in private dinners and companies like Freight Technologies (FRGT) lever up to buy $20 million of TRUMP with explicit intention of influencing tariff policy.

It is impossible to map a complete web of influence — the tendrils are too many, too intertwined. But it might look something like this.

We are verging into manic territory, but hear me out…

Read more

By: Daniel Saedi, Former Banks Investor at Bridgewater and Founder of Minerva, and The Last Bear Standing.

The Saedi Summary

Social Finance or SoFi (SOFI), once a silicon valley success story leading the DeBank movement, has transformed into a regulated specialty consumer lender with all of the same challenges and significantly worse positioning than its new competitors.

In the past three years, SoFi’s business model has shifted dramatically — from high quality student loans to unsecured credit card consolidation and personal loans. An OCC bank charter has allowed for cheaper funding and faster growth, but the company continues to count its profits as if it were simply an originator, not the loan-to-hold bank model it has largely adopted.

By valuing its balance sheet loans well over par — an unusual and aggressive accounting practice — SoFi front-loads its earnings and equity capital accretion. But this aggressive approach also brings mark-to-market volatility with immediate impact on its financial and regulatory standing. Today, nearly a third of its tangible equity capital has come through the cumulative write-up of its loans, even as its credit performance has materially declined. If it were to value its loans at cost with reasonable credit loss assumptions — like all of its peers — SoFi would likely trip bank capitalization requirements, prompting immediate corrective action.

What do you get when you have an undercapitalized, fast growing bank making riskier and riskier loans? History has never been kind at the end of a credit cycle to these players, why should this time be different?

Don’t Bank. Sofi.

Social Finance has been one of the Silicon Valley darlings of the 2010’s fintech revolution. Founded at Stanford in 2011, it began as a student loan refinancer and eventually became a private student loan originator, following the success of peer-to-peer lending startups like Lending Club. SoFi benefited from the FinTech and SaaS craze in the valley at the time, with a bombastic CEO who insisted during the run up that SoFi was a technology platform, not a bank. All across San Francisco, and eventually even Wall Street, during the 2015-2017 era, you could find their marketing slogans like “Don’t Bank. Sofi” and “Even banks don’t want to be banks anymore”.

But what was so special about SoFi, and how was it able to distinguish itself from existing peer to peer players like Lending Club, consumer focused BNPL fintechs like Affirm (at least early) and tech enabled originators like Rocket? Brilliant distribution, a focus on a safer but fast growing segment of the consumer lending market in private student loan refinancing from top institutions and graduate programs, and an inability of older consumer credit models to incorporate income curves and growth for that market.

Back in 2015, with capital cheap and a good story, SoFi was off to the races. It made sense for someone with an MBA from Stanford or Harvard to consolidate their federal undergraduate and graduate loans and get some extra cash to spend at a 3% secured interest rate.

In an era of growth at all costs, SoFi was able to effectively turn cheap debt capital from loan hungry banks in California like East West into its customer acquisition tool. It may seem like forever ago, but companies back then in its cohort of West Coast upstarts were selling useless products at massive losses while wasting equity capital on ineffective marketing campaigns plastered across the internet. At one point we even had a company that promised to send somebody on a scooter to valet your car for cheaper than it was to park on the street.

SoFi was able to flip this model on its head and use ZIRP directly, lending at very thin net interest margins (NIMs) and packaging up loans whenever it needed cash or liquidity. The result? Nearly $3 billion in total capital raised from funds like SoftBank, IVP and Discovery before Chamath took the company public via SPAC in 2021 at a $9.5 billion valuation.

Despite proudly claiming that not even banks wanted to be banks, Sofi eventually became attracted to the prospect of consumer deposits vs. expensive shadow bank financing and acquired a national banking charter through the acquisition of Golden Pacific Bancorp in February 2022. SoFi became the very institution it was trying to displace.

What Happens When a Silicon Valley Darling Becomes a Bank?

Every bank is a company. But not every company is a bank. Once you obtain a bank charter and fall under the purview of the Federal Reserve and OCC, it doesn’t matter if you were funded by Founders Fund, Softbank and QED, you are now more similar to the small mechanics bank in Gary Indiana than you are to OpenAI and Facebook.

As someone who has covered and studied banks for nearly two decades, studying credit cycles going back to the Roman Empire, there is one constant, timeless and universal phenomenon that emerges — somebody always thinks that they can lend better and begins to issue new credit quickly. Usually, they’re actually a bit better than the incumbents at the start of the cycle, but once their perceived edge starts to dry up or they fall under pressure to keep the growth engine going, they make riskier and riskier loans.

This is a tale as old as time — and we see it here again.

While SoFi had an early edge in extending lower interest rates to qualified borrowers at premier educational institutions in the US, that edge and technology does not apply to unsecured personal loans and credit card consolidations, which now make up the majority of SoFi’s growth and loan book. No matter the data, technology, or AI advantage, consumer lending is a game of margins and there is a reason credit card companies extend 20-34% APRs to individuals who have 700+ FICO scores — borrowers default on unsecured loans first and recoveries are very slim.

With an originator model, SoFi would simply continue to borrow money from banks to issue loans to consumers and then turn around and sell those loans back to the banks or to asset backed securitization vehicles clipping the interest difference, origination fee, acquisition fee, and servicing fee for continuing to collect payments.

But if you think you are uniquely good at finding diamonds in the rough borrowers, you want to capture the full economic value of the Net Interest Margin (NIM) through time, which only makes sense if you are a bank. Why? Because banks can create loans funded by deposits, which are much cheaper than lines of credit, and they get to decide when they want to mark things to market or hold until maturity, allowing them to smooth their returns (I wonder where private credit got this idea from).

SoFi told investors that they sought out a banking charter to (1) be able to offer a larger suite of financial products, including saving and mortgages, to their customers and (2) to access deposit funding directly from consumers. It was a major step in fueling the growth engine that began to pay off even as rates were rising.

Despite the Fed hiking 525 Basis points from 2022-2023, SoFi’s cost of funding rose far more gradually as it switched from expensive warehouse financing from other banks and financial institutions to retail deposits.

But being a bank brings new levels of scrutiny. Governments have known that banks have the ability to create money for centuries and therefore have attempted to regulate them much more stringently than regular corporations. This burden has prevented many fintech companies from attempting to be true depository institutions.

Amongst a whole host of regulatory requirements (AML, KYC, liquidity, stress testing and capital planning), the most applicable across time and space has been capital and loss reserving. In the modern version of banking regulations that we see today, all regulated banks have to maintain minimum capital ratios. They need to have positive equity and that equity must be relatively liquid in the event that losses are sustained.

Subscribe now

The Bank’s Books

Despite becoming a bank, SoFi still prefers to think of itself as an originator and distributor, not a balance sheet lender. While nearly all banks value their loans at amortized cost or “par” value with a reserve for credit losses (the Current Expected Credit Loss or “CECL”) approach, SoFi opts to value its loans on a “fair value option” — allowable under GAAP but highly unusual among peers.

The two methods result in dramatically different earnings and capital balances.

Amortized Cost Accounting: Under standard accounting, a bank does not “make money” in its P&L simply by issuing a loan. While the bank expects the loan to be profitable, it takes P&L earnings throughout the life of the loan as it actually performs.

When issuing a new loan, the bank will mark the loan at its amortized cost (or roughly “par” value), with a reserve for expected credit losses. Then, over time, the bank will recognize earnings through ongoing spread between the asset yield (interest received from the loans) and its funding yield (the cost of deposits). If the bank sells or securitizes a loan, it will record a gain or loss on the sale as it actually occurs.

In practice, this means that if a bank makes a five year loan, it will recognize the earnings throughout its duration. Likewise, the bank’s capital position will gradually grow as P&L earnings fall through to the retained equity. The bank can also update its credit risk over time through by adjusted the loss provision associated with the loan.

Fair Value Option: By contrast, a FVO approach effectively takes P&L credit for the loan on the day it’s issued. Instead of holding the loans at par value, the bank makes assumptions around prepayments, defaults, and a discount rate, which define the “fair market value” (FMV) of the loans it holds on its balance sheets.

Since a bank underwrites with the expectation of profitability, the FMV will almost certainly be higher than the par value of the loan, creating a day-one writeup in loan values. That writeup creates an unrealized gain that flows through the P&L and immediately accrues to the equity capital. Faster equity capital accrual also allows the company to grow faster without pressuring its regulatory leverage tests.

The FMV relies on management assumptions and dramatically accelerates earnings recognition and equity capital appreciation. While both methodologies are allowable under GAAP, the fair value option is aggressive and highly unusual for banks – in fact SoFi is the only bank of its peers that takes this approach for a majority of its loans.

Simply, SoFi counts its earnings as if it were a pure play originator and loan seller, despite the fact that its balance sheet loan book has grown 4x in three years. SoFi wants the best of both worlds — it wants the valuation and reporting benefits of the write-and-sell model with the reliable and lower-cost funding of a bank. But with $27 billion in customer deposits, regulators might see things differently.

SoFi’s Marks

Below shows the key input into SoFi’s FMV marks by loan category; personal loans, student loans, and home loans.

Using these assumptions, SoFi’s personal loan book was valued at 105.5% of par value as of its most recent quarterly marks. In other words, for $100 of unpaid principal outstanding, the company has recorded an asset of $105.5. By contrast, CECL accounting would value these loans well below $100, when including a provision for losses. On average, companies like COF, DFS, and SYF value their on-balance sheet loans at ~$92.

While the company provides support for these DCF model inputs, it seems the real justification for auditors is by referencing the “Sales Execution” of its recent transactions. In the most recent quarter, the company calculated a personal sale execution level of 106.2%. In other words, so long as SoFi can offload some loans at 106.2, it can mark all of its balance sheet loans at 105.5.

Indeed, in its most recent investor presentation, SoFi shows the evolution of its reported sales execution and FMV over time, implicitly acknowledging the importance of these figures in justifying its balance sheet marks.

Again, SoFi’s guiding principle here is far more aligned with its shadow banking past rather than the regulated bank it has become. The entire valuation approach assumes efficiently selling loans rather than holding loans. Which is odd, considering the company now holds a whole lot of loans…

Capital Impact

As spreads have tightened and sales execution has improved over the past two years, SoFi’s FMV markups has become an ever greater portion of its total reported profitability and, perhaps more importantly, its regulatory capital.

SoFi’s cumulative FMV adjustment of its loan book has grown to $1.4 billion as of 1Q25 — increasing both as a percentage of loans outstanding and regulatory capital. Today, this $1.4 billion markup constitutes 30% of the company’s Common Equity Tier 1 capital (CET1).

If SoFi was to hold its loans at par, the company’s CET1 ratio would fall to 10.7%, even before considering CECL loss reserves. If we assume a 5% CECL reserve, which is well inside SoFi’s 8% expected loss underwriting criteria for personal loans, the CET1 ratio would fall to 5.8%, well below the minimum capitalization threshold that would induce immediate regulatory action.

In other words, if SoFi simply followed the accounting practice of its peers, it would likely fall below the mandated minimum capitalization of 7%, triggering a Fed/FDIC’s Prompt Corrective Action (PCA) with an immediate halt on distributions and a formal recapitalization plan, backed by escalating supervisory oversight.

This question of accounting therefore is not trivial but critical. A reduction in these marks would materially reduce the bank’s equity capital, requiring the company to either dramatically reduce growth while building capital buffer or raise new equity which would be very negative for the stock.

And as we dig into the details, we see several causes for concern.

Things That Make You Go “Hmmm”

There are three red flags in the company’s FMV approach. We focus on the company’s personal loans, which are the largest – and highest risk – segment of loans. They are also the category held at the greatest premium to par.

Read more

In the ancient spectacle of the bullfight (or corrida de toros), the Matador must understand the bull's nature, anticipate its movements, and recognize the precise moment when its confidence becomes its vulnerability. The corrida is carefully planned and executed, layered into a series of phases that must each be completed skillfully before the killing blow (the estocada).

The Matador must remain calm, knowing that behind the frenzy is the inevitable exhaustion. And the moment that exhaustion appears, perhaps even before it is known to the bull (who has been in a frenzied attack for hours), the death blow is delivered.

In markets, like bullfights, the ferocity continues until it absolutely can no longer. If there is a singular task that every bull market in history has accomplished with ruthless efficiency, it is the silencing of contrarian voices by the end of its run. Those with the audacity to have warned of risks before the point of absolute exhaustion end up derided and ignored.

This effect persists well after prices have begun to crack, and when combined with the lessons of recent successful bouts of dip buying, create the “complacency” phase of perhaps one of the most acknowledged and ignored charts in markets.

While the length of this piece is…manifesto-y, it’s a necessary exercise in examining the market holistically. That’s because I am more convinced than ever that we have entered the complacency phase of the cycle.

While recent events have taken some of the air out of the most egregiously crowded trades in the market, we have so much further to go. Exceedingly few are truly positioned for the bear market and the bulls haven’t given up without a fight. Investors are frozen, even as they realize that things are, in fact, getting worse.

However, there are still a few that see this for what it is. One of those is my friend and legendary strategist Marko Kolanovic, unique among Wall Street’s outspoken voices for not capitulating on his conviction that the market has climbed to unsustainable heights and is complacent, if not ignorant, to the significant downside risk.

In many ways, he is truly one of the Last Bears Standing. He’s been kind enough to author the foreword to Matador: The Bull Killer

Foreword

By: Marko Kolanovic

When my friends at The Last Bear Standing asked me to write my thoughts about markets, I agreed. I have been intensely thinking about recent market developments and the enormous levels of uncertainty across the world—perhaps last seen during the 2008 Global Financial Crisis (GFC) and the 2020 pandemic.

Let me start by saying that global macro and geopolitical developments are so profound and complex that one should assess developments day by day and remain flexible to adjust their outlook based on new developments.

Before formulating a view, one needs to first understand where we are currently in the historical context and, more importantly, how markets arrived at their present state. I will start with the 2020 COVID crisis—a great recession that never materialized due to a massive wave of fiscal and monetary easing across the world. As such, the 2020 recession never performed its role of purging unviable businesses and market excesses, and hence it never set the stage for healthy and sustainable growth. If anything, we saw the emergence of inherently worthless assets and shaky businesses with sky-high valuations. Economies grew increasingly dependent on government projects and handouts. It should not be surprising that the combination of COVID disruptions and these monetary and fiscal excesses resulted in inflation not seen since the 1970s.

When the Federal Reserve (being late) began hiking rates in 2022, air came out of broad markets and various speculative assets. Inflation breakevens dropped quickly by early 2023, but the Fed continued hiking to 5.5%. The yield curve inverted—a historical signal of recession. I did not believe that markets would thrive in 2024 with these high rates, even though recessions typically start with a significant lag (1–2 years after yield curve inversion). However, in 2024, everybody forgot about the inversion, rising defaults and recession risks, and it became extremely rare and unpopular to have a cautious market view. Despite near-unanimous optimism, low-end consumers still faced pressure from high credit card and other debts, small businesses struggled with financing costs, and real estate markets remained paralyzed. These concerns were set aside as the economy thrived on promises of AI, the stock market’s wealth effect, record government spending, and interest rate cuts that began just before key elections (here). Markets ended 2024 with P/E multiple near the highest level since the dot-com bubble burst, with narrow leadership—concentration levels not seen in over 50 years. However, there were many signs that a major rebalancing in markets is about to happen which I discussed here, and here.

It didn’t take very long for a correction to begin. The bubble in momentum stocks (both large and small) started unraveling in January 2025 (here). If a stock has a multiple of ~400, a 40% drop should not surprise anyone (here). What followed, however, was far more devastating for markets and the real economy: the start of an unwinnable trade war between the U.S. and much of the world (here, here). Students of markets should recall the first trade war in 2018—not only President Trump’s tactics and response functions, and his relationship with Chairman Powell (here), but also take into account the differences between his first and second presidential terms. I warned that markets would experience a larger drawdown, with the so-called put option strike lower than most anticipated (here, here). Since the trade war began, damage has grown daily, and most economists now estimate a ~50% or higher probability of an imminent recession. In early April 2025, markets crashed, and the VIX reached levels seen during the Global Financial Crisis (GFC) and COVID-19 pandemic.

The probability of a recession is indeed very high, and we may already be in one. Notably, the yield curve inversion that started in 2022 typically indicates a recession when the curve finally un-inverts. Well, three years after the first rate hikes, we saw that signal in 2025. Declining asset prices (both stocks and bonds) have eroded the wealth effect, negatively impacting economic activity. The new administration intends to cut, rather than increase, spending. Low-end consumers continue to struggle, and with the prospect of layoffs, their situation is likely to worsen. The trade war is expected to slow the global economy, and the Fed is not yet signaling relief just yet (here). In addition there is a risk of inflated private assets and crypto.

Where does this leave us after the recent bounce and partial market retracement? The S&P 500’s forward multiple is 20x, and earnings are not properly reflecting recession risks. With a multiple of 20x, how much upside remains? In many recent crises—such as the 2002 double bottom, the 2011 European debt crisis, the 2015 emerging markets crisis, the 2018 trade war, and the 2022 rate hikes—the S&P 500’s multiple dropped to ~15, suggesting a ~20% decline from current levels. I have not yet heard a compelling argument why the multiple would not fall to ~15x and remain there for more than a few days during this crisis. Even assuming a quick earnings recovery to $300, a multiple of ~15x would place the S&P 500 below 5,000. Worse scenarios are plausible: if earnings decline (as in a recession) and the multiple drops to ~15x, the S&P 500 could fall to the high 3,000s or low 4,000s. Even more severe outcomes, though less likely, are also possible—during the GFC, earnings plummeted, and the multiple fell to ~12, and stayed there for a few years.

Downside risk for markets remains substantial, with limited upside after the recent rally. Current prices assume an optimistic resolution to the significant economic, political, and geopolitical challenges the world faces. Trade with caution and evaluate new developments as they arise.

Marko Kolanovic, PhD

@markoinny

Traje de Luces: The Setup

There is a moment when the writing is on the wall for anyone that cares to look, but few do. I would describe this well-established dynamic, but there is no way I’d do it better than Colm O’Shea in his Hedge Fund Market Wizards interview:

Anyone that has been paying close enough attention can see the echoes of the last market top and reversal. Today, we face some similar concerns to the ones that unwound the 2021 peak – a compression in multiples from an overheated equity market, cost and inflationary pressures, and trade concerns.

But in 2022, the US economy was still in a strong state of rebound with a massive labor shortage and rapid jobs growth. Inflation was largely driven by excessive demand and the negative drag of monetary policy was balanced by a growing fiscal impulse.

Today, economic growth is slowing after reaching unsustainable levels. Trade is coming to a halt and fiscal outlays are being slashed. The easy narrative of inflation reversion and Fed panic has given way to a multitude of threats. The days for this bull market — built on unsustainable capital allocation, misplaced trust in institutions, and a speculative fervor — are over.

Here is the opinion of this Bear. The market’s current decline is not a correction, a bump in the road, or a dip to be bought.

We have entered a protracted bear market that will result in sub-par equity returns from current levels. If investors push this bear market rally higher, I will be happily selling to them the whole way. This unwind has been in the making for nearly four years, and now it is time for the conclusion..

The Three Pillars of the New Bear Market

The market today faces a unique and dangerous confluence of risks spanning three critical categories. They are:

1. Macroeconomic risk: Growth is sputtering: Real GDP is decelerating, credit stress is growing, consumer sentiment hovers near a 70-year low, and factory surveys flash bright red. These cracks, formed before the latest policy shocks, have been split wide open.

2. Policy risk: Washington has flipped the script on globalization: Tariffs have halted Chinese goods and alienated allies, while shocking ports, supply chains and prices. The fiscal impulse has reversed.

3. Market risk: The foundation is brittle: Valuations assume earnings growth that no sober model supports. Leverage in ETFs and options will serve to shatter the glass house they helped construct. Complacency sets the stage for a slide worse than 2022’s shake-out.

Every bear market tells a story. The narrative often crystallizes around a central theme, a dominant fear that grips markets and drives asset prices lower. We saw it in 2022, as the specter of runaway inflation and an aggressive Federal Reserve ended the post-COVID euphoria. We saw it in 2020, when a novel virus (extremely briefly) brought the global economy to a standstill. Go back further: the 2018 taper tantrum, the 2008 Global Financial Crisis born from subprime excess, the 2001 bursting of the dot-com bubble. Each downturn had its defining characteristic, its unique signature of risk.

Today, as markets navigate treacherous crosscurrents, the question arises: what is the theme of this potential bear market?

It is hard to put a neat box around the innumerable and cross correlated factors of market pricing. But, across the landscape, several undercurrents keep appearing – the three pillars of the new bear market.

First, is the Capital Cycle

The market has benefited from both a public and private capital cycle, driven by new technology and stimulative policy. The effects on estimates and earnings alike has been impossible to miss. But the cycle is turning.

Second, is the Trade War.

Without exaggeration, we are witnessing the most dramatic rejection of the free-trade ethos that has defined the post-Cold War era – an ethos that has greatly benefitted corporate valuations and profits.

Finally, there is the Market Mirage.

Driven by crowded trades and the proliferation of leverage, many stock have completely disconnected from fundamentals. What’s more, an overwhelming sense of investor complacency has emerged out of a virtuous cycle of reassurance from the market’s consistent rebounds. Signal after signal, sometimes dire, are ignored.

Calling a bear market is one thing, the real question is how to trade it?

This is my first comprehensive analysis of asymmetric opportunities as we enter what is sure to be one of the most hated bear markets in history. Complacency may last for longer than is comfortable on the short side, it always has. Take, for example, 2008…

The writing was on the wall as early as 2007. Bear Stearns collapsed in March 2008. Still, by mid-May, SPX was a meager 9% off of its highs.

So, as we prepare for the inevitable conclusion to this saga, here’s my cross-sector, high-level takes on the micro aspect of a very macro slowdown. I explore how each of these risks is likely to manifest in single names and discuss setups I feel will serve well as the market weakens.

I. Macro Risk: Tercio de Varas

Even before tariffs throttled trade and policy chaos cracked the confidence of CEOs, the underlying economy was flashing warning signs. Consumers — especially low and middle-income consumers — have been struggling for years though widely ignored. The employment market has been deteriorating since 2022.

Consumers polled by the University of Michigan expecting worse business conditions in a year is at its highest level since the survey began, which began rising in Q1 2025:

Outside of the lucky few, businesses aren’t faring much better. Borrowing costs have remained elevated for nearly the years, dashing the hopes for a refi for those struggling with debt service coverage, and small businesses are reporting the worst outlook on the economy that they have in years. PMIs have been in contraction for five months, trucking volumes are already down ~20% YoY and the wave of negative earnings guidance revisions is surging higher. Long before the recent collapse in new orders, manufacturing optimism had peaked most surveys:

And while we are nowhere near the default rates that were present even before the GFC, it’s evident that any sort of buoyancy to the consumer’s position provided by COVID-era stimulus has been completely exhausted, as evidenced by high levels of serious delinquencies.

Sometimes, it’s as simple as listening to the companies. Downward EPS guidance revisions have increased to the highest level since the 2022 downturn.

Earnings calls have taken a dour turn. Just over the past week:

“Saving money because of concerns around the economy was the overwhelming reason consumers were reducing the frequency of restaurant visits.” — Chipotle CEO Scott Boatwright

“Relative to where we were three months ago, we probably aren’t feeling as good about the consumer now” — Pepsi CFO Jamie Caulfield

“I don’t care if you call it a recession or not, in this industry, that’s a recession” — Southwest Airlines CEO Robert Jordan

Back in February, Equifax flagged a slowdown in hiring that had not yet translated to a meaningful increase in unemployment claims — but it’s hard to imagine that doesn’t come to fruition after what’s occurred since. The labor market has been cyclically tight, but not secularly in the way some would have you think.

Put simply: the real economy can’t hold up given the massive headwinds it faces. The market may have rebounded, but the consumer is strained and the corporate sector is cautious. The sugar highs of fiscal spending are wearing off. And with cost pressure suddenly rising, the Fed is boxed in and government policy backstops will see none of the bipartisan support nor the speed of the COVID bills (and to some degree the IRA) that saved us from recession in 2022.

The Capital Cycle

Capital investment is at the heart of economic growth but also highly cyclical. This duality explains how massive capital buildouts, from railroads in the 1850s to internet infrastructure in the last 1999s, often coincide with market booms and busts. The underlying technology is often a long-term winner, but the pattern of optimism leading to overhype and overcapacity is repeated again and again.

The nature of accounting also adds to these swings. In the early phases, there is a widespread overstatement of earnings as the spender’s costs are capitalized while the providers immediately recognize these sums in earnings. As the expense shows up in lagging depreciation earnings it depressed earnings for future periods.

This cycle we’ve just lived through was driven by a surge in capital goods, fiscal spending on infrastructure, new residential construction, and a massive buildout AI. The timeline looks something like this.

We’ve already seen the boom…

AI: Shifting Narratives

After leading the market for much of 2023 and 2024, sentiment for the AI capex trade has clearly already peaked. Early signs of wavering were amplified by the DeepSeek scare — the first true moment of disbelief.

The AI buildout, a key driver of both economic and market activity over the past two years, is no longer a clean macro story. Capex remains high, but efficiency demands are rising. As LLM commodification concerns lead to more prudent and exacting capital plans, many of the “AI winners” are being reset. The next leg requires proof of ROI, not just ambition.

Hyperscalers, which hold the handle of the bullwhip, are increasingly cautious on excessive capex, after already committing to some of the massive spending ever undertaken in the tech industry. Now, architectural improvement and commodification of the LLM foundational models — not to mention the sheer scale of current spending — threatens the ROI on investment while a wave of depreciation expense is only beginning to drag on earnings.

To be clear, I’m not arguing that AI capex is going to collapse. Rather, I’m arguing that it’s impossible to maintain the rate of change that has occurred over the past two years. For picks and shovels, the rate of change is critical, as they require increasing investment for ongoing growth.

Moreso, market perception and valuation is likely to overshoot fundamentals on both the upside and downside. If there is a moderation, pause or decline in hyperscaler capex, the bullwhip will sting through a long value chain of suppliers and beneficiaries, from chips, to hardware, to integrators, and datacenter hosts.

All of these companies, and their suppliers, owners, and employees, have a significant exposure to the direction of spending.

Betting against the capex cycle is not a bet against AI. It is a bet based on the psychology of the hype cycle, a moderation of spending growth rates, and an evolution of more efficient AI development across both training and inference combined with the ongoing rapid deployment of incremental GPU capacity.

Here are the names most at risk.

First up, after collaborating on the Haves vs. Have Nots, I’ve asked CitriniResearch to weigh in on the risks to the King Kong of the AI trade: Nvidia.

NVIDIA Corporation (NVDA)

Collaborator: CitriniResearch

The semiconductor complex, led by Nvidia, has driven markets for much of the 2020s, but cracks are showing. Nvidia still walks on water at 20x next-year sales and a 75% gross margin, yet markets rarely let one firm own a gold rush forever. Five cracks already spider across the surface: customers rolling their own chips, insurgents routing around Mellanox, software abstractions flattening CUDA, efficiency breakthroughs bending the compute curve, and a product road map that keeps skipping beats. When perfection is priced in, even a hairline fracture can wobble the story.

First: the hardware. After years of writing blank checks to Nvidia, hyperscalers now parade TPUs, Trainiums, and next-gen ASICs that hit “good-enough” throughput at cost. Cerebras welds an entire 300mm wafer into one die and Groq’s deterministic TPUs run 1,400 tokens per second. Both dodge the bandwidth chokepoint that props up Nvidia’s ASP. If a data-center manager can clear the same inference load with fewer boxes, the H-series premium looks fragile.

Second: the efficiency gains. While the parade of Jevon’s Paradox explainers has subsided, it still seems as though there is a disconnect between market understanding and the reality of exponential efficiency improvements. My first interaction with The Last Bear Standing was after he called the top in the AI Power Trade, and those insights (that power efficiency will likely result in power demand not expanding due to AI at the pace the market expects) apply across the technology - not just in the power theme.

Third: the cycle. A refrain I’ve found myself repeating all too often over the past 5 years or so: Just because AI or <insert exciting new technology> is a real thing does not mean semiconductors aren’t cyclical anymore. Nvidia just ate a 5.5 billion-dollar writedown on Hopper-derived H20 parts that never found a home. Blackwell, a stop-gap tuned for jumbo context windows, arrives as customers digest excess gear. Management is already teasing “Rubin,” a lower-precision workhorse aimed at robotics. Three overlapping architectures, swelling inventories, and a softening order book rhyme with the margin punch-out that clipped the stock in 2022. It’s starting to feel like we’ve seen this movie before.

Put this all together and the asymmetry tilts bearish, although NVDA is certainly an excellent operator and a company I’d want to own, the fact is I can’t underwrite these risks currently with the potential for such minimal near-term reward. A modest dip in compute demand or a modest nick to pricing cascades straight through a high-operating-leverage model. Cracks in the armor of NVDA’s CUDA moat or a significant series of breakthroughs on ASICs or TPUs from Amazon or Google. More open-sourced efficiency improvements a la DeepSeek.

Now, NVDA is not just training, and inference demand accelerating can be a big tailwind (although AMD may be better positioned and have more favorable base effects for that occurrence). Rubin could ship cleanly while buybacks plug the valuation gap.

Still, when every pixel of the canvas is priced for brilliance, “good” can trade like “bad” and “not perfect” can still pay.

Oracle Corporation (ORCL)

As early mover Microsoft takes a step back, Oracle is beginning to fill its shoes taking on much of pre-training capex through its Stargate partnership with OpenAI and Softbank. But unlike Microsoft, whose massive growth in profitability over the past several years has provided plenty of organic funding for capex, Oracle is going “all-in”.

Even as Oracle’s current capex spending is smaller than the other hyperscalers, it is much larger relative to its cash flow generation. In the most recent quarter, its $5.8 billion capex spend was equal to its entire cash from operations – reinvesting the entire business’ cash flow into the buildout.

In short, it’s leveraging the company on this bet. The spending will certainly provide incremental growth uplift, but if return on investment falters, forward earnings will be blanketed with the long and lagging drag of depreciation for years with little to show for it. If Microsoft’s hesitance proves wise, there is significant downside risk in the stock.

Historically speaking, ORCL hasn’t needed to rely on aggressive capex, so the current bevy of moves smacks of desperation. I see this aggressive business pivot towards growth-at-big-cost to be a misguided departure from the mature tech-giant playbook which generally includes inorganic growth at reasonable hurdle rates and a lot of buybacks.

Broadcom Inc. (AVGO)

As a major component supplier to datacenters, Broadcom has ridden the AI capex wave, but unlike NVDA which has seen its forward valuation metrics rapidly compress to cycle lows, AVGO’s forward metrics expanded significantly since 2023. On a NTM basis, AVGO trades at double its historical multiple and a 20% premium to NVDA.

While much of the company’s recent growth has been driven by US datacenter builds, AVGO’s business is much broader, with exposure across a range of applications and end uses. The company also has significant revenue exposure to China (and Singapore as a proxy), which could present challenges given the geopolitical landscape. Meanwhile AVGO faces competitive pressures within the AI market as recent news shows Google shifting towards MediaTek to help design its Tensor Processing Units (beginning with next generation designs in 2026).

Long-time CEO/Architect Hock Tan and AVGO management are well aware of their saturated position in the communications chip and embedded interconnect markets, so much so that they have made it a point to use cash generated by the chip business to expand into other verticals. The company went on an acquisition spree between 2018 and 2021, spending more than $110 billion on three software companies, with the bulk of that being on virtualization giant VMWare.

While virtualization is clearly a winning theme in the distributed, heterogeneous computing world, Broadcom is running the subsidiary as if it were a private equity firm, raising unit prices by 1,200% in some instances and increasing minimum core purchases from 12 to 72.

A slowdown in AI spending or sentiment could lead to significant multiple contraction, while its broader business portfolio demonstrates significantly slower growth with global macroeconomic sensitivity.

AI Coattail Riders

Many names adjacent to AI infrastructure have ridden the trend as investors scrambled for any hint of companies utilizing AI in a manner that demonstrably increased revenue. This frenzy to own “AI implementers” has handily overlooked the fact that AI itself poses an existential risk to these businesses. Operating in areas that can quickly be commoditized (thanks to current struggle to generate ROI in AI outside supplying the data centers), these companies are pitched as the “scaffolding”. Perhaps investors have forgotten that the scaffolding is dismantled and hauled away once the building is constructed…

Innodata Inc (INOD): Assisting in One’s Own Demise

Innodata’s core business is paid data grunt work: armies of offshore staff label images, transcripts, medical notes, and code snippets so that clients can train or fine-tune large models. The company wraps this labor in basic workflow software, then markets the bundle as “AI data engineering.” No wonder investors have flocked to the stock - after all, inno- + -data. What a winning combo for the AI era.

Innodata’s revenue spiked to about $59 million last quarter, EPS turned positive at $0.34, and operating margin hit 19% — nice optics, to be sure. But investors appear to believe the company is a picks-and-shovels royalty on generative AI demand rather than a people-per-hour, overearning contract services shop that is easily disintermediated by the very technology that has caused its earnings to spike.

INOD trades at more than 6 times sales, roughly in-line with Salesforce (CRM).

That premium ignores how easily the service can be replaced even if AI demand explodes. Hyperscalers and foundation-model labs are moving labeling in-house, using active-learning loops that cut human clicks by 70 percent. Open-source tools like Label Studio and commercial platforms such as Scale AI make it trivial for customers to spin up their own pipelines. On top of that, synthetic data generation and self-supervised techniques keep reducing the need for meticulously hand-tagged datasets.

Eighty percent of run-rate revenue comes from one hyperscaler, and the rest is “lumpy” by management’s own admission. Additionally, the model depends on thousands of offshore annotators, so wage inflation flows straight to cost of goods. Scale AI, Appen, and in-house labeling tools from the same hyperscalers are undercutting bids every quarter, crimping price and capping margin.

If that flagship customer trims volumes by even 20%, the promised 40% growth in 2025 vanishes, EBITDA could halve, and a services multiple of three times sales puts the stock in the mid-teens. Unless Innodata lands new whales fast or proves it owns defensible tooling rather than interchangeable labor, investors are paying a SaaS valuation for a business that can be automated out of existence.

Another example of this ironic setup is…

Intapp (INTA): Commoditized Offering, Differentiated Valuation

Take, for example, IntApp (INTA), which has risen 500% since 2022 lows despite just a ~50% increase in revenue.

Intapp sells cloud workflow tools to law firms, private-equity shops, and consulting partnerships. Management touts “AI-assisted knowledge management,” yet the letters AI appear far more often in the marketing deck than in the income statement. Even so the stock changes hands at about 9x sales and roughly 44x free cash flow—a richer multiple than ServiceNow commanded at the same size. Net cash is thin, and the company still reports GAAP losses. The bull case says legal and deal advisers cannot afford not to “AI-up,” but legal startup Harvey will likely surpass INTA’s SaaS ARR this year and make it clear that they hardly have an edge.

Don’t get me wrong, these are both businesses that I’d love to have operated for the past couple years. They’ve benefitted hugely from the demand for AI. But, as an investor, the terminal value proposition is dismal. The C-Suite will walk away having paid themselves handsomely for being in the right place at the right time, and investors will get shafted as the companies are relegated to the stock market equivalent of that cardboard box in your garage with all the power adapters for old cable boxes and first generation iPads. We can see that playing out in the insider transactions for INOD already:

Healthcare AI Hope Trades

Both GE Healthcare (GEHC) and Veeva (VEEV) have been bandied about as “AI winners” in the healthcare industry. While the outlook for AI deployments remains dynamic and could certainly accrue to the benefit of these companies, there is nothing material in the near-term, and in the meantime, both have seriously challenged core businesses.

In the case of GEHC, the company supplies capital equipment such as imaging machines and patient monitoring systems to hospitals and other healthcare facilities. A double-digit percentage of their earnings comes from China, putting them squarely in the middle of the trade war, and the bulk of their US business comes from hospitals that face a significant risk of funding cuts under the Trump administration.

Meanwhile, VEEV, a software company, sells technology to drug developers to help manage trials. While drug development activity is picking back up after a long lull, the timing of this rebound is in question, and in the meantime, VEEV faces significant competitive pressure from larger players like Salesforce (CRM) as well as from internally developed solutions from their customers.

Another Healthcare AI hopeful, Doximity (DOCS), is trading near an all-time high selling technology solutions that help pharmaceutical companies market their products to doctors. Not only is this a challenging business with significant potential competition, but pharmaceutical advertising is in the crosshairs. The Trump administration has spoken directly about limiting how the industry markets, but more importantly, a slew of negative potential policy actions, including tariffs targeting the industry and tax reform, will impact bottom-lines in the pharmaceutical industry and, as a result, advertising budgets. This puts DOCS revenue at significant risk.

Below the Paywall: Policy and Market risk to define 2025 and 40 more single stock discussions.

Read more

The Trump administration is losing face and losing leverage. But for those hoping for a trade-war backpedal — don’t hold your breath.

The April 2nd Liberation Day announcement was an intentionally choreographed shock-and-awe affair that worked much better than many admit, at least at first. Swinging aggressively and recklessly, Trump caught everyone — trading partners and market watchers — on the backfoot, commanding global attention. This bad-cop routine also gave room for a good-cop retreat.

A 90-day pause announced on the eve of the “reciprocal” tariffs put a gleeful spin on what would otherwise be considered a paradigm-busting trade policy — a 10% tariff on most of the world, with limited exceptions. The administration has already implemented the highest and most wide-sweeping tariffs in a century, and these levies have been achieved without “retribution” by most countries, at least for now.

But such an aggressive first move has left its flanks vulnerable. While US vs. Everyone made for a dramatic opening sequence, the administration must now act out the Sisyphean script. After tossing the bird to every country in the world, can they sit across the table with these same folks and negotiate a hundred bilateral trade deals in just three months? Today is April 18th; Nine of the 90 days have elapsed with not one deal approaching completion.

Beyond simple bandwidth, there is the question of collective bargaining. Even if these negotiations are “bilateral” they don’t exist in a vacuum. In a one-vs-one matchup the US might hold the cards, but fighting every counterparty at the same time in such a public forum is an impossibly tall order, particularly with the China faceoff turning into a bare-knuckle brawl.

Perhaps most importantly, Trump hasn’t held the steely composure he donned in the Rose Garden. The stock market tantrum and volatility in Treasuries clearly changed the administration’s “take the medicine” tune. Meanwhile, the “Apple Exemptions” announced last weekend revealed inconsistency and pressure points. The soundbites coming from the White House are increasingly sounding clingy and desperate — Xi will call! Great progress with Japan!

The mojo has run dry. While Trump’s most steadfast followers weave narratives about strategy, most observers recognize dysfunction. If the plan was to encircle China, why alienate everyone else first? The shoot-first paradigm has created a situation where it’s not even clear that US negotiators have a specific objective, let alone a means to achieve it.

But the loss of the leverage does not mean the end of the trade-war. As time ticks with little progress on new deals (and a one-way tariff in place), the likelihood of retribution increases. Meanwhile, as Trump’s most aggressive ambitions falter, he will cling harder to the ground he has already gained. Even if there is some room for de-escalation with China, it seems more and more likely that the broad global situation is heading for a stalemate — a prolonged battle in which pain is felt on both sides while global trade suffers. Rather than a blitzkrieg with a swift and decisive outcome, the trade war is entering the trenches.

Read more

For the past week, the White House has rag-dolled the global financial market with its ever-changing whims. The level of volatility is historic, sharing company with the worst market crises of the past century.

And yet this episode is in a category of its own. Whether it’s freewheeling chaos, elaborate negotiation tactics, or a stiff cocktail of the two, there is a level of reality TV drama that suits the parties involved. And for all the round-tripping of this week, the S&P 500 is up a relatively mundane 3.3% on the week, through Thursday’s close.

Considering the titanic trade shifts in question, some bigly swings are justified. But in some way, the drama seems to be crowding out more acute analysis.

The market is stuck in an overly simplistic binary, where “more tariffs = bad”, “less tariffs = good”. As equity correlation has surged to multi-year highs, the vast majority of the market has traded in tandem with the changing tune of the administration. As a knee jerk reaction, this impulse is correct — if a global trade war triggers a recession, it’s bad news for everyone.

But investors should quickly sharpen their pencils. There has been relatively little distinction or discrimination in how these current tariffs will impact individual companies. In some cases, perhaps too much pain is being priced; in others, not enough.

Wednesdays’ historic turnaround is a perfect example. At 1:19pm the White House announced a 90-day pause on reciprocal tariffs for every country except China, while upping the China tariff rate to 145%1. The result was one the largest one-day rallies in market history — a near universal green screen with advancing stocks outnumbering declining stocks at a rate of ten-to-one.

But other than the massively oversold positioning and desperate desire for relief, did that market action make any sense?

Sure, the perception of flexibility and de-escalation was taken as a positive. But as Bloomberg’s Anna Wong calculated later in the evening, the impact of increased China tariffs cancelled out the impact of lower reciprocal tariffs elsewhere, leaving the total tariff stack more or less unchanged on an aggregate basis.

But for any single company reliant on foreign goods, the news was either very good or very bad depending on its relative exposure to China.

As a high-profile example, why did Apple — which sources a large majority of its hardware from China — rally 15% on the day that its tariff rate increased dramatically2? Elsewhere, why are certain dollar stores — which source nearly all of their merchandise from China — still providing a “recession” safe haven?

Moving forward, more critical judgement is set to drive relative performance. Here are twelve stocks that stand on the firing line.

Read more

For months, we’ve been circling the drain — the median stock peaked back in December, DOGE took office in January, Tariff-talk arrived in February, soft data started screaming in March. Now, it’s April and we’ve passed the point of no return. We’ve entered The Vortex.

Markets are only beginning to price this reality.

I’ve written about the turning cycle — untethered euphoria, the turn of sentiment, the moment of AI disbelief, and “max stupid” ringing the bell. So far, we’ve played it tactically, focusing on the lowest hanging fruit by top-ticking nuclear nonsense, and shorting pre-revenue memes.

Fading momentum alone was enough for a market correction, and the S&P 500 did indeed notch a 10% drawdown by early March. Unsurprisingly, dip buyers held the index in limbo for the past three weeks while traders did their best to explain away the flashing warning signs while calling political bluffs.

But this week’s events change the landscape dramatically. Not only is hard data beginning to confirm a material consumer slowdown, but the hopes of Trade War skeptics were violently squashed in the White House Rose Garden.

What lies ahead is much more daunting than a mean-reversion. Trump’s tariffs represent the most extreme and sudden reversal of US trade policy maybe in the nation’s history. Do not underestimate the significance.

It’s not a negotiation tactic to reduce trade barriers, it’s a retreat from free trade for the purpose of national security and realignment. It’s not a policy to juice economic metrics, it’s one that will bring immediate pain. It’s not a stick or a carrot, its a wrecking ball, intended to dismantle the status quo — a status quo that has supported market gains for decades.

The implications are profound and come at a precarious moment. For the past several years, the US economy has been powered by stimulus, fiscal spending, and the wealth effect, leading to complacency and vulnerability. Now, we are seeing slashed spending, government firings, and tax hikes in the form of aggressive tariffs, while the single largest store of paper value crumbles. At the same time, consumer spending is fading and real GDP may already be in contraction.

These factors are reinforcing and accelerating. The Trade War has only just begun — retaliatory measures have yet to be announced and few parts of the global economy will remain untouched. Supply chain disruption, policy-based inflation, and business uncertainty will put even greater downward pressure on economic activity. Against the backdrop of rising import taxes and a weak consumer, corporate profits will come under pressure. Declining earnings will hasten the deflation of market multiples as any last bit of optimism dies.

What we now face is different from the last several market downturns.

In 2020, an exogenous shock was counteracted by extreme monetary and fiscal support, while global actors largely put aside their differences for the greater good. In 2022, demand-driven inflation forced rate hikes, but the underlying US economy was still in a strong rebound, and the war in Europe had minimal impact stateside.

Today, we contemplate something different; a market crash, rising global tensions, supply-side inflation, and even a severe global recession.

Read more

EXECUTIVE SUMMARY

Recession, or “The R-Word” as the financial media likes to call it, has become more of a threat than at any point since the Fed began hiking rates in 2022. For the first two years of that cycle, Fiscal Policy and Monetary Policy engaged in a sort of beautiful dance whereby the efficacy of interest rates in slowing the economy was blunted just enough by the supportive nature of Fiscal Policy.

Now, however, the fiscal impulse has turned decidedly less positive and the chances of a recession have risen. Rather than bandy about meaningless probabilities (I recently saw one forecast that said the chance of a recession in the U.S. this year is 50%, which is potentially the least useful thing I’ve ever read), I think it’s more interesting to explore something a bit more nuanced.

Recessions are not all identical. They each have their own unique characteristics and drivers. If the current uncertainty snowballs into economic weakness, what becomes the defining characteristics of that slowdown?

Despite appearances, the American economy is experiencing an unprecedented bifurcation. It’s a tale of two economies operating in parallel but divergent realities. A vast majority of wealth has accrued to the top 10%, and their exposure to the market makes declines a significant threat. A fundamental restructuring of consumer spending power following the pandemic has created both systemic vulnerabilities and unique market dynamics based on how this could play out.

A Wealth-Effect Driven Recession?

Citrini Research’s latest Macro Memo (After the Trump Bump) introduced the concept that the divide between rich and poor, in terms of both asset ownership & spending, had gotten so stark that the primary threat to the economy has become a decline in asset prices & a wealth-effect driven recession. With the top 10% representing half of all consumer spending (a statistic that is likely lower than reality considering the mechanisms of wealth transfer), this “Macroeconomic Ouroboros” seemed worth exploring further.

As it stands today, the wealth of the U.S. has never been so dependent on the stock market. Nor has financial wealth been so highly concentrated among the top economic percentiles and overall spending so dependent on those few. Therefore, it stands to reason that the economic prognosis has also never been so contingent on the movements in asset prices.

We often assume that stock market performance reflects the underlying economy – some blend of jobs, consumption, investment. Of course, there’s feedback – rising asset prices can boost confidence and consumption, while falling prices have the opposite effect. But this reflexivity is not constant.

Valuation, asymmetry of outcomes, and the relative dependence on asset prices for ongoing economic activity must play an important role. Sometimes the economy drives the stock market, sometimes the stock market drives the economy. We believe we’re firmly in the latter dynamic.

A Government Policy Driven Recession?

Four years ago, the Last Bear Standing shined a light on a $50 trillion asset class that put one of the world’s largest economies on the brink. The Chinese property market – the economic growth engine of the country for two decades – was entering a policy driven collapse. As the primary store of wealth for much of the country, the property bubble’s deflation would have severe knock-on effects for consumer sentiment and consumption. Four years later, the Chinese economy still stands in the shadow of this property downturn, even as the re-routed investment is already bearing clear fruit.

Today, a different $60 trillion asset class is at risk. For the past fifteen years, the U.S. stock market has been the greatest driver of “wealth” in the world’s largest economy (indeed, in the entire world). Rising asset prices have boosted consumption habits from the top earners, even as the uneven distribution has driven social discontent and populism. And despite the many lessons of history to the contrary, there is now an engrained assumption of policy-based protection.

While it’s almost certain that an economic recession would coincide with falling asset prices, it also seems that falling asset prices could cause a recession. Perhaps it is even a prerequisite. Like China, an abrupt pivot in policy stands central to this question.

What if the Trump administration is serious about “Main Street vs. Wall Street”? With so much of the wealth and spending concentrated among the top 10%, it seems likely that any change in their spending habits could be enough to trigger an economic contraction.

From a political perspective, we have a defiant and emboldened President with populist-tendencies and a consolidated power structure who is intent on driving a new economic vision focused on protectionism, closed borders, and the retreat from free trade.

Undoubtedly, tariffs have a greater impact on low and middle-income consumers than they do on high-income consumers.

Even while writing this article, we were hit with this headline right after a terrible Friday for stocks:

Economists can debate the net effects of Trump’s proposed policies (and the pressures of his billionaire backers), but when he blames a falling S&P 500 on “the globalists” we shouldn’t ignore the class-based tenor.

And maybe he’s right – is there a greater emblem of the neo-liberal free-trade world order than the S&P 500?

So far, it’s been a decent idea to listen to the administration when they say they’re focused on something. Already, we’ve seen the total suffocation of illegal migration, an enormous supply of cheap (and legally unprotected) labor for the prior four years. Tariff policy waffles daily, but as we laid out in our piece “Tariff-ied”, don’t expect Trump to retreat before a significant impact. Even in this uncertainty, trade tensions have grown and business decisions remain paralyzed.

On the other side..no taxes on tips? There are very few financial analysts, attorneys and accountants making tip income. Meanwhile, Trump’s Musk-shaped federal hatchet has already begun to hack at government employment and white-collar contractors.

An AI Powered Recession?

At the same time, a large-language-model shaped Sword of Damocles now hangs above the white collar employee.

If you’re a company that primarily employs these kinds of well-educated knowledge workers that has to go from simply pausing hiring due to economic uncertainty to actually cutting your workforce due to economic hardship, well… AI might not be great right now, but it just might be good enough.

What does “good enough” mean? AI has progressed just far enough–to the point where companies, out of necessity, will be willing to underwrite the risk that it can make mistakes or be sub-par at certain tasks and actively seek to implement AI solutions that can result in permanent job-losses for white collar professionals. AI systems are advancing most rapidly in domains that have traditionally commanded premium salaries - legal analysis, financial planning, executive decision-making, and specialized consulting.

Some illuminating statistics from The White-Collar Recession 2025:

• The American Bar Association noted recently that the nation’s largest law firms have cut entry-level hiring by nearly 25% compared to previous years.

• Ad agencies are cutting creative and copywriting positions as generative AI produces marketing materials faster and cheaper. Notably, even Grammarly—a company founded on augmenting human writing—laid off 20% of its staff, explicitly citing AI’s growing capability to autonomously perform editing and writing tasks.

• Vanguard reports that hiring for positions paying over $96,000 annually has reached a decade-low level.

On the other hand, we’re a long way away from people who work with their hands getting replaced by AI. Jobs requiring physical presence, situational adaptability, and human interaction remain less vulnerable to AI disruption. From healthcare support roles to maintenance services, these middle and low-wage positions often involve skills that remain challenging for AI to replicate. We might encounter a temporary inversion of the traditional economic hierarchy as it relates to income & recession resilience in this first era of the “age of AI”.

High-Income Dependency

As we mentioned earlier, top earners drive both an outsized portion of total spending and discretionary spending. They are also highly disproportionately invested in financial assets.

The top 10% of earners now account for approximately 50% of all consumer spending in the United States. Meanwhile, the bottom 90% have had their savings systematically depleted due to the impact of inflation. This concentration of spending power creates a precarious economic dependency on the financial health and sentiment of high-income households.

Low and middle-income consumers have effectively already been in a recession for some time, which is already reflected in the market.

Depressed consumer sentiment readings and a rejection of the incumbent administration reveal this reality even as headline numbers flash green. Inflation has had a significantly asymmetric impact on their financial well-being. Their spending habits have already adjusted. If we were to have a recession, the largest delta in spending behavior would likely come from high-earners.

Even in a market crash, top earners will continue to outspend the rest. But their relative spending decline could be significant — base-effects matter in year-over-year corporate profit growth. It seems like we’re approaching a tipping point where either spending patterns rebalance or social tensions around income inequality intensify (think French Revolution).

The culture of institutional Wall Street — dominated by top 10% income earners — creates blind spots, and therefore market opportunities. Given how lopsided financial asset ownership is, market participants are prone to become disconnected from the economic reality of most Americans.

A Divergent Recession…And How to Play It

All of this leads to an uncomfortable but (we think) well reasoned conclusion.

A sustained market drawdown could trigger a negative feedback loop that transforms financial market stress into real economic contraction through reduced high-income consumer spending. This transmission mechanism operates through direct wealth effect, psychological contagion, income effect, and housing market channels and, in terms of rate of change, will disproportionately exercise its impact on the economy through the spending of high income/net worth individuals vs low and middle.

A mean-reversion in spending patterns by top 20% vs bottom 80% would catch many companies on the wrong foot. After the spending from “aspirational consumers” (LVMH’s word for poor people with a stimulus check) was exhausted, many companies have focused primarily on their higher-end segments in a phenomenon dubbed “premiumization”. This targeting of the highest earners has worked well. The market has rewarded premium-focused companies with handsome multiples, while companies geared to lower segments are often pricing the recessionary conditions their clientele have already been facing for years.

But this article is not another exploration of how necessity spending is more resilient than discretionary spending in a recession.

We are exploring something else — what would the dispersion of stock-specific returns look like if a wealth-effect driven recession were to occur that had a significantly more pronounced impact on the “haves” compared to the “have nots”.

We’re not proclaiming that things would get better for the poor in a recession. However… the rate of change on the spending of the well-to-do would make the changes from the poor pale in comparison. The lower middle class is already in a recession while the upper class has been doing great.

While there’s some defensive and value aspects to this strategy, it’s unique relative to the classic recession trades. This isn’t just long staples/short discretionary — we believe any economic turbulence that caps off this cycle deserves more in-depth work than that.

1. Long Basket: companies either relatively insulated from wealth effect pressures (focused on value positioning, trading-down beneficiary potential, and necessity-based offerings) with asymmetric exposure to the working class and/or the potential to recover from years of underperformance due to their primarily lower-income focus

2. Short Basket: companies with high sensitivity to wealth effect dynamics, characterized by premium pricing strategies, limited trading-down defense, and disproportionately impacted by a stock market decline

The benefit of this approach is that it could navigate either a unique recessionary dynamic or a recovery in low-income spending and a narrowing in income inequality. In other words, it doesn’t require a bone fide recession for this basket to work. Our "Haves vs. Have Nots" investment thesis represents an asymmetric opportunity to capitalize on the economic divergence between high-income and low/middle-income consumers.

(Bonus Point: if you’ve just read all of the above and your reaction is that this is total nonsense, that there is no way the poor are less negatively impacted than the rich by Trump, that any potential economic weakness will be destructive for low and middle-income while being neutral or even positive for higher-income (as it has done many times before), then you can “choose your own adventure” and invert the longs vs. the shorts!)

This project is a cross-publication collaboration between The Last Bear Standing & Citrini Research. You can follow this link to subscribe to CitriniResearch.

Read more

From buying dips to selling rips, the breakdown in U.S. equities continues. After two red weeks, a position that has taught traders to reflexively anticipate a rally, stocks have failed to find footing. Bears have awoken, and they are hungry.

There has been plenty of negative news to devour.

Trump’s tariffs announced on Sunday soured the mood from the start, and the subsequent reversals have failed to rebuild confidence. Beyond the direct effect of trade policy, the daily flip-flops look less like 3D chess and more like a flailing ad-lib. On Wednesday, ADP reported the weakest payroll gains in seven months, missing estimates by a mile. The Challenger report recorded the largest monthly in job cuts since COVID as DOGE clearcuts the Federal workforce. The Atlanta Fed’s1Q25 GDP nowcasts stands at negative 2.4%, even if driven by import oddities like transatlantic gold shipments and tariff front-running.

The turnaround in sentiment has been stark. Just months ago, Trump’s election sent the market and meme-stocks soaring. Now the growing consensus, even parroted by his advisors, is that the President’s intent is to break the market’s back.

Last week, I argued that this dip is different, and from a technical standpoint it is. While the drawdown alone is not extraordinary, it comes on the back of several stalled rallies and now threatens to break the trendlines that haven’t been tested in years. Chartists watch anxiously as the S&P 500 closed yesterday on its 200-day moving average, threatening to break the long-term momentum indicator for the first time in eighteen months.

Yesterday, the Nasdaq 100 broke this line in the sand for the first time in two years.

Regardless of whether these levels provide support in the near-term, something bigger is afoot.

We’ve looked at this market with growing incredulity for months, and fading the more outlandish narratives has proven wise. The Nuclear Narrative top-ticked the power trade with Constellation (CEG) and Vistra (VST) in a 40% meltdown since the publication in late January.

Meanwhile, the near universal collapse of the meme-cos continues. From a wide menu of candidates, the Four Shorts highlighted two weeks have cratered 25% - 40%.

But this growing negative outlook isn’t based on short-term price action, recent headlines or the volatility in the absurdities. Rather, qualitative and quantitative factors visible for months suggest a new regime is forming.

The volatility unwind that began last July was a notable turning point and early warning sign. Up until that point, the rally was supported by consistently declining implied volatility, bringing liquidity to the market and indicating investor’s growing comfort with insuring risk. But the subsequent bounce in equities has come with rising volatility, a shaky foundation.

With multiple expansion driving the lion’s share of appreciation, and valuations stretching to extremes, the market has become as vulnerable to swings in sentiment even absent an underlying downturn. And with surging call options providing much of the juice for gains, a mere retreat in enthusiasm can trigger a correction.

As more supportable theses like the AI buildout began to waver and were supplanted by speculation throughout the winter, it seemed the hype cycle was reaching a predictable peak. Deepseek’s Moment of Disbelief showed that investors were willing to throw in the towel on its most core convictions of the past two years. The Max Stupid blow-off top, with its some striking similarities to last cycle’s apex, provides further confidence.

Now, the more recent awakening of geopolitical realignment, trade and tariffs, and policy upheaval provide more concrete fuel to drive the downturn.

Navigating a New Regime

Read more

On the surface, this latest market swoon is nothing special. Since briefly kissing new all-time highs last Wednesday, the S&P 500 has fallen 5%. But the benchmark still trades within the range of recent chop and remains flat on the year. The defensive Dow Jones has held firmer, while the Nasdaq 100 and Russell 2000 have underperformed.

But under the surface, there has been a neck snapping whiplash in sentiment. Just two weeks ago, momentum junkies chased 10% daily jumps in names like Palantir (PLTR), Robinhood (HOOD), Hims & Hers (HIMS), and Applovin (APP)— stocks that are now yacking up their blow-off tops in the past several sessions.

Meanwhile, Bitcoin is swimming with cement boots, working on completing a mirror image reversal of its post-election breakout.

But this bloodbath didn’t come out of nowhere — there have been warning signs if you’ve been paying attention. Discussed last week in Four Shorts, signs of a momentum breakdown in speculative shares finally set up actionable opportunities on the downside.

Since publishing the short recommendations last Friday, each of these pure-blood meme-cos have fallen between 21% - 30% as of Thursday’s close. (RKLB and ACHR are each down another ~15% in today’s pre-market trading).

Even the AI infrastructure trade — the bulwark of this bull market — seems exhausted. NVIDIA’s earnings beat on Wednesday evening was rewarded with a 8.5% plunge on Thursday, as the stock trades back to its DeepSeek lows. When good news is punished, it’s a telltale sign that expectations have run ahead of reality1.

Still wounded from China’s AI breakthrough, swirling reports of Microsoft’s cancelled datacenter capacity and some sober reflection on AI demand from CEO Satya Nadella in media appearances last week have helped shift the narrative from endless enthusiasm to the overbuild2. As I argued in A Moment of Disbelief, souring sentiment in AI, rather than actual the underlying stories, may be a more important signal for where this market is heading — for what it’s willing to believe.

But haven’t we seen these drawdowns before? Momentum was mushed last April, and the summer unwind in late July brought a far more extreme spike in volatility. Am I cherry-picking a couple of bad apples?

At moments like these, I think it’s best to be as direct as possible. This dip is different.

Read more

For the past several months, this column has shone a spotlight on growing speculation in the market. We might squabble about index valuation, the economic cycle, or macro trends, but there is little doubt that some micro-bubbles have formed beneath the surface. Specifically, we’ve been tracking “new-era” sectors with several traits.

• Negligible revenues, major losses

• Grand narratives based on unproven technology

• Massive price appreciation over the past four months

These stocks have traded higher in-step not because of any fundamental connection, but a stronger unifying force. They each surf on the same wave of speculation — momentum-based trend chasing fueled with options leverage. It’s a wave that we saw building in the early fall and cresting in the post-election euphoria. Some of these companies may have bright long-term futures, but their current trading is more related to Fartcoin than the S&P 500.

I’ve documented the inflation of these bubbles and provided level-headed analyses of the underlying industries, but stopped short of recommending readers fade the hype. Standing in front of a crazy train is a bad idea, so long as it remains on the rails.

But it seems the tide is finally turning. Over the past several weeks, these shares have shown classic churning and hesitation — a rocket that has exhausted its fuel, suspended for a moment as thrust gives way to gravity. Now, the charts are breaking down, and a downside reversal is the most likely next step.

In other words, these are now actionable short opportunities.

Read more

The reality is less impressive. Rather than shrinking the base money supply, the Fed has simply mopped up trillions in excess liquidity, already sitting inert in its own accounts.

Since QT commenced in June 2022, usage of the Fed’s overnight Reverse Repo Facility (RRP) has decreased by $2.2 trillion — fully offsetting the balance sheet reduction. Over the same time, the monetary base, a more meaningful measure of dollars in circulation has actually increased. Reserve balances at commercial banks have increased marginally as well.

To date, the Fed’s QT program has merely moved money from the right pocket to the left. The decline in its aggregate balance sheet is an illusion — the elimination of double counting more than a true shrinking of money supply.

But the overflow tank is running on fumes. With just $68 billion remaining, the RRP now stands at its lowest level since early 2021. As it trends towards zero, so goes a vital liquidity backstop for financial markets.

Unfortunately, commentary on this topic tends to get reduced to extreme proclamations — imminent disaster or a total non-factor. Let’s cut through that noise.

The exhaustion of the RRP does not spell disaster but it should not be ignored. Moving forward, QT will actually begin to tighten — reducing banking liquidity, increasing banking leverage, and increasing pressure in short term funding markets. Indeed, we have already seen increased volatility in short-term interest rates over the past several months as the RRP has dwindled.

Even still, the Fed’s plan to reduce bank reserves to the lowest comfortable level of reserves (LCLOR) suggests that QT is not done yet. Instead, QT may actually begin to bite.

Read more

Homebuilders at inflection.

Since 2020, the housing market has been one of the toughest to navigate. Soaring prices in the stay-at-home economy were quickly tested by rate hikes. But despite calls for a housing bust, the pullback in 2022 proved temporary and demand proved sturdy.

Amid a dearth of existing home supply in the past three years, new construction has filled the void between homebuyers and a resale market frozen by the mortgage “lock-in” effect. For homebuilders, this backdrop has led to major profits and market gains.

But the sentiment is shifting. Over the past two months, homebuilding stocks have puked up much of the prior year’s gains. New completed housing inventory is piling up as new home prices continue a wayward drift, leaving us to question whether we’ve seen this movie before.

Rather than speculate, let’s take a hard look at the nation’s biggest homebuilders and find out. First, we take a look at the industry in aggregate, then sift the wheat from the chaff, picking six winners and four losers. The findings were surprising.

Read more

The official story of NVIDIA’s flush goes like this.

A group of Chinese researchers released two new large language models — Deepseek V3 and R1 — that mirrored the success of the American AI elite. But they did it at a fraction of the cost and compute, touting a headline of just $6 million (in pre-training cost). That a Chinese firm put out its technology with open-sourced transparency was perhaps an extra kick in the shorts for “OpenAI”, which has long abandoned its lofty altruistic claims.

Of course the news came with skepticism and speculation — Deepseek likely does have access to quite a few forbidden NVIDIA GPUs, and the total cost to build the models is far higher than the headline pre-training figure. Credible estimates suggest Deepseek has at least $500 million of GPUs at their disposal, partially acquired prior to export restrictions. Further, the idea that these are some bootstrapping hobbyists is also not true — while not a household name until recently, the firm has been a leader in the Chinese AI space with an impressive research team that has been watched by insiders for some time.

Even still, there is still little debate that models are impressive. The transparent nature and accompanying research papers make the methods clear and traceable. In short, Deepseek struck a blow to the AI thesis, or more accurately, the AI trade that has dominated U.S. equity markets for the past two years.

Arguably, we can trace the ethos of American AI development back to a paper released by OpenAI researchers five years ago in January 2020, titled Scaling Laws for Neural Language Models. The abstract gets to the point (emphasis added):

[Large Language model] loss scales as a power-law with model size, dataset size, and the amount of compute used for training, with some trends spanning more than seven orders of magnitude. Other architectural details such as network width or depth have minimal effects within a wide range.

The paper suggests that three variables — parameters, data, compute — are the keys to model performance and highly predictive of future model results. They likened the relationship to the natural laws of physics, such as the ideal gas law. The implication was straightforward:

Our results strongly suggest that larger models will continue to perform better, and will also be much more sample efficient than has been previously appreciated. Big models may be more important than big data.

Over the past five years, LLMs have indeed gotten much bigger and better. In practice, bigger means more GPUs, more racks, more interconnects, more electricity. Call it what you want — picks and shovels, infrastructure, datacenters, hardware — these companies have been the biggest beneficiaries of the AI trade to date. NVIDIA leads the pack.

But Deepseek has challenged OpenAI’s version of physics. By making a more efficient training architecture, including utilizing a reinforcement learning feedback loop, the model is able to produce impressive results more efficiently, without human supervision.

To be clear, more efficient LLMs are a positive development for the commercialization of AI, reducing the costs of model development and lowering the barriers to entry for competition. But perhaps the road to AI dominance may not be paved with ever-escalating capex forecasts. The next leg may focus more on architectural and training advancements than brute force. In other words, the infrastructure trade may be souring.

This story reached a fever pitch this past weekend, and the market reaction on Monday was brutal.

Gapping down 12% on the open, NVDA ended the day down 17% — the largest single day decline dating back to the March 2020 COVID crash. In some cases, the pain was even worse for tertiary plays riding the same datacenter thesis. (The four nuclear power plays I highlighted last week — SMR, OKLO, VST, CEG — plunged 20% - 30%, though have recovered some of their losses.)

This is the official story. But it’s not the whole story. It may not even be the main story.

Read more

A Nuclear Renaissance is Upon Us! Haven’t you heard?

Last year, the mood in the power sector turned giddy. The voracious energy appetite of datacenters has provided the first long-term bullish case for power demand in decades, transforming stagnant industry players into market darlings (see: Datacenters and Power Demand). Zero-carbon baseload nuclear energy suddenly became a key ingredient in technological ambitions.

The momentum has continued this year. So far in 2025, nuclear power has dominated the market charts. Four of the eight best performing large-cap1 stocks are nuclear plays — Oklo Inc (OKLO), Constellation Energy (CEG), NuScale (SMR), and Vistra Corp (VST) — which have grown 32% - 75% in market value over the past three weeks.

Of course, there are tangible catalysts supporting the hype. Upsized demand forecasts and ever growing datacenter capex budgets have reshaped grid planning and reliability needs. The Inflation Reduction Act (IRA) has also begun providing new critical subsidies to nuclear plants. And announcements of datacenter co-locations and nuclear plant restarts have put weight behind the claims of a nuclear resurgence:

• Palisades Restart: With the support of billions in loans and grants from state and federal agencies, Holtec intends to restart the 800MW Palisades nuclear facility shuttered in 2022 (Sept. 2023)

• Susquehanna Datacenter: Amazon purchased a 650MW datacenter co-located on Talen Energy’s (TLN) Susquehanna nuclear plant facility, entering into with a 20-year power purchase agreement (PPA) (March 2024)

• Three Mile Island Restart: Microsoft has entered to a new 20-year PPA with Constellation that would support the restart the 835MW Three Mile Island I nuclear facility shuttered in 2019 (October 2024)

• Oklo LOIs: Oklo announced 750MW of incremental letters of intent (LOIs) with two unnamed datacenter companies to purchase its prospective small nuclear reactors (November 2024)

• V.C. Summer Nuclear Station: Just this week Santee Cooper announced that it intends to market its partially built 2,200MW nuclear station — a project that was abandoned in 2017 due to cost overruns (January 2025)

For an industry accustomed to economic pressure and massive capital overruns, retirements and project abandonments, this flurry of activity is a welcome change of fortune.

But a Nuclear Renaissance? It’s time to dump some cold water on this hot sector.

Hard Truths

The nuclear power industry in the United States has been stagnant for over thirty years. The number of operable units peaked in 1987 and net generation has been flat since the late 1990s. The average operating U.S. nuclear power plant is 42 years old.

The two most recent new reactors, completed in 2023 and 2024 at Georgia’s Vogtle plant, were not driven by datacenter demands but instead marked the completion of a project than began construction in 2009, delivered fourteen years later at a total cost of over $30 billion for just 2,200MW of generating capacity.

There are currently no new large scale power plants in development in the U.S., and if one were to be announced today, would likely take well over a decade to complete. Simply put, the capital cost of newbuilds is prohibitive.

The federal Energy Information Agency’s (EIA) benchmarking shows that newbuild nuclear costs 4-7x that of new natural gas or renewables generation. Even natural gas with 90% carbon capture — which mirrors the low-carbon benefit of nuclear — is estimated to cost less than half of a newbuild nuclear reactor.

The actual experience at Vogtle was far worse than these estimates, and the more recent V.C Summer project was abandoned entirely due to massive cost overruns. While the stated restart budgets for Palisades and Three Mile Island are more competitive, history cautions against trusting these figures and timelines, particularly since these are the first nuclear restart projects ever attempted.

Even if Three Mile Islands and Palisades meet their budgets, the net impact of these restarts is limited. At a combined capacity of 1,635MW, these two projects would add just 1.7% to the total U.S. nuclear generating capacity. And they won’t happen immediately. While Holtec continues to push an aggressive 2025 timeline, Three Mile Island is not slated for completion until 2028.

And there are very few, if any, additional restart candidates. Most retired plants are too far down the decommissioning and disassembly process — the prospect of repairing and restarting long-idled plants presents an even more formidable challenge.

Beyond capital costs, existing nuclear plants struggle to remain profitable against natural gas and renewables. In fact, the greatest boon for the nuclear sector hasn’t been new datacenter demand, but rather direct subsidies through Production Tax Credits (PTCs) granted in the IRA, which cushion profitability as power prices decline below breakeven costs.

The Microsoft/Three Mile Island’s PPA is telling. At $100/MWh, Microsoft will be purchasing power at a ~150% premium to the current wholesale electricity rates of $40/MWh in NYISO, New York’s power market.

While Microsoft may have pockets to afford this $13 billion expense over 20 years, it speaks to the underlying economic challenge. The two restart projects require off-market PPAs or grants for viability. They are not competitive in pure market terms and must be subsidized by the government or by hyperscalers’ discretionary carbon goals.

Since electricity is the largest variable cost in datacenters, it seems inevitable that power purchasers become quite a bit more price conscious as products are commercialized, profitable margins are required, and players compete.

Meanwhile Small Nuclear Reactors (SMRs) face three critical challenges.

• They don’t exist. NuScale’s first VOYGR plant is now projected to be operational by the early-to-mid 2030s, following the cancellation of its Carbon Free Power Project in 2023 due to cost overruns. Other companies like Oklo and TerraPower Energy are targeting ~2030 for their first reactors, but significant technical, regulatory, and supply chain challenges remain.

• They won’t be cheap. Beyond technical feasibility, there is little evidence to suggest such projects will be economically viable, especially in their early iterations.

• They are small. If the goal is to meet gigawatts of new demand, “small” reactors seem counterintuitive. Even if early prototypes are delivered, they will not reach a threshold of materiality in the power market in the next decade.

Despite the hype, incremental nuclear generation will not supply the electricity that is expected to be required over the coming years. Subsidies, restarts, and new power agreements will simply maintain overall nuclear power generation at its current levels.

So what should we make of public market action? Let’s take a look at some power producers, miners, and SMRs that have been riding the hype.

Read more

It’s a simple question, but impossible to answer precisely. From the top down, we might say that the price is determined by the supply and demand for chocolate, peanuts, and caramel (balanced against the thousands of other tasty alternatives). From the bottom up, we could try to analyze some fifteen raw inputs, manufacturing, wholesale, and retail costs and profits, and the state’s rake.

But endless variables confound even the sharpest pencils — profit motives, productivity improvements, inventory overhangs and discounts, local costs, human preference and expectation, and the immeasurable elasticity of supply and demand. And yet, a Snickers bar has a price; $2.67 at my closest convenience. Last year it probably cost less, next year it will likely cost more.

In 2024, $457 million of Snickers were sold in the United States — roughly 0.002% of Gross Domestic Product (GDP).

This is the problem with inflation. Stable prices are critical to economic stability and a key mandate of monetary policy in a fiat world. But we don’t really know why prices change. The economy is simply too vast, interconnected, and complex to model with precision. Absent perfect models, the ultimate arbiter is observation.

Today, as the concerns about inflation and monetary policy have once again come to the fore, the market has turned to obsessive observation. This week, the market received December Producer Price Index (PPI) and Consumer Price Index (CPI) data — both showing increases in YoY readings from the prior month. These figures were instantly sliced into their smallest components and measured against expectations in basis points. Outliers are identified and discarded. In the end, there was collective agreement that the data was “Good” relative to expectations. Stocks and bonds rallied on the hopes that the Fed will continue to cut.

But maybe we are losing the plot — lost in the weeds of soy lecithin. The endless dissection of monthly numbers ignores root causes — monetary aggregates, cyclical and secular trends, fiscal policy. Even if we can’t translate these forces into precise price changes, perhaps they still provide predictive value.

After all, the surge of inflation from 2021-2022 was not random. It was driven by a rapid expansion in money (however defined), expanding credit (bank, non-bank, and sovereign), and purchasing power (via fiscal transfers), against a backdrop of labor shortages and supply disruptions (energy in particular)1.

Likewise, the rapid disinflation from 2022 - 2023 was not “immaculate” but driven by a confluence of strong disinflationary forces. Money aggregates shrank and asset prices declined. Interest rates rose and risk spreads widened. Credit and capital markets closed. Labor markets cooled. Goods prices declined. Energy prices fell as production returned to pre-pandemic levels.

Yet even today, prices climb faster than at almost any point in the prior decade, certainly faster than the Fed’s 2% inflation target. And as we look forward, there are signs that the disinflationary trend may be ending.

Money, Credit, Assets

The official policy story is that the current monetary stance is restrictive and further cuts are compatible with ongoing disinflation towards a 2% goal. Reality suggests otherwise.

There was indeed a point in which policy was restrictive and helped lower inflation from its peak. But those days are behind us. Many measures suggest the maximum impact of this tightening cycle occurred almost two years ago in early 2023. Since then, the credit and capital cycle has been in an upswing.

Read more

What did you miss in this first, abbreviated, trading week of the year? Not a whole lot, to be frank.

In markets, some meme-stocks meme’d but the indices haven’t bucked their slump. In labor, ADP data came in a touch soft but unemployment claims were strong. ISM surveys skewed positively. A decline in consumer credit gave the bears something to chew.

Don’t worry, things will get more interesting with inflation data and earnings next week. Later in the month, we will see if the Federal Reserve pushes pause. In the meantime, let’s talk about capex, credit spreads, and common sense.

The Tipping Point

This is a bad idea. But hear me out.

Read more