Why Steve Eisman Isn't Buying Anthropic's IPO

SUMMARY

Legendary investor Steve Eisman joins Ed Elson to analyze Anthropic's S-1 filing, focusing on financial commitments, revenue trends, and governance issues. The episode also features Dan Chiolan discussing the competitive landscape of autonomous vehicles and concludes with a critical look at misconceptions surrounding Anthropic's IPO.

MAIN POINTS

• Steve Eisman highlights Anthropic's massive future capital commitments and the lack of Q3 financial data in the S-1 filing. [2:00]
• Discussion centers on the implications of $800 billion in compute commitments and the potential for significantly increased capital expenditures. [6:02]
• Eisman criticizes Anthropic's governance structure, emphasizing the lack of shareholder rights due to special share classes controlled by founders. [10:00]
• Dan Chiolan outlines the scale of vehicle deaths and the promise of autonomous vehicles, comparing major players like Tesla, Waymo, and Chinese companies. [16:02]
• The conversation addresses regulatory fragmentation in the U.S. and shifting public sentiment toward autonomous vehicles. [23:01]
• Ed Elson debunks myths about Anthropic's net losses and revenue growth, clarifying the impact of future compute commitments on the company's financial outlook. [29:01]
• The episode concludes by stressing the risks inherent in Anthropic's business model and the uncertainty surrounding its future profitability. [32:01]

DETAILED ANALYSIS

The episode opens with a sharp focus on Anthropic's upcoming IPO, as Steve Eisman immediately expresses skepticism about investing in the company. Eisman draws attention to Anthropic's S-1 filing, particularly the extensive list of future capital commitments, which he estimates to be in the range of $800 to $900 billion. This figure, while staggering, is presented as a central concern for potential investors, as it signals a significant increase in future expenses that are not yet reflected in the company's current financial statements.

Eisman further notes that the S-1 only includes financial data through the second quarter, omitting the third quarter figures that could reveal a slowdown in revenue growth. He references industry reports suggesting that after a period of aggressive 'token maxing'—where companies maximized AI usage regardless of cost—there was a marked deceleration in revenue growth and a reduction in the ratio of revenue to compute expenses in Q3. This shift, coupled with rumors of a 50% price cut by Anthropic to regain market share, raises questions about the sustainability of its growth and profitability.

The discussion then pivots to the implications of Anthropic's massive compute commitments and anticipated capital expenditures. Eisman and Elson debate whether these commitments will translate into direct expenses or capitalized investments, but agree that such large obligations will inevitably impact the company's income statement, likely through depreciation or other accounting measures. These details are buried in the footnotes of the S-1, suggesting a lack of transparency about the true financial trajectory of the company.

Eisman characterizes the S-1 as backward-looking, capturing performance at the peak of AI demand and before increased competition from Chinese open-weight models. The omission of Q3 data is interpreted as a deliberate move to obscure a slowdown in business momentum, further complicating the investment case.

A critical aspect of the analysis is Anthropic's governance structure. Eisman is unequivocal in his assessment, labeling it 'ridiculous' due to the presence of multiple share classes that effectively strip ordinary shareholders of voting rights. The most significant is a single class F share, controlled by the seven co-founders, which holds 50.1% of the voting power.

This arrangement institutionalizes founder control and leaves outside investors with no meaningful influence over corporate decisions. While similar structures exist in other tech companies like Meta, Eisman warns that such governance models reflect a broader trend of founder dominance and a lack of trust in public shareholders.

The conversation transitions to the broader AI and autonomous vehicle sectors, with Dan Chiolan providing an overview of the competitive landscape. The discussion highlights the persistent dangers of driving, with over 40,000 annual deaths in the U.S. alone, and positions autonomous vehicles as a potential solution to reduce fatalities caused by human error. Major players in the autonomous vehicle race include Tesla, Waymo (referred to as 'Whimo'), Amazon's Zoox, and Uber, each employing different strategies and technologies.

Waymo leads in commercialization, operating in 15 U.S. cities and delivering 500,000 paid rides weekly, while Tesla is rapidly scaling its robo-taxi service but remains behind in operational scope. Internationally, the U.S. and China dominate the sector, with China excelling in manufacturing and the U.S. leading in commercial deployment. However, geopolitical barriers prevent cross-border expansion, relegating competition to neutral markets like Europe and the Middle East, where Chinese companies have gained a foothold.

Regulatory challenges are also addressed, with the U.S. lacking a unified federal framework for autonomous vehicles, resulting in a patchwork of state-level regulations and intense lobbying efforts. Public sentiment toward AI and autonomous vehicles is described as increasingly skeptical, with recent polls indicating a decline in confidence despite technological advancements and improved safety records. Investment opportunities in the sector are identified in component manufacturing, particularly LIDAR sensors, maintenance services, and the trucking industry, where automation could yield significant cost savings.

The episode concludes with Ed Elson revisiting Anthropic's S-1, aiming to clarify misconceptions. He explains that the reported $80 billion net loss is largely an accounting artifact stemming from the revaluation of convertible notes, rather than operational underperformance. While revenue growth was impressive in the first half of the year, it slowed considerably in Q3, signaling the end of the AI investment frenzy.

The most consequential figures are the $800 billion in future compute commitments, including substantial contracts with Amazon, Google, and Broadcom, none of which are yet reflected in the company's financials. Elson cautions that these obligations will soon materialize as significant expenses, fundamentally altering the company's profitability outlook. He emphasizes that, despite Anthropic's technological achievements, the business model remains unproven and fraught with risk, making it a speculative venture rather than a secure investment.

LINKS

• Prof G Markets YouTube Channel
• Prof G Markets newsletter subscription
• Order 'Notes On Being A Man'
• Subscribe to No Mercy / No Malice
• Markets on Instagram
• Scott Galloway on Instagram
• Ed Elson on Instagram
• Ed Elson on X (Twitter)
• Ed Elson on Substack
• Gusto payroll and benefits software

KEYWORDS