OpenAI’s Financials Leaked — The Losses Are Staggering

SUMMARY

Ed Elson hosts a discussion with Ed Zitron and Nicolas Owens, examining OpenAI’s leaked financials and the sustainability of its business model, as well as the soaring valuation of SpaceX amid AI-driven market euphoria. The episode highlights concerns about unprecedented losses, aggressive spending, and speculative valuations in the current market environment.

MAIN POINTS

• OpenAI’s leaked financials reveal $13 billion in revenue but a staggering $39 billion net loss, with questions about the sustainability of such spending as the company prepares for an IPO. [1:30]
• Ed Zitron emphasizes the importance of focusing on actual costs and revenues, dismissing attempts to downplay OpenAI’s losses as accounting maneuvers. [6:00]
• Discussion centers on the challenges OpenAI faces in achieving profitability, with skepticism about its ability to become a sustainable business given its spiraling costs. [12:00]
• Nicolas Owens from Morningstar presents a valuation of SpaceX at $780 billion, significantly lower than its current $2.6 trillion market value, citing uncertainties in AI and space infrastructure. [18:00]
• The conversation addresses how SpaceX’s repositioning as an AI company has fueled investor enthusiasm, despite questions about the realism of its projected market opportunities. [24:00]
• Owens discusses the implications of SpaceX’s $60 billion acquisition of Cursor and the potential impact of insider lockup expirations on the stock’s future performance. [28:00]
• The episode concludes with concerns about market-wide euphoria, noting that valuation metrics have reached levels reminiscent of the dotcom bubble, particularly in the AI and space sectors. [33:00]

DETAILED ANALYSIS

The episode opens with a detailed breakdown of OpenAI’s recently leaked financials, which have sparked widespread debate in the technology and investment communities. OpenAI, a company at the forefront of artificial intelligence development, reported $13 billion in revenue last year, representing a more than 250% increase from the previous year. However, this rapid growth came at a steep cost: the company posted a net loss of $39 billion, with $21 billion lost from operations alone.

These figures, independently verified and audited by the Financial Times, have raised significant concerns about the sustainability of OpenAI’s business model as it prepares to go public.

A closer examination of the financials reveals that OpenAI’s expenditures are heavily weighted toward research and development ($19.18 billion), cost of revenue ($7.5 billion), and notably, sales and marketing ($5.73 billion). The sales and marketing spend represents 44% of total revenue, an extraordinarily high proportion compared to tech giants like Facebook and Google at their respective peaks. This aggressive outlay, particularly in sales and marketing, suggests that OpenAI is investing heavily not just in building AI technology but also in promoting and selling its products to achieve market dominance.

Another point of concern is the concentration of revenue sources. Approximately $867 million, or 6.6% of OpenAI’s revenue, came from a single partner, SoftBank, through a project called Crystal Intelligence. The project, still in its nascent stages, raises questions about the reliability and diversity of OpenAI’s revenue streams. Such dependence on a single partner for a significant portion of income introduces risk, especially if the project fails to deliver on expectations.

The episode also addresses the complexity of OpenAI’s reported losses. While some sources argue that the real net loss is closer to $8 billion after adjusting for stock-based compensation and other accounting factors, Ed Zitron dismisses these explanations as attempts to obscure the true scale of the company’s financial challenges. He argues that the focus should remain on the actual cash outflows and operational losses, which are far more indicative of the company’s health than adjusted figures.

Looking ahead, OpenAI’s plans to spend $50 billion on compute in 2026 suggest that its losses could escalate even further. The company’s transition from a nonprofit to a for-profit entity has not yet translated into profitability, and there is skepticism about whether this will change in the foreseeable future. The discussion highlights that OpenAI’s business model, centered around ChatGPT and its API, lacks the diversified revenue streams and operational leverage seen in other tech giants.

The ability of CEO Sam Altman to convince investors of the company’s long-term potential is seen as critical, especially given the daunting financials.

The conversation then shifts to SpaceX, which recently became the fifth most valuable company in the world, tying with Amazon after a significant rally in its stock price. Nicolas Owens, an equity analyst at Morningstar, provides a grounded assessment of SpaceX’s valuation. His analysis, based on a weighted average of three scenarios, values the company at $780 billion—substantially lower than its current market capitalization of $2.6 trillion.

Owens breaks down SpaceX’s business into three main components: rockets, Starlink, and AI. While the rocket and Starlink businesses are seen as mature and possessing strong cost advantages, the AI segment is viewed as highly speculative, with uncertain financial benefits from ventures like data centers in space.

Owens attributes the disconnect between his valuation and the market price to the low public float of SpaceX shares and the market’s willingness to price in highly optimistic, “moonshot” scenarios. He notes that only about 4% of shares are available to the public, making the stock susceptible to meme-like trading behavior. The upcoming expiration of insider lockups is expected to test the stock’s resilience, as a significant number of shares could enter the market, potentially putting downward pressure on the price.

The episode also covers SpaceX’s recent $60 billion acquisition of Cursor, an AI coding startup. While the deal is paid in stock and does not materially alter Morningstar’s valuation, it underscores SpaceX’s strategic pivot toward AI and its efforts to capitalize on investor appetite for artificial intelligence opportunities. Owens suggests that while there is potential for SpaceX to extend its cost advantages into AI infrastructure, the evidence for a sustainable competitive moat in this area remains limited.

In the final segment, the discussion broadens to the overall state of the stock market, which is described as having entered “crazy town.” The Schiller PE ratio for the S&P 500 has climbed to 42, the second-highest level in history, rivaled only by the dotcom bubble of 1999. While not all stocks are seen as overvalued—Microsoft and Meta are cited as relatively reasonable—there is a clear warning that certain sectors, particularly those driven by AI and speculative growth narratives, are at risk of correction. The episode concludes by noting that AI euphoria has firmly taken hold, echoing past periods of market excess.

LINKS

• Prof G Markets YouTube Channel
• Prof G Markets newsletter subscription
• Order 'Notes On Being A Man'
• No Mercy / No Malice newsletter subscription
• Markets on Instagram
• Scott Galloway on Instagram
• Ed Elson on Instagram
• Ed Elson on X (Twitter)
• Ed Elson on Substack
• SoFi private student loans
• ODO all-in-one business software
• Granola AI-powered notepad for meetings

KEYWORDS