Deep Dive
SpaceX and the Warning Signs of Excess
Chanos opens by directly comparing SpaceX to the Enron era — not in specifics, but in timing and pattern. He argues Wall Street is about to unleash massive equity issuance in 2026, with SpaceX kicking off a wave that will include OpenAI, Anthropic, and others, shattering records for IPOs and secondaries. Historically, these periods coincide with market peaks: 1999-2000, 2021 meme stocks and SPACs. The key observation is that supply is finally meeting demand after three years of restraint, signaling a regime shift. Chanos treats this IPO surge not as a positive for markets but as a warning flag that investors should reduce risk. The valuation is his smoking gun: at 110 times revenues, SpaceX sits at a level that has historically produced terrible returns.
The Mysterious Pivot to Data Center Leasing
A week before filing, SpaceX made a sharp strategic turn. Previously their prospectus highlighted enterprise solutions through Grok — developing and selling AI software. Instead, they pivoted to leasing compute capacity to Anthropic and Google, a model Chanos calls the 'Neo Cloud' strategy. This is fundamentally a rate-of-return business: buy chips from Nvidia, lease them out. The problem is that 22 of their 29.5 trillion dollar TAM now rests on this commoditized model, which commands far lower valuations on public markets than actual model development or hyperscaler status. Chanos finds it odd that SpaceX would deliberately choose a lower-margin business model right before going public, unless the higher-growth narrative no longer held water. This pivot represents a retreat from hopes and dreams to financial reality.
The Starlink Reality and Starship's Unproven Path
Chanos does credit Starlink as a real, growing business worth a couple hundred billion dollars. But he's quick to note the launch service still loses money after 22 years, and Starship — the rocket everything depends on — has failed to achieve Earth orbit in 12 missions, with half resulting in mishaps. The entire valuation castle rests on a vehicle that hasn't proven it works at scale. Beyond Starlink, SpaceX's real hopes and dreams center on the X (formerly Twitter) AI division, a company purchased in February for 250 billion in stock that the market now values well over a trillion and a half. But that's pure optionality, not established cash generation. So the portfolio consists of one real business (Starlink), one loss-making legacy business (launch services), and two unproven moon-shots (Starship and X AI) being valued as if all four are thriving.
The Dot-Com 2.0 Pattern and Earnings Illusion
When asked if this resembles the dot-com bubble, Chanos says it's much bigger relative to GDP and the economy than 1999-2000 was. He identifies a crucial accounting dynamic: when Tom builds a data center and buys chips from Nvidia, Nvidia recognizes that as immediate revenue and profit. Tom, meanwhile, capitalizes those expenses and amortizes them over 5 to 10 years. The same dollar gets counted as earnings by Nvidia and deferred by the buyer. This mismatch creates an optical earnings boom. In 1998-2000, the S&P operating earnings rose 30% over two years as the CapEx frenzy peaked. Then, when order books collapsed from mid-2000 to mid-2001, those same earnings dropped 40%. The current AI build-out is following the exact same pattern, suggesting a reckoning lies ahead.
Tesla as a Counterpoint and the Elon Premium Question
Hosts raise Tesla as a comparison point: it's returned 53% per year since COVID, trading at roughly 14 times revenues with 100 billion in annual revenue. Bloomberg's BQ function shows 27 times revenues, which Chanos finds historically poor, but it's a different order of magnitude from SpaceX's 110x. The question becomes whether a Tesla and SpaceX merger could make sense, a possibility Chanos acknowledges could happen but doubts would meaningfully add value through cost cutting. He frames the real issue as the Elon Premium — the market is valuing two high-growth hopes-and-dreams companies (Tesla with Optimus and autonomous driving, SpaceX with AI and Starship) at aggressive multiples. Does owning both double the premium or create an either-or choice? Chanos's view is clear: neither is being valued on actual operations, and the market is betting entirely on unproven future breakthroughs.