The AI industry's private financing era is ending. SpaceX filed the largest IPO prospectus in history this week, seeking $80 billion and pulling back the curtain on its AI operations, which posted $6.4 billion in losses. OpenAI plans its own public offering within days, targeting a $1 trillion valuation by September.
For nine years, venture capitalists and sovereign wealth funds have controlled AI valuations behind closed doors. Public markets operate differently. They demand profitability or a credible path to it. They scrutinize unit economics. They punish hype.
SpaceX's filing reveals the cost of scaling AI infrastructure at the frontier. The company runs Grok, a chatbot competing with ChatGPT and Claude, alongside its core space business. The AI division's massive losses show what happens when you build language models and serve them at scale without clear monetization. The broader SpaceX business props up these bets.
OpenAI's trillion-dollar target raises sharper questions. The company lost money last year despite generating over $3 billion in revenue. It burns through billions on compute. Its user base remains modest compared to traditional software giants. A trillion-dollar valuation assumes either explosive growth in revenue, dramatic cost reductions, or both.
Public markets care about cash flow and return on invested capital. Private markets accept losses if growth metrics look exponential. Retail investors and pension funds operate with different risk tolerances than Sequoia and Saudi Arabia's Public Investment Fund.
The timing matters. AI hype has peaked. Recent research has shown diminishing returns in scaling language models. Regulatory scrutiny is mounting. Several AI startups have underperformed their private valuations in public debuts.
Neither company needs public money to operate. SpaceX generates actual revenue. OpenAI has wealthy backers. These IPOs are about legitimacy and