OpenAI Is Worth Less Than Its Younger Rival

The Surprising Inversion at the Heart of IPO Summer

The company that made the world obsessed with AI chatbots is, by most available metrics, the weakest business heading into public markets this summer. OpenAI's confidential IPO filing, announced this week, lands with a paradox baked into it: Anthropic, founded six years later and widely considered the quieter, more cautious lab, now holds a higher valuation, a stronger revenue run rate, and a commanding lead in the enterprise market that actually pays the bills.

That is not the story the hype cycle told. But it is the story the numbers are telling.

The Gap Is Bigger Than the Headlines Suggest

OpenAI currently carries a pre-IPO valuation of $852 billion. Anthropic sits at $892 billion. The gap looks narrow until you look underneath it. Anthropic's revenue run rate is reported at $47 billion. OpenAI's, depending on which figure you trust, is somewhere between $13 billion and $25 billion. The $13 billion figure came from CEO Sam Altman himself at the end of 2025. More recent reports pushed that number up to $25 billion, but even at the higher end, Anthropic is generating roughly twice as much revenue.

The enterprise market tells an even starker story. Anthropic has been almost single-mindedly focused on B2B customers and now controls about 40% of that segment. OpenAI controls 27%. Enterprise is where durable, recurring, high-margin revenue lives. It is also where institutional investors will be looking hardest when S-1 filings go fully public.

To justify its current valuation, PitchBook estimates OpenAI would need to produce $95 to $105 billion in free cash flow by 2030. Right now, the company is projected to lose $10 to $30 billion a year. That is not a rounding error. That is a business model question.

Why OpenAI Cannot Wait

The filing timing is not coincidental. OpenAI is in a race it cannot afford to lose on the calendar. Public markets are being asked to absorb an extraordinary sequence of mega-offerings in a compressed window. SpaceX goes public Friday, targeting $75 billion at a fixed price of $135 per share. Anthropic is next. OpenAI follows.

Collectively, these companies are seeking hundreds of billions of dollars from markets that historically source around 90% of IPO capital from institutional investors. Institutional appetite is finite, and it tends to flow toward the strongest business. If Anthropic prices first and draws the serious institutional money, OpenAI faces a much harder room.

The third-phase vision OpenAI published this week, covering an automated AI researcher, broad economic acceleration, and personal AGI for every person on earth, reads like an investor roadshow deck designed to fill the gap where hard financials would normally go. Whether that narrative holds under scrutiny from institutional analysts is another matter entirely.

The Retail Investor Mechanism

SpaceX is solving a different version of the same problem. With institutional demand uncertain, the IPO has been engineered to attract retail investors at an unusual scale. Up to 30% of shares are being offered to retail buyers, compared to the typical 5 to 10 percent. Fidelity is lowering its account minimum to $2,000 to participate. There is a 15-day lockup on those retail shares, which is not coincidental: SpaceX needs investors to hold for at least 15 days to qualify for inclusion in the Nasdaq 100 and Russell indexes.

The Nasdaq 100 changed its rules to allow SpaceX entry within 15 days of listing. The S&P 500, to its credit, did not. It kept its profitability and governance requirements in place, which means passive S&P investors will not be automatically exposed. But anyone holding a Nasdaq 100 index fund will own SpaceX whether they chose to or not.

The strategy echoes Tesla's evolution into a stock where price and underlying business fundamentals became largely disconnected. That can work for years. It also creates real risk for retail participants who do not understand what they are buying into.

The Long-Run IPO Problem

The historical record on IPO investing is not comforting for anyone who buys in after the first-day pop. Research from University of Florida professor Jay Ritter, covering 9,300 US-listed IPOs from 1980 to 2024, finds that buying at the end of an IPO's first day and holding for three years produces returns roughly 21% below a value-weighted market index. The people who benefit most are pre-IPO investors, and the lockup schedules confirm exactly that priority.

What Comes Next

The IPO wave of 2026 is genuinely historic. That does not make it a safe entry point for ordinary investors. The companies going public are burning cash at scale, racing each other to markets for structural reasons as much as business readiness, and relying on index rule changes and retail enthusiasm to fill gaps that institutional demand may not.

OpenAI's filing is the most revealing moment of the sequence so far. It confirms that being first in the public imagination does not mean being first in the fundamentals. Anthropic built a quieter business and ended up in a stronger position. The public markets will now have to decide what that is worth, and what OpenAI's vision of AGI for all is worth against a spreadsheet that does not yet add up.