Anthropic Is Slowing. OpenAI Is Catching Up. What Happens to Those Trillion-Dollar Valuations?

Anthropic has the bigger revenue number; OpenAI has the faster growth. Neither has an entirely comfortable pitch for investors considering trillion-dollar valuations. TickerTrends estimates put Anthropic’s annualized revenue run rate at $76.8 billion as of September 24, up just 1.8% month over month. OpenAI stood at $50.44 billion, up 13.4%. Anthropic still led by more than $26 billion, but its explosive growth was showing signs of leveling off. With both companies heading toward eventual IPOs, investors face an awkward combination: substantial sales, a slowing leader and a challenger that has yet to close the gap. The revenue is impressive. The trajectory behind the price tags is getting harder to explain.

These figures measure annualized revenue run rate: current sales extrapolated over a full year. The companies still have to earn that money. TickerTrends updates its estimates between company disclosures using signals of demand and adoption. Accounting also complicates the comparison. As Sacra explains, Anthropic books the gross value of certain sales through cloud partners, then records the partners’ share as an expense; OpenAI books only its retained share of certain partner sales. Anthropic leads on the published figures, but the size of that lead depends partly on the accounting. Public-market investors will need to establish how much of the headline turnover stays with each business.

An earlier snapshot of coding products shows the competitive pressure beneath those company-wide totals. For the week ending August 10, TickerTrends estimated Claude Code’s annualized revenue at $15.12 billion, against $8.83 billion for OpenAI Codex. Claude Code was 1.71 times larger. Yet between July 6 and August 10, Codex’s tracked figure grew about 50%, compared with roughly 6% for Claude Code. The reported gap narrowed from $8.38 billion to $6.30 billion. Those are August product estimates, separate from the September company totals. They nevertheless raise a question for the IPO pitch: how long can Anthropic command a leadership premium if a rival is gaining ground that quickly in one of AI’s most commercially convincing markets?

October’s disclosures supplied something firmer than a tracker estimate: quarterly revenue. Bloomberg reports that Anthropic generated $16.8 billion in the third quarter, about 45% more than in the second, following growth of roughly 140% the previous quarter. The company also told prospective investors it had positive free cash flow and an adjusted gross margin of 85%. That materially strengthens its IPO case: there is evidence of financial returns from the business already being built. Bloomberg’s analysis, however, separates a successful Anthropic offering from a revival of the broader AI trade. A good business can go public for considerably less than its prospective shareholders had penciled in.

Expectations explain why such strong numbers can still disappoint. MarketWatch cites Jefferies’ assessment that Wall Street was probably pricing in an Anthropic revenue run rate of at least $80 billion, with hopes of exceeding $105 billion by year-end. In Bloomberg Opinion, Shuli Ren treats the slowdown as a threat to the trillion-dollar AI trade itself. When the valuation already assumes another extraordinary leap, growth of several dozen percent can prompt complaints about how little the company delivered.

Anthropic is delivering an immediate shock to Wall Street: a dramatic revenue slowdown that threatens to deflate the trillion-dollar AI trade. — Bloomberg Opinion

At the roughly $2 trillion valuation under discussion, Anthropic would cost about 26 times its September annualized revenue, using TickerTrends’ estimate. Annualize the $16.8 billion actually earned in the third quarter and the multiple rises to about 30. Both calculations show how much investors would be paying beyond today’s sales. As we examined in our earlier IPO analysis, the valuation rests on a forecast of $190 billion to $200 billion in annual revenue in 2028. A prolonged slowdown would make that target harder to reach and push the expected cash flows further into the future.

OpenAI has more time to deliver its next surge, but little reason to assume its valuation is insulated from a sector-wide repricing. Bloomberg has reported that the company pushed its IPO beyond 2026 and was discussing a bridge round of roughly $30 billion at a valuation of about $1.4 trillion before the new money. Against September’s roughly $50 billion revenue run rate, that is a multiple of about 28. OpenAI has also told investors it expects to reach or exceed a $70 billion run rate by year-end. Delivering that forecast would bring the same valuation down to about 20 times annualized sales. The next few months are doing a lot of work in the financing pitch: they must help justify both the private round and the eventual public offering.

If Anthropic keeps slowing and OpenAI remains the challenger, investors lose a convenient story about two exceptional future giants, each deserving its own exceptional premium. A lower-priced Anthropic IPO would establish a public benchmark for OpenAI and other model developers. Faster growth could improve OpenAI’s standing against its rival even as the sector’s valuation multiple falls. Sales could rise while the company’s value declines. Raising the same amount through an IPO would then require selling more shares. Owners unwilling to accept that dilution could seek another bridge round and postpone the listing. Construction bills can keep climbing even as equity valuations fall.

Contracted spending is particularly bad at following revenue downward. According to Reuters’ reporting on Anthropic’s prospectus, about 80% of its infrastructure commitments, under a program of at least $518 billion over roughly a decade, are either non-cancelable or payable regardless of usage. Weaker demand would reduce cash coming in without necessarily reducing payments for reserved computing capacity. One quarter of positive free cash flow improves Anthropic’s position. Sustaining it depends on whether future demand justifies the contracts already signed. The capacity was booked for a much more confident growth curve.

The consequences extend well beyond the two listings. We previously covered the roughly $500 billion in market value lost on October 8, when OpenAI told investors its revenue run rate was around $50 billion rather than the widely circulated $70 billion. Even an accounting clarification was painful enough to hit AI infrastructure suppliers’ shares. A sustained sales slowdown would put the underlying economics in question: data-center utilization, future accelerator orders and model developers’ ability to pay for computing. Microsoft, Amazon, Google, Oracle and the chipmakers collect different portions of the money moving through that chain. Their expansion plans nevertheless depend on a shared forecast of demand.

Successful IPOs for Anthropic and OpenAI are therefore perfectly compatible with a lower valuation for the industry as a whole. Public investors could willingly buy both companies while refusing to pay the old premium for years of exceptional growth. Preserving trillion-dollar valuations will require repeatable profitability, sustained increases in customer spending and returns on the computing capacity already contracted. If the leader levels off before the challenger catches up, rising revenue becomes a much less convincing excuse for any price. Going public will give the promised future a ticker symbol—and let investors reconsider what they are willing to pay for it every day.