Stripe's Reported OpenRouter Price Is a $3 Billion Range

Bloomberg reported on 16 August 2026 that Stripe had finalized an agreement to acquire OpenRouter for more than $7 billion, citing people familiar with the matter. Twenty-four days earlier, on 23 July, The Wall Street Journal reported that the two companies were in talks at a valuation near $10 billion.

Both figures are still the only prices in public circulation, and both rest on unnamed sources. Stripe's on-record position has not moved: a spokesperson told TechCrunch the company does not comment on rumors or speculation. OpenRouter declined to comment.

HaiPay reported the July talks at the time and wrote then that no agreement or price had been announced. That sentence is still accurate on 17 August 2026.

The confirmation check, run on 17 August

Both companies operate their own announcement channels, and both were checked directly rather than through coverage of them.

Stripe's newsroom index carried no OpenRouter acquisition item. Its most recent entry was dated 21 July 2026, about Ramp using Stripe for stablecoin payments and accounts — four days before the first deal report and 27 days before the check.

OpenRouter's blog was equally quiet on the subject. Its most recent post, dated 14 August 2026, was a developer guide on sending images to a language model. The eight most recent posts covered benchmarks, routing, evaluation tooling and team spend controls. None mentioned Stripe, an acquisition, or a change of ownership.

That is zero acquisition confirmations across two owned channels, 24 days after the story first broke and one day after it was reported as finalized.

Two reported numbers that cannot be subtracted

The temptation is to read $10 billion in July and more than $7 billion in August as a $3 billion discount. The public record does not support that reading, because the two figures are different kinds of statement.

The Journal's figure described a valuation under discussion, reported as approximate, in talks that the same report said could still collapse. Bloomberg's figure describes an agreed deal price and is open-ended upward: "more than $7 billion" is a floor, not an estimate. A deal at $9.5 billion would satisfy both reports simultaneously.

What the two reports jointly establish is therefore a band, not a point. The values consistent with both run from roughly $7 billion to roughly $10 billion. Fortune added that sources indicated the final price might still change.

That band is at least $3 billion wide. OpenRouter's entire disclosed valuation at its May funding round was $1.3 billion, so the uncertainty in the reported price is about 2.3 times the company's most recent priced valuation.

The implied multiple moves with it. Against the $1.3 billion Series B, more than $7 billion is at least 5.4 times, and about $10 billion would be roughly 7.7 times. Reported annualized revenue of about $50 million as of March — itself a figure attributed to unnamed sources rather than disclosed by the company — would put the range at roughly 140 to 200 times revenue. None of these multiples is calculated from a confirmed price.

The framing hardened as the number loosened

The sourcing behind these reports did not change between July and August. The language describing them did.

The Journal's July report was written as talks. TechCrunch's 16 August headline said Stripe "will reportedly acquire." SiliconANGLE said Stripe "reportedly finalizes." Fortune said Stripe "clinches." Downstream aggregators dropped the hedge entirely, publishing that Stripe "acquires" and that Stripe "completes" the acquisition.

By the time the story crossed into Chinese-language coverage, the qualifier was gone. Tencent Technology's WeChat account published at 00:31 Beijing time on 17 August that Stripe had acquired OpenRouter at a price above $7 billion, stated in plain past tense with no attribution to anonymous sourcing and no note that Stripe had declined to confirm.

Every one of those items traces back to the same anonymous sourcing. The confidence increased; the evidence did not.

The numbers nobody is quoting are the ones that moved

Almost every deal story describes OpenRouter the same way: 8 million users and more than 400 models. Those are the figures from the company's Series B announcement, published on its blog on 28 May 2026, with the accompanying wire release dated 26 May. That round raised $113 million led by CapitalG at a $1.3 billion valuation, and reported weekly throughput of 25 trillion tokens, up fivefold from 5 trillion six months earlier.

OpenRouter's own About page, checked on 17 August 2026, gives different figures: more than 10 million global users, more than 500 models, more than 80 providers and more than 200 trillion tokens a month.

Converting the Series B disclosure to the same monthly basis makes the gap legible. Twenty-five trillion tokens a week is about 108.6 trillion a month, using 4.345 weeks per month. The company's current published figure is therefore at least 1.84 times its Series B run rate, alongside roughly 25 percent more users and 100 more models.

The company's scale disclosures moved substantially in the same 82-day window that the coverage spent recycling the older set.

That 82-day figure, repeated widely as the interval between the funding round and the deal report, is itself date-dependent. Measured from the 26 May wire release to 16 August, it is 82 days. Measured from OpenRouter's own blog post of 28 May, it is 80.

Why the buyer is a payments company

The strategic logic is easier to verify than the price, because it sits in documentation both companies publish.

OpenRouter's own FAQ shows where the money is taken. Buying credits through Stripe carries a 5.5 percent fee with an $0.80 minimum. Buying them with cryptocurrency carries 5 percent. Inference itself is passed through at provider pricing "without any markup." Customers bringing their own provider keys pay 5 percent on usage above a monthly allowance of $25,000 on pay-as-you-go plans, or $200,000 on enterprise plans.

The consequence is specific: the platform's take is charged at the moment a balance is funded, not at the moment a model is called. That is a top-up fee. It is the economics of a payments business wrapped around model routing, which is a concrete reason the buyer is a payments company rather than a model developer or a cloud provider.

The two companies were already coupled at that layer. Stripe's newsroom described the relationship on 29 January 2026: OpenRouter uses Stripe Invoicing to bill customers, Stripe Tax for global tax calculation, and Radar for Fraud Teams for fraud controls, while accepting cards alongside Alipay, WeChat Pay, Amazon Pay, Cash App and Google Pay. That announcement put OpenRouter at more than 5 million developers, and described Stripe automatically tracking usage and applying pricing as model costs change.

PYMNTS reported that Stripe completed a purchase of the usage-metering company Metronome in January 2026. Read alongside the OpenRouter relationship, that gives a consistent direction of travel — metering and billing for consumption-priced software — though the sequence is assembled from reporting, not from a disclosed acquisition strategy.

What is established and what is not

Established: two major outlets have reported a transaction, 24 days apart, at two different figures. Stripe declines to comment on either. OpenRouter declines to comment. Neither company has published an announcement. OpenRouter's disclosed scale has grown materially since May, and its fee structure is documented and charged at funding time.

Not established: the price, the structure, whether the agreement is signed, whether it has closed, any regulatory or closing conditions, whether the reported revenue figure is accurate, and what happens to routing neutrality, provider treatment or customer data governance under new ownership.

The last of those is the question with the most direct bearing on anyone building on the platform, and it is the one on which there is currently no public information at all — not a disclosure, not a comment, and not an anonymous report.

Until one of the two companies publishes something, the defensible statement is narrow. A transaction has been reported, twice, at prices that do not match, by outlets citing people who are not named, about companies that will not confirm it.