It fits a couple of things — 1) Stripe has a history of M&A around developer tooling, dating back to them buying runkit. 2) Hard to argue with the theory of being payments for developers when they're the clear market winner for devs who need card acceptance + they've translated that mindshare to having every major AI company start taking payments on Stripe. Anthropic and Open AI become winners take most? Stripe processes the lion's share of their payments, so they squeeze the payments tool on each subscription payment, enterprise payment and per-usage payment. 3) B2B payments is huge. 4) The future of B2B payments is going to be agents buying things, and sometimes from agents. Who's going to build the agents? Developers. How are they going to be built and operate? On AI tokens. Will OpenRouter deliver the kind of revenue to justify their ~$8B purchase price? Unclear. Stripe needs to get a software or other non-payments multiple on its total revenue, even though that revenue is largely derived from payments. Otherwise, it can fall back to earthly valuations like Block, FIS, Fiserv, Jack Henry and others. Or even —shudder—PayPal. How do you get that not-payments multiple when you're a payments company? One key theory is to diversify, especially with more data and software based revenue. I think Shopify has done the best here. And P, J and Will talking about "more than 10" business lines doing $100M+ (presumably each) seems to hint that that's the strategy. To the extent Stripe is using cashflow from what is historically a low-multiple business (payments) to buy higher multiple revenue businesses (orchestration software), that feels like a good trade. And doing that trade, into a growing space, while you're private and the rest of us have to speculate like this, feels like a solid move. Lets you keep buying with cashflow and potentially inflated stock (only the public markets can really tell). But there's a greater question about this bet, beyond the nuts and bolts work to try to meaningfully diversify away from payments revenue in a business where payments revenue is tracking 2%+ of GDP and all the associated growth — Is this a true business need? Everyone needs payments, and Stripe made payments easy to get in the form you want them (API, POS, dashboard, etc . . .). The letter to investors (my copy must be coming in the mail) touts token spend as something everyone will need, especially corporates. But I doubt it's as high a need as payments, or even as ubiquitous. And it's not clear the Fortune 500 CFOs are going to have their teams adopt OpenRouter as the tool to manage spend. And why would they when they're probably buying 1 to 3 AI tools on an enterprise basis + setting spend limits per seat or per team. The thesis is it's going to be hard to budget and manage AI token spend. But enterprises with lots of money to spend on tokens will need to be convinced they need a tool to do it, and also that OpenRouter is that tool. And this is where there's a chance to out-Stripe, Stripe. Someone else (I think some devilishly handsome founders in NYC) bought and launched router.com. People in my mentions are saying DataBricks does this. It looks like others are also in on the game, or that you could see Amazon or a coding harness eat this feature as well. We're early innings, so if token routing and orchestration is going to continue to grow Stripe doesn't have to be right or have picked THE singular winner. But in a world where they think they're going to sell this into enterprises — the same enterprises whose CFOs and Payments VPs go find multiple payment processors for redundancy and negotiating leverage purposes — it's hard to see how this offering doesn't turn into a leaky back of funnel business that needs to constantly be maintained by an army of human AEs and customer success managers.