Harry Stebbings gives OpenRouter as a standalone product less than five years to survive

Harry Stebbings does not think OpenRouter survives as a product. The five-year window he puts on that call is explicit, which makes it checkable rather than impressionistic, and his assigned probability clears 51 percent, which means he is treating it as a likely outcome rather than a tail risk.

The structure of the prediction matters as much as the verdict. Stebbings is not arguing that OpenRouter fails. He is arguing that product survival and revenue survival are separate questions, and that the routing capability has a good chance of becoming a meaningful revenue stream for Stripe, somewhere in the range of 20 to 30 percent of the company’s revenue. What he doubts is whether it carries the OpenRouter name and identity through that transition rather than becoming a Stripe primitive that people stop identifying by its original brand. The distinction between “the product disappears” and “the product gets absorbed and renamed” is important: one implies failure, the other implies success of a particular, quiet kind.

The underlying logic connects to a familiar pattern in infrastructure acquisitions. Tools that route, aggregate, or normalize access to underlying services tend to be absorbed by the platforms that have the most to gain from owning that layer. OpenRouter’s value to developers was access to a wide range of AI models through a single interface, a routing layer that reduced dependence on any one provider. Stripe’s existing business already sits between merchants and money, performing a structurally similar function in payments. An AI routing layer that sits between developers and models fits the same architectural template. From Stripe’s perspective, the capability is legible even if the domain is new.

I don't even think this product will exist in five years. But I think it will be I think there's a high chance more than 51% chance it builds into a into a 20 or 30% revenue stream for Stripe Harry Stebbings

The tension Stebbings’s call surfaces is one that any infrastructure acquisition introduces when the acquired product had built its identity around independence. OpenRouter’s pitch to developers included the argument that using it avoided lock-in to any single large platform. Stripe is now among the largest financial infrastructure companies in operation. A product that sold freedom from platform risk is now owned by a platform. Whether developers treat that as a meaningful contradiction or as an acceptable trade for stability and scale will shape how much of the user base holds together, and how long the brand remains worth maintaining as a distinct entity.

None of that settles the call. Stebbings’s 51 percent threshold is deliberately modest. It signals conviction without certainty, which is the honest register for a five-year prediction on a product category that matured quickly and changed hands faster than most observers expected. The modesty of the threshold is not a hedge; it is a statement that the outcome is genuinely uncertain but that the direction is clear enough to name. What Stebbings is really saying is that the routing layer is worth keeping and that Stripe knows it, but that the act of keeping something valuable inside a large platform is usually the beginning of the end for the brand that carried it into the acquisition.

Whether OpenRouter follows that pattern or becomes an exception to it is the question the next five years will answer. The routing capability itself, in Stebbings’s reading, is durable. The product as a named, independent thing facing the developer market on its own terms is what he doubts will last. That is a specific enough claim to be wrong, and the five-year window he has set makes it a call that will eventually close rather than one that quietly fades from memory without a verdict.