Baseten Nears $1.5B Round at $13B in Inference Boom

AI inference startup Baseten is close to finalizing a roughly $1.5 billion funding round at a $13 billion valuation, according to a TechCrunch report. The deal is split-priced — some investors are reportedly entering at an $11 billion valuation — and is led by Spark Capital, Sands Capital, Altimeter Capital and Wellington Management. None of that is officially confirmed yet, but the trajectory is the story.

The pace is the eye-catching part. Five months ago Baseten raised a $300 million Series E at a $5 billion valuation; nine months before that, a $150 million round. If the new numbers hold, the company's valuation has jumped about 160% in under six months. Founded in 2019, Baseten has gone from a quiet infrastructure name to one of the most aggressively funded companies in the AI plumbing layer.

What Baseten actually does

Baseten doesn't build its own large models. It runs other people's — the work known as inference, everything that happens after you hit send on a prompt. Its pitch is routing each request to the best model for the task, and in particular steering work toward capable but cheaper open-source alternatives instead of always reaching for the priciest frontier model. The selling points are speed and cost control: same answer, less money, lower latency.

That is why investors are circling. For most of the AI boom the money and attention went to training — building ever-larger models. But once a model exists, the recurring cost is serving it to millions of users, and that bill never stops. The industry has taken to calling the rush to fund this layer the "inference gold rush," and Baseten's back-to-back mega-rounds are a clean example of where capital is now flowing.

There's a note of caution worth keeping. A 160% valuation leap in months, on a split-priced round structured to flatter the headline figure, is the kind of move that looks brilliant in a rising market and brittle if the AI spending mood cools. The demand for cheap, fast inference is real and growing — but valuations climbing this fast are a reminder that the AI infrastructure market is being priced for a future that still has to show up.