Baseten, the San Francisco company that helps businesses deploy and serve machine-learning models in production, has raised $1.5 billion, according to funding data published June 19, 2026. The round ranks among the largest single AI-infrastructure raises of the year and underscores how aggressively capital is flowing toward the "serving" layer of the AI stack rather than only frontier model labs.
Baseten's platform focuses on inference — the step where a trained model actually answers requests in real time. As enterprises move generative AI from pilots into customer-facing products, the cost, latency, and reliability of inference have become a defining constraint. Companies increasingly want to run open and custom models on dedicated infrastructure rather than route every request through a closed API, and Baseten positions itself to handle that orchestration, autoscaling, and GPU efficiency.
The raise lands during a frenzied stretch for AI infrastructure financing. Recent weeks have seen nine- and ten-figure rounds for GPU-cloud and inference startups, with investors betting that demand for compute will keep compounding even as model prices fall. Falling per-token costs tend to expand usage rather than shrink the market, which makes efficient serving platforms more valuable, not less.
For Baseten, the new capital provides room to expand GPU capacity, deepen support for the latest open-weight models, and compete with hyperscaler offerings from AWS, Google Cloud, and Microsoft, all of which are pushing their own managed inference and agent platforms. It also signals continued investor conviction in the "picks and shovels" thesis: that durable returns in AI may accrue to the companies enabling deployment at scale, not only those training the models.
The figure should be read as a reported funding total rather than a confirmed single-round valuation, and Baseten had not detailed investor composition at publication time. Still, the scale of the raise places the company firmly among the best-capitalized independent inference providers heading into the second half of 2026.
Source: [Fundup AI](https://fundup.ai/recently-funded-startups)