Gimlet Raises $300M for AI Inference Cloud

AI infrastructure is shifting from training scale toward the cheaper, faster and more flexible deployment of models.

What happened

Gimlet Labs raised a $300M Series B led by Andreessen Horowitz. Sapphire Ventures, Menlo Ventures, Arm, M12, Samsung Ventures, Tiger Global and others also participated.

The round values Gimlet at $3B and brings total funding to $392M. The company is building a multi-silicon inference cloud for AI workloads.

Why it matters

Inference is where models become products. As AI moves into everyday software, companies need infrastructure that can run models reliably and cost-effectively across different chip architectures.

Gimlet’s positioning matters because the AI market is moving beyond one-size-fits-all GPU clusters. A flexible inference layer could help companies route workloads to the right hardware for performance, cost and availability.

The bigger picture

The next AI infrastructure winners may not only be training superclusters. They may be companies that make inference cheaper, faster and easier to deploy across many models, chips and enterprise environments.