The AI Infra MapFROM WATTS TO TOKENS

Baseten

San Francisco, US · founded 2019
L6 · AI Runtime, Serving & Model Access
inference & servingstartupprivateseries C+

Baseten provides enterprise AI model serving infrastructure that enables software teams to deploy, manage, and scale machine learning models in production. It supplies an open-source model packaging framework alongside serverless GPU infrastructure supporting fast cold starts and auto-scaling rules. Engineering teams deploy both open-weights models and proprietary PyTorch architectures on its platform.

Ecosystem functionDelivers production model serving software coupled with elastic GPU hosting designed for custom model execution logic.

Business modelPay-as-you-go GPU compute runtime charges

Revenue—
Valuation$13Breported
Last round$1.5Bled by Altimeter Capital · 2026
Sources · 3 ↓
  1. https://www.baseten.co
  2. https://forgeglobal.com/martian_stock/
  3. https://www.baseten.co/blog/why-inference
Suggest an editLast verified 2026-07-26
← Back to L6