Baseten just raised $1.5B because your AI inference bill is about to become your largest IT line item

Baseten just closed a $1.5B Series F at $13B valuation.

Not for training. For inference.

The part of AI where trained models actually generate responses for your users. The part that runs 24/7. The part that scales with every feature you ship.

Revenue grew 20x year-over-year. 1 billion inference calls per day. 87 clusters. 18 clouds.

Here's what your CFO needs to hear.

Leading application-layer companies now direct 30-50% of model spend toward custom and post-trained models. Not frontier APIs. Custom models built on their own data, optimized for their own workflows.

That's not a trend. That's a structural shift in how AI budgets move.

Baseten processes inference for Cursor, Mercor, Notion, HubSpot, and Abridge. These aren't experiment teams. These are production AI systems generating real revenue.

The inference layer is becoming the cloud computing of this decade. And the companies that lock in their infrastructure partners now will have cost advantages competitors can't replicate in 2027.

Audit your AI inference spend today. If you're still routing everything through a single frontier API, you're paying a premium for lock-in. Multi-model strategies aren't optional anymore. They're the new cost floor.

SOURCE: https://www.businesswire.com/news/home/20260622645563/en/Baseten-Raises-%241.5-Billion-to-Power-the-Next-Era-of-AI-Inference
VERIFIED: BusinessWire, Pulse2, Verdict.co.uk, Dealroom.co
SIGNAL: Inference infrastructure just became a $13B category. Your AI cost structure is now a competitive moat or a liability.