Baseten is a US AI inference infrastructure company founded in 2019. It markets itself as an "AWS for inference" — a vendor-neutral inference layer for AI-product builders. As of May 26, 2026 it was reported to be in talks to raise $1 billion at an $11 billion valuation, more than double the $5 billion valuation set roughly three months earlier (Source: techstartups.com).
Type: AI inference infrastructure company. Founded: 2019. HQ: United States. Founders: Amir Haghighat, Tuhin Srivastava, Philip Howes, Pankaj Gupta.
Snapshot
Valuation
| Date | Amount | Round / Event | Source |
|---|---|---|---|
| 2026-05-26 | $11B (in talks) | $1B raise reportedly in talks per The Information (via TechStartups); more than double the $5B valuation set three months earlier. (Source: techstartups.com) | The Information / TechStartups |
| 2026-02 | $5B | Series E ($300M led by IVP + CapitalG; $150M anchor from Nvidia). | The Information / TechStartups |
Business
Baseten positions itself as an "AWS for inference" — a vendor-neutral inference layer that abstracts away the compute substrate (Nvidia GPUs, custom silicon, hyperscaler infrastructure) for AI-product builders (Source: techstartups.com). The pitch is framed against a market in which inference is forecast to reach roughly two-thirds of AI compute demand by end-2026, making the inference layer rather than training the dominant cost center (Source: techstartups.com).
Customers
Per TechStartups (May 26, 2026), named customers include Notion, Cursor, Writer, Gamma, Patreon, Descript, and HeyGen (Source: techstartups.com). The customer profile is weighted toward AI-native applications that build on top of frontier-model APIs but operate their own inference for cost, latency, or product-specific fine-tuning reasons.
Funding and valuation
In February 2026 Baseten raised a $300 million Series E at a $5 billion valuation, led by IVP and CapitalG, with a $150 million anchor investment from Nvidia (Source: techstartups.com). As of May 26, 2026 it was reported to be in talks to raise $1 billion at an $11 billion valuation, per The Information relayed by TechStartups — a 2.2× move from the February figure over roughly three months (Source: techstartups.com).
The pace of that valuation move, at the inference-middleware layer specifically rather than the frontier-lab layer, is treated as a data point in the AI bubble debate. A $1 billion raise at $11 billion for an inference middleware company, following Nvidia's $150 million anchor in the prior round, has been read as a signal that the inference layer is consolidating into a small number of large platforms, alongside CoreWeave on the hyperscaler-substitute side. Nvidia's Series E anchor places Baseten within the broader Nvidia strategic-investment pattern discussed in AI Bubble Debate under "circular financing."
Relationships
- related: AI Bubble Debate — valuation-move data point.
- related: CoreWeave, Nvidia & TSMC — AI Compute Infrastructure — parallel-stack and strategic-investor relationships.
- instance-of: Inference Economics and Token Pricing.
Sources
- TechStartups (relaying The Information), May 26, 2026: "AI inference startup Baseten in talks to raise $1 billion at $11 billion valuation" — (Source: techstartups.com)