Baseten in talks to raise $1B at $11B valuation as inference boom accelerates
What's the deal? Baseten, a San Francisco-based AI inference startup, is in talks to raise $1 billion at an $11 billion valuation, according to a person familiar with the discussions. If completed, the deal would more than double the company's $5 billion valuation from January 2026, when it raised $300 million in a Series E led by IVP and CapitalG with a reported $150 million anchor from Nvidia.
Founded in 2019, Baseten pitches itself as "AWS for inference," helping enterprises deploy and run AI models in production. Its customers include Notion, CursorDealroom has a profile for this one. Try Dealroom →, Writer, and HeyGen. In December 2025, Baseten acquired Parsed, a startup that helps companies customise open-source models for niche tasks.
The company's annualised revenue surged from $200 million to $600 million during Q1 2026 — 20 times what it was generating a year earlier. That growth reportedly encouraged some investors to make offers valuing Baseten at around $15 billion.
Why now? The AI infrastructure build-out is shifting from training models to serving them at scale. Analysts expect inference to account for roughly two-thirds of AI compute demand by the end of 2026, up from one-third in 2023.
An explosion of apps built on open-source AI models has been a boon for inference providers. Baseten isn't alone: rival Modal recently raised at a $4.65 billion valuation after its annualised revenue hit $300 million, while Together AI raised at a $7.5 billion valuation after exceeding $1 billion in annualised revenue.
What could go wrong? GPU scarcity and high prices could squeeze Baseten's margins. The company rents Nvidia AI servers from larger cloud providers before leasing them to developers — a capital-intensive model that depends on favourable hardware economics. Demand for chips from the biggest AI developers, such as OpenAI and Anthropic, has driven up spot prices for GPUs, and most established businesses still prefer closed-source models from those two providers.
The signal: The real money in AI is moving downstream. Training frontier models still grabs headlines, but serving those models to millions of users is becoming the larger business. Baseten's trajectory — from roughly $30 million in annualised revenue in March 2025 to $600 million by Q1 2026 — illustrates how fast inference infrastructure is scaling. Baseten has previously raised $585 million in funding from investors including Nvidia, CapitalG, IVP, and Spark Capital.
Read more: The Information · Tech Startups · PYMNTS
