Baseten Raises $1.5B Series F at $13B Valuation as AI Inference Becomes the Hottest Infrastructure Layer
AI inference platform Baseten closed a $1.5 billion Series F round led by Altimeter Capital, Conviction, and Spark Capital, valuing the company at $13 billion. The San Francisco startup now processes over 1 billion inference calls daily across 87 clusters worldwide.
Baseten just raised $1.5 billion. Not from a frothy moment of hype — from a company that grew revenue 20x year-over-year and now runs more than 1 billion AI inference calls every day.
The Series F, announced June 22, was led by Altimeter Capital, Conviction, and Spark Capital, with Sands Capital and Wellington Management as co-leads. IVP, Greylock, 01A, Blackbird, Durable Capital Partners, Battery Ventures, and D.E. Shaw Ventures also participated. Total raised to date: over $2 billion. The round closed across two tranches at $13 billion and $11 billion respectively, reflecting strong investor conviction even as the broader VC market remains selective.
What Baseten Does
Baseten builds the infrastructure layer that sits between AI models and the applications that call them. That means GPU orchestration, autoscaling, observability, billing, and developer tooling — all the operational plumbing that teams deploying multi-model AI strategies would otherwise build themselves.
The platform spans 87 clusters across 18 clouds globally. It handles workloads from frontier models (GPT-5.x, Claude, Gemini) alongside custom fine-tuned variants that enterprises increasingly rely on for cost, latency, or compliance reasons.
CEO Tuhin Srivastava put it plainly: “The future of AI will be built on millions of specialized models. Companies using Baseten build a system that compounds in value over time.”
Why This Round Matters
The timing reflects a structural shift in how AI money flows. Model training consumed most of the capital in 2023 and 2024. Inference is where 2025 and 2026 investment is concentrating — and for good reason. Every ChatGPT message, every Copilot suggestion, every agentic API call hits inference infrastructure. As model usage scales, inference cost and latency become the primary levers organizations can actually pull.
Baseten’s pitch is that managing inference across a mixed fleet of frontier and custom models — at scale, with reliability — is genuinely hard, and that most engineering teams shouldn’t be building it themselves.
The $13 billion valuation puts Baseten in the top tier of AI infrastructure companies, behind only the hyperscalers and a handful of well-capitalized foundation model labs.
What Comes Next
The capital will go toward expanding engineering and research teams, adding compute capacity, and scaling go-to-market. The company plans to triple headcount during 2026.
The inference infrastructure battle is now a three-way fight between startups like Baseten and Modal, hyperscaler managed services (AWS Inferentia, Google Cloud TPUs), and the model labs themselves building proprietary serving infrastructure. Baseten’s bet is that enterprises deploying multiple models from multiple providers will always need a neutral layer to manage it — and that the company that owns that layer has durable value regardless of which foundation models win.
One billion inference calls a day suggests the thesis is already holding.