Baseten
Baseten is an inference infrastructure company that hit $600 million in annualized revenue by March 2026, up roughly 1,900% year-over-year, according to Sacra.
Profile
Baseten provides cloud infrastructure specifically for running AI models in production, handling deployment, scaling, and GPU access across multiple cloud providers.
Baseten is an inference infrastructure company that hit $600 million in annualized revenue by March 2026, up roughly 1,900% year-over-year, according to Sacra. The San Francisco-based startup has raised $585 million in total funding, closing a $300 million Series E in January 2026 at a $5 billion valuation led by IVP and CapitalG, with a $150 million investment from Nvidia. As of May 2026, Baseten is in talks to raise another $1 billion at an $11 billion post-money valuation, with reported offers as high as $15 billion.
Founded in 2019 by Tuhin Srivastava (CEO), Amir Haghighat, Philip Howes, and Pankaj Gupta, the company emerged from the founders' long-standing obsession with machine learning, including work on the Netflix Challenge and Kaggle competitions. Baseten's platform lets customers deploy open-source, custom, or fine-tuned AI models on infrastructure optimized for inference, abstracting away GPU procurement, autoscaling, and multi-cloud reliability. Customers include Abridge, Clay, Cursor, Decagon, Descript, EliseAI, Gamma, HeyGen, Lovable, Notion, OpenEvidence, Parallel, Poolside, World Labs, and Writer.
The company monetizes usage-based pricing for API consumption and GPU capacity. Baseten competes with DigitalOcean's AI-Native Cloud, which reported $170 million in AI customer ARR in Q1 2026, and with other inference-focused platforms. The company employs roughly 200 people as of early 2026.
Who buys this
- AI-native startups building and deploying proprietary models
- Enterprise SaaS companies integrating generative AI features
- Large language model developers needing inference APIs
- Organizations seeking cost savings versus proprietary model APIs like OpenAI
- Companies requiring multi-cloud GPU access and high availability
Publicly disclosed clients
- Abridge
- Clay
- Cursor
- Decagon
- Descript
- EliseAI
- Gamma
- HeyGen
- Lovable
- Notion
- OpenEvidence
- Parallel
- Poolside
- World Labs
- Writer
Strengths and what to watch
Strengths
- Revenue growth from $30 million to $600 million annualized in roughly one year, per Sacra estimates, indicating strong product-market fit in the inference market.
- Multi-cloud GPU aggregation across AWS and Google Cloud provides customers with access to scarce Nvidia chips and geographic redundancy that single clouds cannot match.
- Nvidia's $150 million investment in the Series E round signals deep strategic alignment and likely preferential access to next-generation hardware and software optimizations.
Watch for
- Customer concentration risk: a handful of large AI-native clients (e.g., Notion, Cursor) could represent a significant share of revenue; churn by any one would materially impact growth.
- Intensifying competition from DigitalOcean's AI-Native Cloud, which reported $170 million in AI customer ARR in Q1 2026 and is raising its 2027 revenue growth outlook to over 50%.
- Valuation escalation: from $825 million in February 2025 to a potential $11 billion in May 2026 implies extreme multiples on current revenue, leaving little room for execution missteps or market slowdown.
Recent moves
Key Information
- Industry
- GPU Cloud / Infra
- Founded
- 2019
- Headquarters
- San Francisco
Frequently Asked Questions
What is Baseten and what does it do?
Baseten is a cloud infrastructure company that helps businesses deploy and scale AI models in production. It handles GPU access, autoscaling, and multi-cloud reliability across providers like AWS and Google Cloud, so customers can run models without managing hardware.
How much funding has Baseten raised and what is its valuation?
Baseten has raised $585 million in total funding, including a $300 million Series E in January 2026 at a $5 billion valuation led by IVP and CapitalG. As of May 2026, it is in talks to raise $1 billion at an $11 billion valuation.
What is Baseten's annual revenue and how fast is it growing?
Baseten hit $600 million in annualized revenue by March 2026, according to Sacra, representing roughly 1,900% year-over-year growth. This rapid increase from about $30 million the prior year indicates strong product-market fit in the AI inference market.
Who are Baseten's main customers and what industries do they serve?
Baseten's customers include AI-native startups like Cursor and HeyGen, enterprise SaaS companies like Notion, and healthcare firms like Abridge. These organizations use Baseten to deploy proprietary or open-source AI models with optimized inference infrastructure.
How does Baseten's multi-cloud GPU access benefit customers?
Baseten aggregates GPU capacity across AWS and Google Cloud, giving customers access to scarce Nvidia chips and geographic redundancy. This multi-cloud approach helps avoid single-provider bottlenecks and ensures high availability for production AI workloads.
What are the main risks facing Baseten's business?
Key risks include customer concentration, as a few large clients like Notion and Cursor could represent significant revenue. Competition from DigitalOcean's AI-Native Cloud, which reported $170 million in AI ARR, and extreme valuation multiples also pose challenges.
Sources
- sacra.com — Revenue estimate of $600M annualized, $5B Series E valuation, $11B potential valuation, total funding $585M, investor list, headcount 200
- www.cnbc.com — Series C $75M at $825M valuation, sixfold revenue growth, multi-cloud strategy, DeepSeek R1 support, customer cost savings of 40%
- www.baseten.co — Series D $150M led by BOND, founding story, investor list including Conviction, CapitalG, 01A, IVP, Spark, Greylock, Premji Invest
- research.contrary.com — Founding date September 2019, CEO Tuhin Srivastava, employee count 200, product description, market context
- www.baseten.co — Customer logos including Abridge, Clay, Cursor, Decagon, Descript, EliseAI, Gamma, HeyGen, Lovable, Notion, OpenEvidence, Parallel, Poolside, World Labs, Writer; product offerings Dedicated Deployments, Model APIs, Training, Frontier Gateway