Baseten, a San Francisco–based AI inference company supporting AI applications, raised $300 million in a funding round led by IVP, CapitalG, and NVIDIA.
The company is valued at a $5 billion.
This funding round increases the total capital raised to $585 million.
Read More:Capital One Acquires Payments Startup Brex
Baseten’s third fundraiser in the past year reflects surging demand for high-performance infrastructure to run modern AI models in production reliably.
“In a world where every ambitious AI team wants to run many models and fully own its IP, Baseten gives them the freedom, reliability, and economics to do that at scale.”
After six years of building, Baseten has become the inference platform powering many AI products, reshaping how people work and build software, including Cursor, OpenEvidence, Abridge, Notion, and Clay.
“Baseten lets us run the models we need, the way we need to run them,” says Shiv Rao, Co-founder & CEO of Abridge. “The performance is best in class, but what sets them apart is everything else, the reliability, the developer experience, the fact that they are constantly finding ways to lower our costs. They’re a partner, not a vendor.”
The AI industry is entering a new phase at breakneck speed. After years of focusing on training ever-larger models, attention has shifted to inference—running AI models that reason and generate outputs in real time within real workflows. Analysts estimate inference will make up two-thirds of all AI compute by the end of 2026, up from one-third in 2023.
Baseten is built for this shift, enabling customers to run more of their models reliably in production.
“If cloud computing laid the foundation for the last generation of great technology companies, inference will underpin the next,” said Tuhin Srivastava, co-founder and CEO of Baseten.
“Every breakout AI application depends on fast, reliable, and cost-effective inference. We’ve spent six years building the infrastructure to make that possible, and we’re ready for the next chapter—supporting hundreds, and soon thousands, of models.”
“Baseten is rapidly becoming default infrastructure,” said Sarah Guo, General Partner at Conviction. “As more AI teams aim to run multiple models and fully own their IP, Baseten gives them the freedom, reliability, and economics to scale. With open runtimes, multi-cloud resilience, and a thoughtfully designed developer experience, Baseten is setting the new standard that top companies expect.”
Baseten believes the future of AI is multi-model. As organisations build custom and domain-specific models, they need an independent inference layer with strong security, guardrails, and observability, running on infrastructure they control. Baseten enables companies to own their differentiation and IP through fully open runtimes, with no model lock-in and multi-cloud flexibility delivering the performance, reliability, and developer experience that fast-growing AI companies standardise on as they scale.
About Baseten
Founded in 2019 and based in San Francisco, Baseten is the inference platform behind a new generation of AI products. The company builds systems software that runs the entire AI workload from GPUs and autoscaling to observability, billing, and developer tools so teams can focus on building great models and user experiences instead of managing infrastructure. Baseten partners with leading AI companies including Cursor, Mercor, Clay, OpenEvidence, Lovable, and Abridge.
Read More:Mendra Raises $82M in Series A Funding
