Baseten, the AI inference company powering a new wave of category-defining AI applications, has announced a $300 million financing round anchored by IVP, CapitalG, and NVIDIA. The funding values Baseten at $5 billion and marks the company’s third fundraise in the past year.

The investment reflects accelerating demand for high-performance infrastructure capable of reliably running modern AI models in production. After six years of focused development, Baseten has become the inference platform behind many rapidly growing AI products.

Companies relying on Baseten include Cursor, OpenEvidence, Abridge, Notion, Clay, and others building production-grade AI systems.

“Baseten lets us run the models we need, the way we need to run them. The performance is best-in-class, but what sets them apart is the reliability, the developer experience, and the constant focus on lowering costs.”

— Shiv Rao, Co-founder & CEO, Abridge

Inference as the Defining AI Story of 2026

The AI industry is rapidly entering a new phase. After years centered on training ever-larger models, focus has shifted to inference—the ability for AI systems to reason and generate outputs in real time within production workflows.

Industry analysts estimate that inference will account for nearly two-thirds of all AI compute by the end of 2026, up from one-third in 2023. Baseten’s platform is purpose-built to support this transition, enabling customers to run more models reliably in production.

“If cloud was the foundation that enabled the last generation of great technology companies, inference is the foundation for the next.”

— Tuhin Srivastava, Co-founder & CEO, Baseten

Srivastava emphasized that every breakout AI application depends on fast, reliable, and cost-effective inference, and that Baseten has spent years building the infrastructure needed to support hundreds—and eventually thousands—of production models.

The funding reinforces Baseten’s belief that the future of AI will not be dominated by a small number of giant models, but by thousands of specialized models built for specific industries and use cases.

Powering a Multi-Model AI Future

Baseten envisions a multi-model ecosystem where organizations build and deploy custom, domain-specific AI models. To support this future, companies need an independent inference layer that offers strong guardrails, security, observability, and control over infrastructure.

The platform enables customers to fully own their intellectual property. Its runtimes are open, not tied to proprietary model weights, and the infrastructure supports multi-cloud deployment to optimize for both reliability and cost.

“Baseten is quickly becoming default infrastructure. In a world where AI teams want to run many models and fully own their IP, Baseten gives them the freedom, reliability, and economics to do that at scale.”

— Sarah Guo, General Partner, Conviction