Baseten Raises $300 Million Series E at $5 Billion Valuation as Nvidia Invests $150 Million in AI Inference Infrastructure

Baseten, an AI inference infrastructure startup powering deployment for leading AI applications, has raised $300 million in Series E funding at a $5 billion valuation, more than doubling its worth from just four months earlier. The round was led by venture capital firm IVP and Alphabet's CapitalG, with chipmaker Nvidia contributing $150 million as a strategic investor, according to the Wall Street Journal.

The rapid valuation increase from September's $2.15 billion Series D reflects surging demand for AI inference infrastructure as companies race to deploy large language models and other AI capabilities at scale. Baseten's platform converts machine learning models into production-ready APIs, functioning as AWS Lambda for AI workloads and enabling companies to move from experimentation to deployed products without managing complex GPU orchestration.

Nvidia's $150 million investment represents half the total funding round and continues the chipmaker's strategy of taking equity positions in companies heavily dependent on its GPU computing hardware. The investment ensures Nvidia maintains proximity to demand drivers while supporting the infrastructure ecosystem generating revenue for its core semiconductor business.

Baseten already provides support for Nvidia's Nemotron 3 Nano model and recently announced text-to-video inferencing services across the United States, Finland, and France through partnerships leveraging Nvidia's computing infrastructure. The relationship exemplifies Nvidia's approach of vertical integration through strategic investments rather than direct competition with infrastructure providers.

The funding arrives as AI inference emerges as a critical bottleneck in the industry. While model training garners headlines and massive capital expenditure, inference represents the ongoing process of running trained AI models to generate predictions, decisions, and outputs for end users. Unlike one-time training costs, inference expenses scale directly with application usage, creating recurring revenue tied to AI adoption.

Baseten's customer roster demonstrates the platform's role in mission-critical deployments. Medical information company OpenEvidence uses Baseten to serve healthcare providers across major facilities nationwide, processing billions of custom fine-tuned language model calls weekly. The platform enables OpenEvidence to deliver medical information to physicians with sub-160 millisecond latency, reducing infrastructure development time by a factor of eight.

Other prominent customers include Abridge, which transforms medical conversations into clinical documentation for tens of thousands of clinicians generating over one million notes weekly, and Clay, an AI-powered go-to-market platform. These deployments highlight Baseten's focus on applications requiring exceptional reliability, performance, and cost efficiency at significant scale.

CEO Tuhin Srivastava positions Baseten as foundational infrastructure for the AI economy, analogous to how cloud computing enabled the previous generation of technology companies. "As model-driven products become ubiquitous, we will be the invisible infrastructure behind the AI-first economy," Srivastava stated in announcing the funding.

Baseten's technical approach leverages a multi-cloud capacity management system spanning over ten cloud providers rather than owning GPU infrastructure directly. This asset-light model enables dynamic workload allocation based on GPU availability and pricing, addressing supply constraints in high-end AI chips while avoiding vendor lock-in risks common with single-provider strategies.

The platform's open-source Truss framework, with over 6,000 GitHub stars, enables developers to package models into auto-scaling HTTPS endpoints managing GPU orchestration, caching, and monitoring automatically. Baseten recently expanded beyond inference to include training capabilities for multi-node fine-tuning jobs and embedding inference for retrieval-augmented generation workloads, covering the full machine learning lifecycle.

The company's revenue has grown more than tenfold over the past twelve months according to previous disclosures, though specific financial figures remain private. Total funding now exceeds $585 million across six rounds since the company's 2019 founding.

For the broader AI infrastructure landscape, Baseten's rapid valuation appreciation signals continued investor confidence in picks-and-shovels businesses serving AI application developers rather than foundation model builders, as inference workloads scale with adoption regardless of which models ultimately dominate the market.