AI 인프라 기업 Baseten이 300억 달러 규모의 기업가치를 목표로 하고 있다.

A new funding round could lift the valuation of AI infrastructure startup Baseten to about $30 billion, reflecting growing investor appetite for companies that supply the tools and computing capacity needed to run artificial intelligence models as their use spreads across institutions and corporations.

According to a Bloomberg report published on October 30, 2026, the company aims to raise about $2 billion in a round split into two tranches, each tied to a different valuation. The first tranche is expected to value Baseten at $26 billion, and the second would lift the valuation to $30 billion.

That would be a major jump from the company's last disclosed valuation of $13 billion in June, when it raised $1.5 billion. If the round closes at the valuations being cited, the company's value will have more than doubled in no more than four months.

According to Bloomberg, the round will be carried out in two stages: the first at a $26 billion valuation, followed by a second tranche at $30 billion. The round is expected to be led by Sands Capital, which is among Baseten's existing investors.

The June 2026 round, in which the company raised $1.5 billion at a $13 billion valuation, was led by Sands Capital and Wellington Management, according to Reuters.

The gap between the two valuations reflects the scale of growth the company has achieved in the AI infrastructure market. It also signals investors' readiness to commit large sums to companies working on one of the main challenges to the spread of AI applications: providing the computing capacity needed to run models efficiently and affordably.

However, the figures for the new round remain estimates from media reports citing people familiar with the discussions. They do not mean the funding has been closed or the valuation finalized, pending an official announcement from the company.

Founded in 2019, Baseten provides software and infrastructure that help companies deploy, run and manage AI models without having to build a complete compute stack from scratch.

Its platform allows enterprises to run multiple AI models in a single environment, along with the tools and technical capabilities needed to handle operational requirements. It targets companies that want to integrate AI into their products and services but do not necessarily have specialized infrastructure teams or the computing resources such operations require.

The importance of this activity stems from the fact that building an AI model is only one stage of service delivery. After training a model, companies need to run it continuously so that it can receive and process user requests and deliver results. This is where Baseten comes in, providing a runtime layer that helps customers use models and manage inference operations, the process through which a model produces answers or outputs based on the inputs it receives.

Baseten's growth comes as the technology and investment sectors increasingly focus on the inference market, or AI inference, which refers to running AI models after their training is complete.

This stage differs from model training. Training requires massive computing resources to process data and build model capabilities, whereas inference is tied to running the model when it is actually used, whether to answer questions, analyze data, execute tasks or power intelligent applications.

As AI applications and autonomous agents see wider use, the need to run models continuously is growing, pushing up demand for processing units, cloud resources and software that ensure fast response times and efficient operations.

Industry estimates suggest that spending on computing chips and infrastructure dedicated to inference could eventually overtake spending on model training, as AI moves from the model-development stage to widespread use inside companies. That shift is an opportunity for companies such as Baseten, which focuses on providing the tools needed to run models, particularly as enterprises look for ways to lower usage costs and gain more flexibility in choosing which models to rely on.

Baseten is not alone in this field. It competes with a group of companies seeking to provide more efficient infrastructure for running AI models. Among the most prominent competitors is Fireworks AI, which raised $1.5 billion last July at a valuation of $17.5 billion, according to information published about its funding round.

These deals reflect the scale of investor interest in the inference market, as companies compete to offer solutions that help developers and enterprises run models faster and at lower cost.

These companies also benefit from the wider use of open-source models, which give enterprises options beyond total reliance on the closed commercial models offered by companies such as OpenAI and Anthropic. As open-source models improve, some companies have become more interested in running custom models tailored to their needs while keeping greater control over data, costs and the technology stack they use.

Inference platforms provide the link between these models and the applications that use them, supplying the runtime environment and computing capacity needed to deliver the service to users.