EL7.AI
Macro EconomyMedium•30 October 2026•
4 min read

Baseten in talks at potential $26 billion valuation, 3 months after previous round

Key Facts

1Axios reported on September 22, 2026, interest in a round that could value Baseten at about $26 billion, 3 months after its previous round.
2Baseten raised $1.5 billion in June 2026 at a $13 billion valuation.
3Baseten said its revenue grew 20 times in the year before its June 2026 round.
4Baseten raised $300 million in January 2026 at a $5 billion valuation.

Axios reported on September 22, 2026, that Baseten had received interest in a funding round that could value it at about $26 billion, just 3 months after its previous round. The figure points to the possibility of the company’s value doubling, but was cited in the context of talks and no completion of the round was announced. The company raised $1.5 billion in June at a valuation of $13 billion. Baseten operates infrastructure that enables artificial intelligence models to run and deliver outputs in applications. The talks come as inference companies’ needs for capital, computing capacity and the chips required to serve customers are increasing.

Bloomberg said in September that inference platform companies need additional funding to secure the computing power and chips their services require. This reflects the nature of a business that depends on extensive operating resources to deliver models to customers. In June, Baseten said its revenue had grown 20 times in the year before the round, as demand for inference increased. The company defined this stage as the point at which trained models produce outputs inside real-world applications. The reported revenue growth provides context for expanding financing interest in the company, without by itself establishing the terms of a new round or its completion.

The June round was Baseten’s fourth capital raise in 18 months, Reuters reported. It was led by U.S. investors Sands Capital and Wellington Management. The company said it would use the funds to expand computing capacity, software and hiring; these uses relate to the June round and do not constitute an announcement about how potential funds from another round would be deployed. A Blackbird partner also described his firm’s increased stake in Baseten as “a signal of conviction.” He said the company’s expansion gives it an opportunity to compete with model providers at lower cost.

Baseten had previously announced in January 2026 that it raised $300 million in a Series E round at a $5 billion valuation. The round was led by IVP and CapitalG, with participation from NVIDIA and other investors, according to the company’s announcement. It said at the time that its inference volume had grown 100 times in the previous year. This earlier round illustrates the path of announced valuation increases ahead of the June round, from $5 billion in January to $13 billion in June. The $26 billion level, meanwhile, remained a potential valuation in September, cited as part of funding talks.

In September, Baseten announced a partnership with OpenAI that allows enterprise customers to use their existing OpenAI commitments to access open models served by Baseten through Codex or the Responses API. The company said its infrastructure operates across more than 90 compute clusters and in more than 20 clouds, and that it has relationships with more than 200 compute providers. The partnership expands the channels through which enterprise customers can use its infrastructure, while the computing figures disclosed by the company reflect the scope of its operations. On September 10, Baseten announced its acquisition of Blaxel, which provides isolated runtime environments, persistent storage and connectivity for running agents. It said inference should run close to computing, storage and networking, because distance between them can add latency in every operating cycle.

On October 23, 2026, Baseten announced the launch of a partner network to build custom reinforcement learning environments for customers, expanding its services beyond training and inference. The company said the network would build custom environments based on customers’ data and domains. After those environments are prepared, Baseten will train models and deploy them through its inference infrastructure. The announcement places the company’s latest disclosed expansion into post-training services in the context of the funding talks Axios reported in September. It did not confirm the completion of a new round or specify how its funds would be used.