PaleBlueDot AI raises $200M to add computing capacity

PaleBlueDot AI supplies the computing power developers and enterprises use to build and run AI systems. Customers can use company-operated GPU clusters, draw on marketplace capacity from outside suppliers or send inference work to a serverless service. The Palo Alto startup has closed a $200 million Series C led by ComputeCore at a $3.2 billion valuation.

For large or specialized workloads, PaleBlueDot says it owns and operates dedicated clusters customized for customers. Marketplace users can tap computing capacity made available by suppliers around the world. The serverless product provides a third route for developers and enterprises that need inference computing.

PaleBlueDot said it had signed more than $5 billion in customer contracts by the end of September. That figure describes agreements, not a disclosed revenue total. The company separately said customers in the United States and Japan accounted for more than half of monthly revenue, but it did not report the revenue amount, contract duration or delivery schedule.

The company said the Series C will fund additional computing capacity, giving customers more choice of locations and hardware. It did not say whether that expansion will come through company-operated infrastructure, marketplace suppliers or some combination.

The financing follows a $150 million Series B announced in January. B Capital, an existing shareholder, also participated in the Series C with other investors. PaleBlueDot says it was founded in 2024.

PaleBlueDot cited its B300 cluster in Japan as one example of the infrastructure it operates itself. The cluster earned Nvidia Exemplar Cloud status after testing against Nvidia’s reference benchmarks, according to the announcement. CEO Stephen Watts said the company will focus on frontier labs, Neolabs and enterprises in the United States while investing in its full-stack and go-to-market teams.