Skip to content
TrustList
Funding news

PaleBlueDot AI raises a $200 million Series C led by ComputeCore for its GPU cloud and inference platform

Editorial

By TrustList Editorial

Palo Alto-based PaleBlueDot AI, which runs its own GPU clusters, a GPU marketplace and serverless inference, announced a $200 million Series C led by ComputeCore at a stated $3.2 billion valuation on 1 October 2026.

About PaleBlueDot AI raises a $200 million Series C led by ComputeCore for its GPU cloud and inference platform

PaleBlueDot AI raises a $200 million Series C led by ComputeCore for its GPU cloud and inference platform

1 October 2026 — PaleBlueDot AI, a Palo Alto company founded in 2024 that rents out AI computing capacity, announced on 1 October 2026 that it has completed a $200 million Series C. The company sells access to its own GPU clusters, runs a marketplace for other providers' GPUs and offers serverless inference, and it says customers in the United States and Japan account for more than half of its monthly revenue.

Not yet independently verified. The company's release exists only on a wire; no copy on its own site was found. The valuation, customer-contract and revenue figures are the company's own. The business is partly hardware: it owns the GPU clusters it rents out. We will update this when it can be confirmed, and remove this note.

The round

According to the company's release, the $200 million Series C was led by ComputeCore at a valuation of $3.2 billion, with B Capital, an existing shareholder, participating alongside other global investors. The round follows a $150 million Series B announced in January. PaleBlueDot says the money will fund additional computing capacity, giving customers more choice of locations and hardware.

What the company does

PaleBlueDot describes three businesses under one platform. It owns GPU clusters that it configures for customers with large or specialised workloads, and it says its B300 cluster in Japan has been recognised by NVIDIA as an Exemplar Cloud. Its marketplace lets customers rent capacity from other providers through the same account. Its serverless inference service runs models for customers without them managing machines. The company states that it has signed more than $5 billion in customer contracts; that figure has not been checked.

Why it matters for buyers

Specialist GPU clouds have become a real alternative to the large public clouds for training and inference, often with faster access to new chips and lower prices. They are also younger companies, heavily financed by investors, with concentrated customer bases and expensive hardware on their books. A large round improves a provider's capacity and runway, but it does not remove the questions any buyer should ask before moving a production workload.

What to check

  • Where your workloads would physically run, and which jurisdictions' rules apply to the data.
  • Whether capacity is guaranteed in a contract, and what happens if a marketplace provider withdraws.
  • Pricing basis for reserved, on-demand and serverless use, and the cost of moving data out.
  • Security certifications and who operates the hardware in each location.
  • Exit terms: how quickly you can move models and data to another provider.

Sources

Categories & features