Nscale Acquires Anyscale in $1.65 Billion Deal to Build Full-Stack AI Cloud

A London-based AI cloud company most people have never heard of just made one of the biggest infrastructure deals of the summer. Nscale announced on July 30, 2026, that it has agreed to acquire Anyscale—the company behind Ray, one of the most widely used open-source frameworks for scaling AI workloads—in a deal Bloomberg estimates at roughly $1.65 billion.

Neither company has disclosed the official purchase price, but the move signals something bigger than a single acquisition: the "neocloud" wave of AI-focused data center companies is racing to stop just renting out GPUs and start owning the entire software stack that sits on top of them.

Quick Summary & Key Takeaways

• The Deal: Nscale is acquiring Anyscale in a transaction valued at approximately $1.65 billion, according to Bloomberg.
• What Anyscale Brings: A commercial platform built on Ray, the open-source framework used to distribute AI training, fine-tuning, and inference across thousands of GPUs at once.
• The Strategic Shift: Nscale currently owns its own power generation, data centers, and compute—this deal adds the software orchestration layer, completing a fully vertically integrated AI cloud.
• 200 Employees Move Over: Anyscale's roughly 200 staff across the US, Europe, and India will join Nscale, continuing to serve existing customers including Coinbase, Runway, and Bedrock Robotics.
• Deal Expected to Close: The transaction is subject to regulatory approval and is expected to close in the second half of 2026.

Deal at a Glance

Detail	Information
Acquirer	Nscale (London-based AI cloud company)
Target	Anyscale (commercial steward of the open-source Ray framework)
Reported Value	~$1.65 billion (Bloomberg estimate; official terms undisclosed)
Employees Affected	~200 staff across US, Europe, and India move to Nscale
Financial Advisors	Goldman Sachs International (lead) and Morgan Stanley for Nscale
Expected Close	Second half of 2026, pending regulatory approval

What Happened? Why Nscale Is Buying Its Way Up the Stack

Nscale grew out of a cryptocurrency-mining business in early 2024 and has since repositioned itself as one of the "neoclouds"—a wave of AI-focused data center newcomers competing to power the AI boom. Unlike many rivals that simply lease GPU capacity from elsewhere, Nscale owns its own power generation, data centers, and compute infrastructure, including a flagship 2,250-acre campus in Mason County, West Virginia capable of theoretically hosting more than eight gigawatts of computing capacity.

What Nscale lacked was the software layer that decides how efficiently that hardware actually gets used—and that's exactly what Anyscale provides. Anyscale's platform, built on Ray, spreads data preparation, training, fine-tuning, and inference across an entire fleet of machines, with the company claiming it can cut total cost of ownership by up to 90% compared to stitching together a patchwork of separate tools.

Nscale CEO and founder Josh Payne framed the acquisition as a direct extension of the company's core philosophy, explaining that while most infrastructure providers simply buy GPUs and rent them out, Nscale builds and owns every layer itself—the power, the data centers, the compute, and now the software that turns all of it into a usable AI cloud. Anyscale CEO Keerti Melkote echoed that framing, noting that as enterprises shift from merely consuming AI services to building their own proprietary AI systems, there's growing demand for infrastructure and software designed to work together as a single unit rather than separate purchases.

What Happens to Ray, the Open-Source Framework?

One detail worth noting for developers: Ray itself isn't being swallowed up as a closed product. The framework was donated to the PyTorch Foundation in 2025 and remains open-source and community-governed. As part of the deal, Nscale has committed to joining the PyTorch Foundation itself, a move meant to reassure the tens of thousands of developers who rely on Ray that its open governance won't change just because its commercial steward has new owners.

Why It Matters: The Neocloud Land Grab Is Accelerating

This acquisition fits into a broader pattern reshaping the AI infrastructure market in 2026:

• From Renting to Owning: The deal marks Nscale's shift from simply renting out raw compute to controlling the full pipeline—power, hardware, and now the software that orchestrates workloads across it.
• Vertical Integration Is the New Battleground: As AI companies increasingly want a single vendor for training, fine-tuning, inference, and agent deployment, infrastructure providers are racing to become genuine one-stop shops rather than piece-part suppliers.
• Funding Signals Confidence: Nscale raised $2 billion in March 2026 from a consortium that included Nvidia, giving it the capital firepower to make a deal of this size just months later.
• Existing Customers Come Along: Anyscale's client base—including Coinbase, Runway, and Bedrock Robotics—gives Nscale an instant foothold with companies already running serious production AI workloads.

💡 AI Tech Safar Insight

Deals like this rarely make front-page headlines the way a new chatbot launch does, but they matter just as much to where the AI industry is heading. As the "who has the best model" race gets increasingly crowded and commoditized, a quieter battle is happening one layer down—over who controls the infrastructure and orchestration software that makes any of these models actually usable at scale. Nscale betting $1.65 billion on owning that layer, rather than just renting out chips, suggests infrastructure providers see software orchestration—not raw GPU access—as the next real differentiator in the AI cloud market.

Frequently Asked Questions (FAQs)

Q1: How much is Nscale paying for Anyscale?
The official terms weren't disclosed by either company, but Bloomberg reported the deal is valued at approximately $1.65 billion, citing a source familiar with the matter.

Q2: What does Anyscale actually do?
Anyscale is the commercial platform built by the creators of Ray, an open-source framework used to distribute AI training, fine-tuning, and inference workloads across large fleets of GPUs.

Q3: Will Ray remain open source after the acquisition?
Yes. Ray was donated to the PyTorch Foundation in 2025 and remains open-source and community-governed. Nscale has committed to joining the PyTorch Foundation as part of the deal.

Q4: When will the acquisition be finalized?
The transaction is expected to close in the second half of 2026, pending standard closing conditions and regulatory approval.

What Do You Think?
Is vertical integration—owning power, compute, and software all in one company—the winning strategy for AI cloud providers, or does specialization still have an edge? Share your thoughts in the comments below!

Source: Reporting based on Nscale's official press release and SiliconANGLE.