Anthropic is reportedly leasing an Nscale data center in West Virginia for $45 billion

Anthropic is reportedly leasing an Nscale data center in West Virginia for $45 billion

Key points

  • Financial news wires reported on August 26 that Anthropic will lease an Nscale data center in West Virginia for $45 billion and receive 460 megawatts of power.
  • Nscale is expected to spend about $71 billion to develop the site, its Monarch Compute Campus in Mason County, West Virginia.
  • Anthropic and Nscale are both private. Nscale, backed by Nvidia (NVDA), is reportedly seeking a US IPO of up to $3 billion.

Anthropic will lease a data center in West Virginia from Nscale for $45 billion, financial news wires reported on August 26. Under the reported terms, Anthropic will receive 460 megawatts of power from the site, and Nscale is expected to spend about $71 billion to develop it.

The site is Nscale's Monarch Compute Campus in Mason County, West Virginia. Nscale acquired the 2,250-acre, gas-powered project earlier this year and has said it could hold more than 8 gigawatts of computing capacity. Microsoft (MSFT) was the campus's first anchor tenant and commissioned 1.35 gigawatts there in March.

The lease would add to a long run of compute deals for Anthropic, the maker of the Claude AI models. The company has committed to renting more than 10 gigawatts of servers, and Google (GOOGL) has agreed to backstop billions of dollars of its data center leases, according to reports. Anthropic's other recent deals include a $19 billion lease with TeraWulf (WULF), $10 billion with Volta and $9.1 billion with Riot Platforms (RIOT). Anthropic itself could file to go public within days.

Nscale is a Nvidia (NVDA)-backed AI cloud company that has grown quickly this year. It struck a deal worth up to $14 billion to supply Microsoft with about 200,000 Nvidia chips, agreed to buy software startup Anyscale for $1.65 billion, and is reportedly seeking a US initial public offering of up to $3 billion.

Is there a stock that benefits?

Not directly. Neither Anthropic nor Nscale is publicly traded, so there is no clean way to own either one. The listed name closest to the deal is Nvidia, whose chips fill Nscale's sites, and we have tracked the same buildout through the neocloud companies that rent GPUs to the AI labs. Nscale's planned IPO would give investors a way in later. The open question is whether all this announced demand turns into real usage, which several states have started to doubt.

As of midday on August 26, neither Anthropic nor Nscale had issued a formal announcement, and the terms came from breaking wire reports.

Sources

Frequently asked questions

What did Anthropic and Nscale reportedly agree to?

According to financial news wires on August 26, 2026, Anthropic plans to lease an Nscale data center in West Virginia for $45 billion and receive 460 megawatts of power, while Nscale is expected to spend about $71 billion to develop the site. Neither company had issued a formal announcement.

What is the Monarch Compute Campus?

It is Nscale's data center project in Mason County, West Virginia, a roughly 2,250-acre, gas-powered site that Nscale has said could hold more than 8 gigawatts of computing capacity. Microsoft (MSFT) was its first anchor tenant, commissioning 1.35 gigawatts in March 2026.

Is there a stock that benefits from the Anthropic-Nscale deal?

Not directly, because Anthropic and Nscale are both private. The closest listed name is Nvidia (NVDA), which supplies the chips in Nscale's data centers. Nscale is reportedly seeking a US IPO of up to $3 billion. This is general information, not investment advice.

Is the Anthropic-Nscale deal confirmed?

As of midday August 26, 2026, the terms came from breaking financial news wires and had not been confirmed by a formal announcement from Anthropic or Nscale.

More on NVDA and MSFT

Dennis Singleton
Dennis Singleton

Dennis Singleton has spent years following the markets, but what keeps his attention is how AI is built. He writes about the companies behind the technology, from semiconductor designers and advanced packaging to photonics, memory, networking, and the hardware powering modern AI. His approach starts with filings, earnings, and industry research, then translates the important details into clear, straightforward analysis without unnecessary hype.