The AI race is no longer just about building better models. It is also about securing the enormous computing power needed to train them, run them, and serve millions of users.
That is becoming increasingly clear as Nscale, an AI cloud computing company backed by Nvidia, enters talks with Anthropic for a potential $45 billion computing deal.
According to a Bloomberg report published on October 20, 2026, the proposed agreement would involve a second facility at Nscale’s Monarch data center campus in West Virginia. The deal could be finalized within days, although negotiations are ongoing and no final agreement has been announced.
If completed, it would add to Anthropic’s growing list of multibillion-dollar infrastructure commitments and give Nscale another major customer agreement as it prepares for a potential US stock market listing.
But there is a bigger story here. The deal highlights how the AI boom is creating enormous opportunities for companies that provide computing infrastructure, while raising important questions about the cost of building and sustaining this industry.
Let’s break down what is happening and why investors should pay attention.
1. What Is the $45 Billion Deal Actually About?
At its core, the proposed agreement is about computing capacity.
Anthropic, the company behind the Claude family of AI models, needs access to powerful computing infrastructure to support its products and expand its AI capabilities. Rather than building every data center itself, the company is securing capacity through infrastructure providers such as Nscale.
Under the proposed arrangement, Nscale would develop a second building at its Monarch campus in West Virginia for Anthropic.
Here are the key details:
- Potential deal value: Approximately $45 billion.
- Infrastructure provider: Nscale Ltd., an AI cloud computing company backed by Nvidia.
- Customer: Anthropic, the company behind Claude.
- Location: Nscale’s Monarch data center campus in West Virginia.
- Expected computing capacity: 460 megawatts for the proposed second building.
- Hardware: Nvidia’s next-generation Vera Rubin AI chips.
- Current status: Negotiations are ongoing, with no final agreement announced.
The potential deal is particularly significant because Anthropic already has infrastructure agreements with Nscale worth as much as approximately $44.6 billion, according to Nscale’s regulatory filing for its planned initial public offering.
The new proposal would expand an existing business relationship rather than establish an entirely new one.
Anthropic also reportedly holds a right of first refusal on the campus’s expected third building, potentially giving it an opportunity to secure additional capacity in the future.
The bigger takeaway: Access to computing power is becoming a major strategic priority for AI companies. Securing infrastructure today can help them prepare for future demand, but it also creates substantial financial commitments.
2. Why Does Anthropic Need So Much Computing Power?
To understand the scale of this deal, it helps to understand what happens behind the scenes when people use AI.
Building and operating advanced AI models requires powerful processors, enormous amounts of data, specialized software, and data centers capable of running intensive workloads around the clock.
And the demand does not stop once a model has been trained.
Every time someone uses Claude to write, analyze information, generate code, or solve a problem, computing resources are needed to produce the response.
As more individuals and businesses adopt AI tools, the infrastructure required to support those services can grow substantially.
Anthropic is competing in a market where the ability to deliver more capable models and serve more customers can become a competitive advantage. That makes reliable access to computing resources increasingly important.
There are three major reasons why the company is pursuing large infrastructure commitments.
A. Supporting growing demand
As AI adoption expands, companies need enough computing capacity to serve customers without running into infrastructure constraints.
Additional data center capacity can help Anthropic support more users and workloads.
B. Developing more advanced AI models
Training and improving sophisticated AI models requires substantial computing resources. Access to advanced chips and large-scale infrastructure can support this work.
C. Planning for the future
Building data centers takes time. Securing capacity in advance allows AI companies to plan for future needs rather than waiting until existing infrastructure becomes insufficient.
However, there is an important distinction between securing computing capacity and making money from it.
More computing power creates opportunities for growth, but it does not automatically translate into higher profits. Anthropic still needs sufficient revenue and efficient operations to justify the cost of its infrastructure commitments.
That is one of the central questions investors should keep in mind as AI companies continue to sign increasingly large deals.
3. Why Nscale Could Be a Major Beneficiary
While Anthropic is securing the infrastructure it needs, Nscale is positioning itself as a provider of the computing capacity that AI companies increasingly require.
The proposed agreement could be valuable for Nscale for several reasons.
A. A major customer relationship
Anthropic is already one of Nscale’s largest customers. A second major agreement would deepen that relationship and potentially provide greater visibility into future demand for Nscale’s infrastructure.
For a company investing heavily in data centers, large customer commitments can be important when planning construction, financing, and hardware procurement.
B. Strengthening its IPO story
Nscale is preparing for a potential US stock market listing and is seeking a valuation of approximately $30 billion.
Another multibillion-dollar agreement with Anthropic could strengthen its pitch to prospective investors by demonstrating demand for its infrastructure.
However, a large contract alone does not determine what a company is worth.
Investors will also want to understand how much revenue Nscale can recognize, the cost of delivering the contracted computing capacity, its financing obligations, and the profitability of its operations.
C. Attracting investor interest
Nscale raised $3.36 billion in a pre-IPO financing round last month. The round was led by Third Point and included participation from Nvidia, Apollo Global Management, Citadel, Hudson Bay Capital, and the Abu Dhabi Investment Council, according to the report.
The participation of major investment firms and Nvidia reflects investor interest in AI infrastructure.
Still, investors have reportedly encouraged Nscale to lower its targeted IPO valuation. This suggests that enthusiasm for the AI infrastructure opportunity does not eliminate concerns about pricing.
A major deal may improve Nscale’s growth prospects, but prospective shareholders will still need to assess whether its expected earnings and cash flows justify its valuation.
For investors, the distinction is important: a large contract can strengthen a company’s business outlook without automatically making its stock a good investment.
4. Nvidia’s Role Goes Beyond Selling Chips
Nvidia is another important part of this story.
The company supplies advanced processors used to train and operate AI models, while its investment in Nscale gives it exposure to another part of the AI infrastructure ecosystem.
The proposed Monarch facility is expected to use Nvidia’s Vera Rubin AI chips, according to people familiar with the discussions.
This creates an interesting connection between the companies involved.
- Anthropic needs computing power to develop and operate its AI models.
- Nscale provides the infrastructure needed to deliver that computing capacity.
- Nvidia supplies the advanced chips expected to power the proposed facility and has backed Nscale.
Each company plays a different role, but their business interests are connected.
As demand for AI computing grows, infrastructure providers may purchase more advanced chips. At the same time, AI developers need access to the data centers in which those chips operate.
This helps explain why the AI investment story extends beyond the companies developing chatbots and other AI applications.
The opportunity also reaches semiconductor manufacturers, data center developers, cloud computing providers, and businesses supplying the power and equipment needed to operate these facilities.
For Nvidia, infrastructure expansion can support demand for its products. But the eventual financial benefit depends on whether these projects move forward, the chips are delivered, and customers generate enough business to sustain their commitments.
5. The AI Infrastructure Boom Is Becoming a Capital-Intensive Business
The potential Nscale-Anthropic deal is part of a much larger trend.
Anthropic has reportedly entered into several other multibillion-dollar infrastructure agreements, including:
- A $50 billion agreement with Fluidstack.
- A $10 billion agreement with Volta Infrastructure Holdings.
- A $45 billion agreement with SpaceX, according to the report.
These agreements illustrate the scale of infrastructure commitments being made to support AI development.
They also raise an important question: Who will finance the infrastructure required to keep the AI boom growing?
Data centers are expensive to build and operate. They require land, specialized equipment, advanced chips, electricity, cooling systems, and reliable network connections.
These costs often need to be financed well before the infrastructure generates its full economic return.
That creates a business model with both significant opportunities and considerable risks.
On one hand, long-term customer commitments can give infrastructure providers greater confidence when investing in new facilities.
On the other hand, the companies building those facilities may take on substantial debt or other financial obligations. If customer demand grows more slowly than expected, or if financing becomes more expensive, the economics of those investments could weaken.
The same challenge applies to AI developers.
Signing large infrastructure contracts may help secure the resources needed to grow, but those commitments also create future financial obligations.
The question is not simply whether AI demand will increase. It is whether the revenue generated by AI products will be sufficient to cover the cost of the infrastructure supporting them.
That is where the distinction between an exciting growth story and a sustainable business becomes especially important.
6. Why Investors Are Becoming More Cautious
Despite the excitement surrounding AI, investors are increasingly examining the financial risks associated with the industry’s infrastructure expansion.
The concern is not necessarily that demand for AI will disappear.
Instead, the issue is whether companies are spending too much, borrowing too heavily, or committing to infrastructure that may take longer than expected to generate adequate returns.
Several risks deserve attention.
A. High capital expenditure
Building data centers and purchasing advanced chips require significant upfront investment.
If infrastructure costs rise faster than revenue, profit margins can come under pressure.
B. Financing and debt exposure
Some infrastructure projects depend on borrowing or complex financing arrangements.
Higher interest costs, tighter credit conditions, or weaker investor appetite could make it more difficult to finance future expansion.
C. Uncertain returns
A facility can be technically impressive and still deliver disappointing financial results if its operating costs are too high or demand does not meet expectations.
D. Valuation risk
Companies associated with AI infrastructure may attract substantial investor interest because of their expected growth.
However, if valuations already assume years of strong expansion, even a successful business may disappoint investors if its actual performance falls short of expectations.
These risks are particularly relevant for companies preparing to enter the public markets.
Nscale’s proposed IPO valuation of around $30 billion will need to be evaluated against its expected revenue, margins, capital requirements, and long-term financial prospects.
The same principle applies to Anthropic if it pursues a public listing of its own.
AI infrastructure may be a major growth opportunity, but investors still need to understand how the spending translates into sustainable cash flow.
7. Why West Virginia Matters to the Story
The proposed deal also highlights the growing importance of physical infrastructure in the AI economy.
Nscale’s Monarch campus in West Virginia is designed for substantial expansion. According to the company’s IPO filing documents, the site could eventually support more than 8 gigawatts of gross power capacity when fully developed.
For perspective, the proposed second building alone would deliver 460 megawatts of computing power.
These figures demonstrate the scale of the infrastructure being planned to support future AI workloads.
But building a facility of this size involves more than installing processors.
Developers must secure adequate power, establish cooling systems, build network connections, obtain necessary approvals, and coordinate the delivery of specialized equipment.
Electricity availability is particularly important because large AI data centers can require enormous amounts of power to operate reliably.
The development also illustrates how the AI boom is creating investment opportunities outside traditional technology hubs.
As demand for computing capacity expands, the economic impact can reach local construction companies, utilities, energy providers, equipment suppliers, and other businesses involved in developing and operating data centers.
However, the benefits depend on whether proposed projects are completed and operate as planned.
It is also worth noting that Microsoft had previously considered leasing part of the West Virginia site but walked away from the location during the summer, according to Bloomberg’s reporting. Anthropic subsequently emerged as the tenant.
That development shows how quickly demand, customer relationships, and infrastructure plans can change in a rapidly evolving industry.
8. What About Anthropic’s Potential IPO?
Anthropic is also attracting attention because of expectations surrounding a potential public listing.
The company was valued at $965 billion in May, according to the report, and is expected to pursue a potentially record-breaking IPO in the coming months.
That valuation is extraordinary, making the company’s ability to turn AI adoption into a sustainable business especially important for prospective investors.
A company can have a widely used product, strong technology, and significant customer demand while still facing questions about its long-term profitability.
For Anthropic, investors will likely want to examine several factors:
- Revenue growth: Is demand for Claude translating into consistent revenue growth?
- Customer economics: Can the company serve more customers without allowing computing costs to rise too quickly?
- Infrastructure spending: How much capital is required to support its expansion?
- Competition: Can Anthropic maintain its position as other AI companies develop competing models and products?
- Profitability: Can revenue eventually cover computing costs, research expenses, and other operating expenses?
The potential Nscale agreement adds another dimension to this discussion.
It demonstrates the scale of infrastructure Anthropic is seeking to secure, but it does not establish whether the company will generate sufficient returns on that investment.
For investors evaluating a possible IPO, that distinction matters.
A large infrastructure commitment can signal ambition and anticipated demand. It can also create financial obligations that need to be justified by future business performance. Both sides belong in the investment analysis.
9. What Should Investors Watch Next?
The next developments could provide a clearer picture of how this deal fits into the wider AI investment story.
Here are five things worth watching.
1. Will the agreement be finalized?
The reported $45 billion figure relates to ongoing negotiations, not a completed transaction. Investors should wait for confirmation of the final terms before treating the deal as certain.
2. What will the contract actually deliver?
The headline value is only one part of the story. The duration of the agreement, delivery schedule, payment structure, and conditions attached to the contract will help determine its economic significance.
3. How will Nscale finance its expansion?
Building additional infrastructure requires substantial capital. Investors should examine how much funding is needed and whether Nscale can manage those obligations without placing excessive pressure on its finances.
4. What will happen with Nscale’s IPO?
The company’s progress toward a public listing, its proposed valuation, and the financial information disclosed in its offering documents could help investors assess how the market values AI infrastructure businesses.
5. Can Anthropic turn AI demand into sustainable profits?
Ultimately, Anthropic’s infrastructure spending needs to support a business that can generate sufficient revenue and, over time, attractive returns.
Strong demand for AI is important, but it is not a substitute for financial performance.
These developments will help investors distinguish between companies benefiting from the expansion of AI and companies that can convert that expansion into durable shareholder value.
The Bigger Picture: AI’s Next Battle Is About Infrastructure
The proposed $45 billion agreement between Nscale and Anthropic offers a glimpse into how the AI industry is evolving.
The competition is no longer limited to developing the most capable models or attracting the largest number of users. It increasingly involves securing the computing resources, energy, and physical infrastructure needed to support those models at scale.
For Anthropic, the potential agreement could help secure additional capacity as demand for its AI products grows.
For Nscale, it could strengthen its position as an infrastructure provider and support its efforts to enter the public markets.
For Nvidia, it highlights the interconnected nature of the AI ecosystem, where demand for advanced chips is closely linked to investment in data centers and AI services.
But the deal also captures the central tension in the current AI boom.
The opportunity is enormous, but so is the cost of pursuing it.
Investors should look beyond headline deal values and ambitious valuations to understand who is paying for the infrastructure, who is taking on the financial risk, and who is ultimately positioned to earn a return.
Because the companies that benefit most from the AI boom may not necessarily be the ones making the biggest promises. They may be the ones that can turn massive investments into lasting, profitable businesses.
What do you think? Is the AI infrastructure boom creating a sustainable investment opportunity, or are companies committing too much capital before the returns are clear?