Nscale, a data-center developer, has laid out a spending plan that underscores the enormous capital demands of the artificial-intelligence boom. In a filing with the U.S. Securities and Exchange Commission, the company said it expects to spend $10.0–10.5 billion on data-center and power infrastructure for Monarch's Building 1, plus another $21–22 billion on graphics processing units (GPUs) and other compute equipment tied to a services agreement with Anthropic, an AI model developer. The filing, cited by Reuters, also revealed a net loss of $1.69 billion for the nine months ended September 30.
The numbers highlight a fundamental tension in the AI infrastructure business: building capacity costs a fortune upfront, while revenue often arrives only gradually as that capacity comes online. Nscale reported preliminary nine-month revenue of $307.5 million—a fraction of the billions it plans to spend. That gap is why the story quickly becomes about funding.
What Nscale is building and why it costs so much
Data centers are not just warehouses for servers. They require massive land, power connections, cooling systems, and physical security. The $10–10.5 billion figure for Monarch's Building 1 covers those costs. The separate $21–22 billion is for the compute equipment itself—the GPUs that do the heavy lifting for AI workloads. GPUs are the most expensive component of modern AI data centers, and demand for them has driven prices up sharply.
Anthropic, the company behind the Claude AI assistant, is a major customer. Nscale said Anthropic exercised an option on September 29 to contract for additional compute at Building 2, a sign that demand is real. But contracted demand is not the same as paid revenue. The project's economics depend on how quickly Nscale can turn those contracts into working, revenue-generating capacity.
This is a classic timing problem. Developers pay for land, power, buildings, and chips upfront, while customers often pay over time as capacity ramps. So even if demand holds up, the cost of financing can make or break the project.
Funding is the swing factor
Nscale flagged a further $2.09 billion commitment from banking partners for 300-megawatt SIN02 projects in Portugal and Norway, but only under certain conditions. That means milestones and credit availability are key variables. If timelines slip or financing gets pricier, the pain is more likely to show up as a bigger financing bill than as an immediate drop in customer demand.
This dynamic is not unique to Nscale. The AI data-center boom has shifted risk into private credit's shadows, as private lenders take on more of the funding burden. And delays are common: Nebius recently cut its 2026 revenue outlook because of data-center delays pushing into 2027.
For Nscale, the size of the bill is staggering relative to its current revenue. The company's $1.69 billion net loss over nine months reflects heavy upfront spending and interest costs. Investors will be watching how Nscale plans to fund the remaining gap—whether through debt, equity, or customer prepayments.
What it means for investors
For everyday investors, the key takeaway is that AI infrastructure is a capital-intensive business with a long payback period. Even companies with strong customer demand can face financial strain if borrowing costs rise or lenders tighten conditions. The Federal Reserve's rate path will influence how expensive this kind of borrowing becomes.
Nscale's situation also illustrates why some investors are cautious about the AI trade. The potential is enormous, but so are the costs. Companies like Enel are adding data-center services to their growth plans, and Poste Italiane has launched a €600 million fund for data centers in Italy. But the economics depend heavily on financing terms.
For Nscale, the next few quarters will be telling. If it can secure the necessary funding and hit its construction milestones, the Anthropic deal could generate substantial revenue. If not, the financing bill could grow, and the loss could widen. Investors should watch for updates on funding commitments and project timelines.
In short, Nscale's plans are a bet that AI demand will keep growing. But the company's ability to finance that bet is just as important as the demand itself.
