NSCALE: the AI cloud race also hinges on access to capital

With roughly $103 billion in contracts presented to investors, Nscale must raise the capital needed to deliver the promised compute. Its search for funding ahead of a possible initial public offering highlights a condition for competing in AI cloud: winning orders requires the means to fulfil them. Private capital, debt and Nvidia’s backing are helping to build that capacity.

Contracts worth $103 billion, and still more capital to raise. According to Bloomberg, Nscale is in talks to secure up to $3.5 billion in financing ahead of its initial public offering. The London-based company is reportedly seeking to sell up to $1.5 billion in convertible notes to investors and obtain around $2 billion in financing from Nvidia. Discussions remain ongoing.

Nscale provides compute to train and run AI models. It brings together power, data centres, GPUs and software services at facilities it owns or operates with partners. Its expansion depends on assembling these resources in step with its commitments to customers.

Contracts need financing before they can be fulfilled

The contracts span several years, while some expenditure must be incurred before services begin. A growing order book can therefore increase financing needs well before the corresponding customer payments arrive.

Access to capital consequently influences an operator’s ability to take on new orders. Combining several sources of financing can give it greater flexibility to expand its offering.

The Anthropic contract, valued at $45 billion according to Bloomberg, would alone account for nearly 44% of the reported total contract value, assuming comparable scope. While a commitment of this size can help secure financing for several deployments, it also concentrates a substantial share of commercial risk in a single customer, making payment guarantees and delivery terms critical.

Nscale combines equity, debt and Nvidia’s backing

To support this growth, Nscale is drawing on several forms of capital. In March, the company raised $2 billion in a Series C round led by Aker and 8090 Industries, with Nvidia among the participants. This equity supports the group’s expansion, while debt financing funds its equipment and deployments.

In February, the company announced a $1.4 billion delayed-draw loan secured by GPUs. In late August, it announced around $3 billion in secured loan commitments for Ward County, Texas, and Madison, North Carolina. A $900 million revolving credit facility agreed in July provides additional liquidity flexibility. These figures combine capital raised with credit available subject to conditions; adding them together does not reveal Nscale’s cash position.

Nvidia’s role creates a direct link between technology and financing. Already an investor in Nscale, the chip supplier could increase its support if the talks result in an agreement. Private capital, loans and the planned IPO are all potential sources of funding for the investments needed to expand the operator’s business.

Deployment timelines become a financing constraint

The partnership with Figure illustrates how long financing must support a project. Announced in early September, it involves an initial commitment of $3.5 billion in compute, with the intention of expanding beyond $6 billion. Initial deployments are targeted for the second half of 2027 in Barstow, Texas, under an arrangement that could eventually encompass up to 100,000 Nvidia GPUs.

From contract signing to delivery, funding must remain available to execute the project. Depending on the contractual terms, a delay can push back customer payments even after some costs have been incurred. Financing maturities and terms therefore matter as much as the amounts announced.

European financiers support expansion in Narvik

European institutions are already contributing to this financing effort. In May, Nscale announced $790 million in committed financing to continue developing its data centre in Narvik, Norway. ABN AMRO, DNB, Nordea, SEB and Eksfin are involved in the arrangement, which also includes an option for an additional facility of the same size that remains uncommitted.

Nscale’s expansion shows that competition in AI cloud also depends on the terms on which operators can finance their commitments. The availability, cost and timing of capital influence the capacity they can promise customers. For European operators, securing this financing is becoming a condition for building their industrial capabilities, and, in turn, for meeting demand and gaining market share.