By Olivia Hart

Naver has signed a $9 billion computing-infrastructure contract with Brookfield, the company said on the 27th, as construction began on an AI factory it is pursuing with NVIDIA and Brookfield.

Naver stated: “We have signed a 9 billion dollars contract with Brookfield to secure computing infrastructure,” adding that the deal “fulfills the preconditions for NVIDIA’s strategic equity investment.”

Earlier in July, Brookfield announced that it would invest up to $9 billion, approximately 12 trillion won, in the factory. Once third-party paid-in capital allocation to NVIDIA is finalized, NVIDIA will become Naver’s third-largest shareholder.

The capital structure links three positions: Brookfield supplies project funding and infrastructure, NVIDIA supplies technology and proposed strategic equity, and Naver Cloud supplies the operating business. Naver expects that division to reduce the initial capital burden while keeping payments tied to customer revenue.

What does Brookfield’s $9 billion fund?

Brookfield will establish a special purpose vehicle tentatively named “Asset Co.” The SPV will invest up to $9 billion to secure AI infrastructure, including GPUs and data centers, before providing those assets to Naver Cloud, the factory operator.

In operational terms, the arrangement resembles a project-finance lease: Asset Co. funds the hardware and facilities, Naver Cloud runs the service, and customer receipts support the SPV through usage fees. Naver will sell computing infrastructure to external customers, then pay usage fees to the SPV as that business generates revenue.

This structure gives Naver a route to large-scale computing capacity without carrying the full upfront purchase alone. It also places Asset Co. at the center of the cash-flow chain, linking its returns to infrastructure usage.

When does capacity arrive, and what does NVIDIA contribute?

Naver plans to operate 55 megawatts (MW) in the first half of 2027, then expand operations to 200 MW by 2028. Those dates give customers and investors a staged view of the factory’s ramp rather than one undifferentiated capacity target.

NVIDIA will provide technical collaboration centered on next-generation high-performance computing platforms. That support connects the equity investment to the systems Naver Cloud will operate, while Brookfield’s SPV finances the underlying GPUs and data centers.

How does Naver limit early-stage financial risk?

Naver expects its cooperation with Brookfield to mitigate the initial investment burden required for large-scale AI infrastructure. The company also expects to manage financial risks during the early stages of the business and enhance operational stability by linking usage fees to business revenue.

The structure separates the initial capital requirement from later operating payments. Revenue from external customers therefore becomes central to the SPV’s fee income, while Naver sells computing infrastructure through Naver Cloud.

Naver has also provided an eight-year payment guarantee for Naver Cloud’s usage-fee obligations. A Naver official said: “The payment guarantee is a measure to secure trust for smooth contract fulfillment and does not translate into actual cash outflows or financial burdens for Naver.”

Naver describes the guarantee as contractual assurance rather than cash spending; the revenue-linked fee mechanism remains central to the arrangement. The distinction separates Naver’s stated credit support from the usage income generated by Asset Co.’s assets.

Why does the structure matter amid neo-cloud concerns?

Investor confidence in neo-cloud companies had been weakening after CoreWeave, a representative global operator, saw its corporate bond yields surge once to 15%. The episode placed financing structures for capital-intensive cloud operators under closer scrutiny.

An investment-banking industry insider contrasted the pact with circular financing, which the insider described as artificially delaying debt to create an illusion. The insider said the Naver, NVIDIA and Brookfield alliance instead allocates capital and technology efficiently through a strict division of roles.

Naver also has a wider earnings base. Unlike neo-cloud operators, which lack revenue streams beyond cloud and GPU rentals, Naver stably generates operating profits from advertising and commerce. That distinction separates Naver’s financing case from operators dependent only on rental income.

What should investors and cloud buyers watch?

The immediate gating step is the final allocation of third-party paid-in capital to NVIDIA, which determines whether NVIDIA becomes Naver’s third-largest shareholder. Asset Co. must then be established and fund the GPUs and data centers that Naver Cloud will operate.

The operating milestones remain clear: 55 MW in the first half of 2027 and 200 MW by 2028. The financial test is equally straightforward: external computing revenue must support usage fees tied to the SPV’s assets, while the eight-year guarantee protects contract fulfillment.

For investors, the relevant issue is whether staged usage revenue can sustain Asset Co. amid the financing strain seen in neo-cloud operators. Naver has reduced its upfront burden, but utilization, capital allocation and delivery dates remain the three tests that matter.