Naver said on 27 October that its contract with Brookfield for its AI factory has met the precondition for NVIDIA’s strategic equity investment, Chosunbiz reported. In a filing with Korea’s DART disclosure system the same day, Naver set out what backs it: a payment guarantee of 18.74 trillion won for its cloud unit, equal to 64.73% of Naver’s equity, running to the end of 2034.
The filing says the guarantee covers Naver Cloud’s payments for GPU computing from a Brookfield vehicle with an investment limit of $9 billion. It calls the 18.74 trillion won a nominal total that assumes the full $9 billion is deployed, and says no payment obligation was fixed on the filing date.
Video explainer · Naver’s AI factory guarantee
A 16:9 video explainer on Naver’s AI factory guarantee, narrated from the published text of this piece and four related FourWeekMBA pieces.
The key insight: As we read it, the filing moves the AI factory’s risk onto Naver’s own balance sheet twice: Asset Co. is to be consolidated as a Naver subsidiary, and Naver guarantees its cloud unit’s payments to it.
Naver’s amended business-plan filing says the $9 billion of computing infrastructure will be built by Asset Co., a tentatively named company to be set up by an investment vehicle of the Brookfield AI Infrastructure Fund, and says Naver Cloud will use it.
The filing says Naver Cloud will buy GPU computing as a service, or GPUaaS, from Asset Co. and sell AI data center services to its own customers. GPUaaS means renting GPU capacity rather than owning the chips.
Asset Co. will be 100% funded by the Brookfield vehicle, but Naver’s guarantee filing says that under the investment and operating contract and K-IFRS accounting rules Naver’s control is recognised, so Asset Co. is to be brought in as a Naver subsidiary and affiliate.
The guarantee filing names the creditors as Green Holding L.P., a special purpose vehicle of the Brookfield AI Infrastructure Fund, and Asset Co. It also says GPU investment will be made in stages, batch by batch, as demand comes in.
Naver’s staged AI-factory plan: 55MW in the first half of 2027, 100MW cumulative by end-2027 and 200MW cumulative in 2028, which Naver calls tentative targets. Source: NAVER Corp. amended business-plan filing on DART, 27 Oct 2026.Naver’s 18.74 trillion won payment guarantee for Naver Cloud, equal to 64.73% of Naver’s equity, and the Brookfield vehicle’s $9 billion investment limit. Source: NAVER Corp. DART filings, 27 Oct 2026.
What the Guarantee Covers
The 18.74 trillion won is the nominal total of supply payments Naver Cloud is expected to make through the longest contract term, to the end of 2034, if the $9 billion investment limit is fully spent, the filing says. It adds that the amount can change with the GPU investment schedule, interest and exchange rates and actual cost settlement.
Each payment obligation runs for six years from the day supply of that stage begins, so the end date can move with the last stage, according to the filing. The guarantee also covers other obligations, such as damages and penalties, which the filing says were left out of the figure because they cannot be calculated yet.
Naver Cloud will pay Naver a guarantee fee of 670,725 million won over the guarantee period, a separate filing says, estimated at an annual rate of 0.4375% on the assumption that the $9 billion limit is fully executed.
For scale, the guarantee filing lists Naver Cloud’s 2025 revenue at 1,554,458 million won and net income at 165,643 million won. The 18.74 trillion won guarantee is about 12 times that revenue, by our arithmetic, though the two cover different periods: one year of sales against payments that can run to 2034.
The Plan Behind It
Naver’s filing lists a staged build: 55MW in the first half of 2027, 100MW cumulative by the end of 2027 and 200MW cumulative in 2028, with gigawatt-scale infrastructure as the final aim. It calls the schedule a tentative target that is still under discussion.
The filing says NVIDIA will take part in a third-party share allotment of Naver common stock worth about $1 billion. The 24 July release from Naver, Brookfield and NVIDIA said “Brookfield will fund up to $9 billion as the exclusive capital partner, NVIDIA will invest $1 billion and NAVER will fund the remaining amount to finance the $10 billion project.”
That release said the 200-megawatt AI factory, “featuring NVIDIA Vera Rubin and Blackwell platforms,” will be built with the NVIDIA DSX platform at Naver’s GAK Sejong data center in Sejong. Chosunbiz reported that Naver plans to quickly complete the share issue to NVIDIA, with specific terms to be disclosed later.
Connecting the Coverage
The structure echoes a deal we covered on 26 October under the headline $63B SPV will buy Broadcom chips and lease them to Anthropic, which describes how a capital partner owns the hardware and the AI operator pays to use it. Here the operator’s parent guarantees the payments, and the vehicle sits on Naver’s books.
As we read it, the $9 billion is Brookfield’s money but Naver’s commitment: Naver’s filing says the 18.74 trillion won is what Naver Cloud is expected to pay through 2034 if the full limit is spent.
As we read it, the staging is the safety valve: the filing says GPU investment is made batch by batch as demand comes in, and each payment obligation runs six years from that stage’s start.
As we read it, the guarantee fee is small next to the exposure: 670,725 million won over the period, at 0.4375% a year, according to Naver’s separate filing.
NAVER, Brookfield and NVIDIA joint release, 24 July 2026
“Brookfield will fund up to $9 billion as the exclusive capital partner, NVIDIA will invest $1 billion and NAVER will fund the remaining amount to finance the $10 billion project.”
Three Implications
FOR NAVER The guarantee filing puts the nominal guarantee at 64.73% of Naver’s equity, running to the end of 2034.
FOR BROOKFIELD The filing names Green Holding L.P., a special purpose vehicle of the Brookfield AI Infrastructure Fund, and Asset Co. as the creditors.
FOR NVIDIA The business-plan filing says NVIDIA will take part in a third-party allotment of Naver common stock worth about $1 billion.
The framework sets out its question: “How the money flows, where it cracks, what it buys, who can absorb it, and what happens when revenue can’t catch up.”
As we read it, Naver’s filings give one answer to “who can absorb it”: a capital partner funds the hardware, and the operator’s parent guarantees the payments. The framework was published in May 2026 and does not discuss Naver.
What Is Not Established
The 18.74 trillion won is a nominal total: Naver’s filing says no supply payment obligation was fixed on the filing date and the amount depends on how much of the $9 billion is invested. The schedule is a tentative target. The terms of NVIDIA’s share purchase are to be disclosed later, and the filings do not name Naver Cloud’s customers for the capacity.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Naver’s 27 October filings show it guaranteeing a nominal 18.74 trillion won, 64.73% of its equity, of its cloud unit’s payments to a Brookfield-funded vehicle investing up to $9 billion, a contract that Naver says meets the precondition for NVIDIA’s investment.
A note on sourcing. Guarantee, plan and fee figures are from NAVER Corp.’s three DART filings of 27 October 2026, read in Korean on the DART site on 27 October; statements attributed to the filings are our paraphrase of the Korean text. The July figures are quoted from the joint NAVER, Brookfield and NVIDIA release of 24 July 2026. Naver’s statement on the NVIDIA precondition is as reported by Chosunbiz. Won figures are not converted. We did not contact Naver, Brookfield or NVIDIA.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
