THE CORRIDOR · Infrastructure
Nvidia is helping pay for the data centre that will buy its chips
A Korean internet company, an American chipmaker and a Canadian asset manager are building Korea's national AI capacity together. Look at who is funding whom.
Chandni Melwani
Founder & Editor
Jul 24, 2026 · 2 MIN READ
The News
On July 24, 2026, NAVER, NVIDIA and Brookfield announced an expansion of Korea's national AI factory infrastructure, growing the NVIDIA DSX deployment at NAVER's GAK Sejong data centre to 200 megawatts by 2028. That more than triples the 55-megawatt buildout announced the previous month. NAVER has said it intends to expand its NVIDIA AI infrastructure to 1 gigawatt over time. The announcement was made during Korean President Jae Myung Lee's AI Summit visit to San Francisco.
Know More
- — The site: NAVER's GAK Sejong hyperscale data centre in Sejong, South Korea, running the NVIDIA DSX platform.
- — The step change: 55 megawatts announced in June, 200 megawatts targeted by 2028 — roughly a tripling within two months of the original announcement.
- — NAVER's stated ambition beyond this deal: 1 gigawatt of NVIDIA infrastructure over time. No date attached to that figure.
- — In the joint release: NVIDIA plans to invest $1 billion into NAVER, and Brookfield has entered a nonbinding term sheet to fund up to $9 billion. NVIDIA's investment is conditional on customary closing conditions and on NAVER first securing at least $9 billion of committed financing separate from it.
- — Reported after the release, from NAVER itself: the $1 billion takes the form of 7.2 million new shares issued to NVIDIA by third-party allotment, giving it about a 4.5% stake. That percentage is not in the joint announcement.
- — Who NAVER is: South Korea's dominant search and internet company, the closest domestic equivalent to Google, and the operator chosen to host national AI capacity.
- — The political setting: announced while Korea's president was in San Francisco for an AI summit, which is why this reads as national industrial policy rather than a private procurement.
- — Not disclosed: the split of financing between debt and equity, the terms of Brookfield's participation, and what NAVER pays for the hardware.
Two months ago the plan was 55 megawatts. On July 24, standing up during Korea’s president’s AI summit visit to San Francisco, NAVER, NVIDIA and Brookfield said they would take it to 200 by 2028, at NAVER’s data centre in Sejong.
Megawatts is the right unit here, and it takes a moment to explain why. A building full of AI chips is constrained by electricity long before it runs out of room, so capacity gets quoted the way a power station is rather than the way an office is. Tripling to 200 megawatts is a statement about how much electricity Korea intends to point at model training, not about square footage.
The financing is the part worth slowing down on. NAVER brings the site and the national mandate; Brookfield brings capital, which is the scarcest input at this scale, since data centres this size are financed like toll roads rather than bought out of cash flow. And NVIDIA, which in the same release says it plans to invest $1 billion into NAVER — a stake NAVER later put at roughly 4.5%, or 7.2 million new shares — brings both the chips and a slice of the buyer.
That last arrangement is becoming ordinary and is worth naming anyway. When the supplier holds equity in the customer, part of the demand is financed by the seller, and an order book stops being a clean read on independent appetite. It is not evidence of anything wrong — capital-intensive industries have always worked this way — but it does mean the Korea number and the Gulf numbers and the European ones should be read as one question rather than three: how much of the world’s AI buildout is being funded by the people selling into it.
Related
Frequently Asked Questions
What is an "AI factory," and why is it measured in megawatts?
An AI factory is a data centre built specifically to train and run AI models rather than to host general computing. The distinction is mostly about density and power. Racks draw far more electricity, cooling is engineered around it, and capacity gets quoted in megawatts rather than servers or floor space, because power is the binding constraint on how much AI a building can actually do.
Why does this need three companies?
Each supplies something the others cannot. NAVER has the site, the operating experience and the national mandate. NVIDIA has the hardware and the platform. Brookfield has the capital, which is the scarcest input once a buildout reaches this size — data centres at this scale are financed like infrastructure projects rather than bought out of operating cash.
Why does it matter that NVIDIA is reported to be taking a stake in NAVER?
Because it changes how you read the demand. When a chip supplier also holds equity in the company buying the chips, some portion of that demand is being financed by the seller. That is a normal arrangement in capital-intensive industries and it is not by itself a warning sign, but it does mean an order book is a weaker signal of independent appetite than it looks. Worth watching as a pattern rather than judging as a single deal.
Which parts of this are officially confirmed?
More than the money coverage implied. The joint release carries the capacity expansion, the site, NVIDIA's planned $1 billion investment and Brookfield's nonbinding term sheet for up to $9 billion. What it does not carry is the shape of that investment — the 7.2 million new shares and the roughly 4.5% holding came from NAVER separately, after the announcement. Hold all of it lightly, because the release calls the expansion proposed and the investments planned, and NVIDIA's money is conditional on NAVER first securing at least $9 billion of committed financing of its own. The debt-equity split and hardware pricing were not published.
Sources
- NVIDIA Newsroom, "NAVER, NVIDIA and Brookfield to Expand Korea's National AI Factory Infrastructure Buildout"
- Brookfield Asset Management, "NAVER, NVIDIA and Brookfield to Expand Korea's National AI Factory Infrastructure Buildout"
- NVIDIA Investor Relations — the same release, filed
- Korea JoongAng Daily, "Nvidia to buy 4.5 percent stake in Naver in $1 billion deal" — the share count and the 4.5%, which the joint release does not carry
- Reuters, "South Korea's Naver jumps 10% on Nvidia's $1 billion investment plan" — NAVER confirming the third-party allotment
Chandni Melwani
Chandni Melwani is the founder and editor of New in AI, covering AI agents, M&A, and enterprise adoption. She holds a Master's in Management of Artificial Intelligence from Queen's University and brings a practitioner's perspective from her work in Data and AI leadership.
Follow ↗The Corridor Briefing
Get the two-beat briefing before the market opens.
One email. Corridor capital moves and Commerce platform shifts, decoded daily.
No spam. Unsubscribe anytime.
More from The Corridor
Policy
The White House says Moonshot AI trained its model on Anthropic's, and accessed restricted Nvidia chips in Thailand
Washington built export controls to keep advanced chips out of Chinese hands. The allegation is that a Chinese lab reached them in Thailand instead, and that the model capability never needed to cross a border at all.
Chandni Melwani · China
Infrastructure
The UAE spent two years winning access to American AI chips. Now it is renting them out.
For a Gulf bank, the question was never whether sovereign AI was a good idea. It was what it costs, and who has to file the export paperwork. There is finally an answer to both.
Chandni Melwani · GCC
Infrastructure
Microsoft is paying to put Azure's European customers on compute it does not own
Europe is trying to build its own AI sovereignty and rent it back. A US hyperscaler is underwriting the capex, and structuring the cheque to stay clear of Brussels.
Chandni Melwani · Europe