This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.
If you searched “Korea Stargate,” here’s the one-line version: Korea has announced an AI-and-chip mega-plan on the scale of America’s Stargate — but it is built in a fundamentally different way, and the president is pushing it personally. The number everyone will quote is ₩1,000 trillion. That is not the number that matters. This is not financial advice.
This piece is the pillar for our series on Korea’s AI data center power supply chain — the big picture the company-by-company episodes feed into. For the map, start with Korea’s AI power supply chain (Ep.0).
The short answer: a real mega-plan, but a blueprint — not committed capex
On June 29, 2026, Korea’s president chaired a national briefing on “three mega projects,” with the chairmen of Samsung and SK in the room. The plan splits the country by role: a semiconductor cluster in Honam, AI data centers in Chungcheong, and physical AI in Yeongnam. The headline scale is more than ₩1,000 trillion (~$720B) over ten years, with some reports floating a combined Samsung-plus-SK figure of up to ₩2,000 trillion (~$1.45T) as a projection.[verify] It has been dubbed a “Korean Stargate” — but the label hides more than it reveals.
Why now — and why the president is driving it himself
AI needs compute, compute needs data centers, and data centers need power and chips at a national scale. Korea, the memory powerhouse, is trying to climb the value chain from selling parts to running AI. What’s different from a normal industrial-policy announcement is the execution structure: the presidential office has reportedly stood up a dedicated team for this project, with the president overseeing progress directly.[verify] The message is hard to miss — this administration intends to move at maximum speed within its current term.
🎩 Under the Gat — Outsiders will call it “Korea’s Stargate.” The comparison is fine, but it misleads. America’s Stargate is private Big Tech spending its own money. Korea’s version is the president announcing it with chaebol chairmen beside him — and now driving it personally, with a dedicated team. It isn’t the scale that differs; it’s the operator. That makes it fast. It also means the momentum is tied to a single administration — the part you can’t forget.
The part outsiders miss — the government lays the table, the chaebol fill the plates
The distinctly Korean feature here is the government-chaebol partnership. In the US, the government hands out subsidies (the CHIPS Act) but companies lead the projects. In Korea, the government designs the regional and policy frame directly, and the chaebol supply the capital and execution. The upside is speed — the state can push through permits, power and sites (ppalli-ppalli, Korea’s “hurry-hurry” culture). The downside is that a 10-year blueprint becomes hostage to the political cycle, and the government’s goal (balanced regional growth) does not always match the companies’ goal (profit and global competitiveness).
Both sides, fairly — the bull and the bear
The bull case. Over ten years, ₩1,000 trillion-plus of demand flows through chips, then power, then construction. AI data centers in particular pull the power infrastructure — transformers, cables, generation and cooling — and new memory fabs feed Samsung Electronics and SK Hynix capex plus the parts-and-materials ecosystem. A credible “Korean Stargate” narrative could also trigger a re-rating of Korea’s AI theme by foreign investors.
The bear case. An announcement is not a commitment — the headline mixes totals, combined plans and forecasts, and must be separated from this year’s actual capex. A global build-out happening at the same time raises the risk of oversupply that damages the memory cycle.[verify] And the structure that gives speed — direct presidential drive — also concentrates political risk.
🎩 Under the Gat — So the headline number is not an investment signal. Track three things instead: which pieces are already broken ground or budgeted, which listed company’s order book they land in, and whether the memory cycle can absorb the new capacity. Watch the capex, not the ₩1,000-trillion headline.
The comparison — same idea, different machine
| Korea (three mega projects) | US Stargate (anchor) | |
|---|---|---|
| Lead | Government-designed, chaebol-executed (Samsung, SK); president drives directly | Private Big Tech consortium (OpenAI, Oracle, SoftBank) |
| Scale | ₩1,000 trillion+ over 10 years (~$720B)[verify] | ~$500B by 2029[verify] |
| Build-out | Honam chips, Chungcheong data centers, Yeongnam physical AI | Large US data centers |
| Driver | Balanced growth + AI sovereignty | Securing AI compute |
| Decisive difference | Policy and politics are the swing variable | Private capital and demand are the swing variable |
Not a perfect parallel — both are giant AI-infrastructure pushes, but Korea’s leans more on a government blueprint, so tracking execution matters more.
What it really means — and where it’s heading
Strip away the headline and this is less a quarterly trade than a 10-year national bet: Korea trying to move from memory supplier to full-stack AI nation — chips, then data centers, then physical AI (robotics and manufacturing). The rate-limiting step for all of it is power; a data center is a machine that eats electricity, and the blueprint can’t be filled without nuclear, renewables and grid build-out.
- Bull scenario (2030s): Korea becomes an AI-infrastructure export package — memory plus data center EPC, power equipment (transformers, cables, cooling) and SMR — using the mega-project as a domestic reference.
- Bear scenario: the blueprint stalls on power, funding and politics, ends up repackaging existing memory capex, and collides with global oversupply that dents the memory cycle.
The timeline asymmetry that matters. Direct presidential drive plus a dedicated team means milestones are concentrated inside the current term — sites, groundbreaking and budgets get pulled forward, while the back half of the 10-year roadmap depends on the next administration’s continuity. Early acceleration, later uncertainty. For an investor, that argues for separating the near-term beneficiaries (power, construction, equipment) from the long-tail continuity risk.
🎩 Under the Gat — The center of gravity may shift from the chip itself to what runs the chip — the power and infrastructure layer. That’s exactly the value chain this series maps, company by company. If you only remember one line: this is a roadmap, and roadmaps are tracked by milestones, not headlines.
A watch-list, not a recommendation
After the announcement, these are the markers to follow — not a buy list:
- Actual groundbreaking and budget allocation — MOUs and forecasts versus real spending.
- Power progress — nuclear/SMR licensing, HVDC and grid expansion (the bottleneck).
- Foreign Big Tech and capital participation — whether a real “consortium” takes shape.
- The memory price cycle — does new capacity support it or break it?
- Policy continuity — the single biggest variable for a government-led project.
Which of these moves first will tell you far more than the ₩1,000-trillion headline ever could.
Disclaimer: This article is for information only and is not financial advice. Figures marked for verification — including the headline investment totals, the existence and remit of the presidential project team, and FX conversions — should be re-checked at publish time against official sources. FX uses roughly ₩1,380 per US dollar as of June 2026.
Frequently Asked Questions
What exactly is the “Korea Stargate”?
Its official name is the “three mega projects” — a semiconductor cluster in the Honam region, AI data centers in the Chungcheong region, and physical AI in the Yeongnam region. Its scale and ambition earned it the comparison to America’s Stargate. This is not financial advice.
Is the figure really ₩1,000 trillion?
That number mixes a 10-year cumulative total, combined corporate plans, and forecasts. Some reports cite a combined Samsung and SK figure of up to ₩2,000 trillion (~$1.45T) as a projection. Committed annual capex is a much smaller, separate thing — treat the headline with care. Figures should be re-checked at publish time. Not financial advice.
Why is it moving so fast — is this a short-term trade?
The presidential office has reportedly set up a dedicated team and the president is overseeing it directly, which points to maximum speed within the current term. But it’s a 10-year roadmap, so it is a milestone story to track, not a short-term trade — and political continuity is the key variable. This is not a recommendation to buy any security.
Which stocks would benefit?
Directly, the chipmakers Samsung Electronics (KRX: 005930) and SK Hynix (KRX: 000660). Indirectly, the AI power supply chain — transformers, cables, generation and cooling — which our series covers company by company. This is not financial advice.
How is this different from the US Stargate?
The US Stargate is led by a private Big Tech consortium (OpenAI, Oracle, SoftBank). Korea’s version is designed by the government and executed by the chaebol, with the president driving it directly. That makes it faster but more tied to one administration.
This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.