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.
The market hunts the “AI trade” in chips. Korean insiders read something quieter — the order slips for the hardware that moves the power. On July 2, 2026, one of those slips became public. This is not financial advice.
A spin-off from our series on the AI data center power supply chain — this time, not the theory, but the contract.
The event, in one line
On July 2, 2026, HD Hyundai Electric (KRX: 267260) disclosed a long-term contract with a single unnamed global big-tech firm to supply up to ₩1.12 trillion (~$721 million) of power infrastructure for data centers across North America. It is the first piece of contract evidence bolted onto the “AI needs power” story — so instead of the narrative, we read the document.
What the filing actually says
[Filing, in brief] Master supply agreement with a global big-tech firm. Total value up to ₩1.12 trillion. Split: distribution equipment ₩553.9 billion + power equipment (extra-high-voltage transformers) ₩567.3 billion. Delivery: sequential through 2028, tied to the customer’s construction schedule. Customer: undisclosed (confidential).
Break the numbers apart and three things read out.
It’s split roughly in half. Distribution ₩553.9bn and power ₩567.3bn — almost even. That means this isn’t a single-product order; it’s a package (transformers plus distribution gear), the kind that wires a whole data-center block.
It isn’t one-off. “Sequential through 2028” is multi-year revenue visibility. As of Q1 2026, HD Hyundai Electric’s order backlog stood at $7.89 billion (up 17.2% from end-2025) — equal to more than two years of revenue.
The customer is hidden. The filing says only “global big tech.” The list of firms building data centers at this scale in North America is short — but a guess stays a guess. (YMYL: no speculation dressed as fact.)
Timeline — what happened when
| When | What |
|---|---|
| ~2023–24 | US extra-high-voltage transformer lead times stretch to years as domestic capacity thins — orders concentrate on a few Korean and European makers |
| FY2025 | Record profitability: full-year operating profit ₩995.3 billion (+49% YoY) on revenue of ₩4.08 trillion, with a Q4 operating margin of 27.6% — driven by high-margin North American transformers |
| Q1 2026 | Operating profit ₩258.3 billion (+18.4% YoY), 24.9% operating margin; order backlog $7.89 billion; North America revenue +26.6% |
| 2026-07-02 | ₩1.12 trillion big-tech contract disclosed (distribution + power package) |
| ~2028 | Sequential delivery expected to complete, tied to customer build-out |
The point of the timeline: July 2 is one dot on a line that was already moving. The bottleneck — America’s shortage of high-voltage transformer capacity — has existed for years. July 2 is simply the day it showed up as a single, signed contract.
Why it matters now
Two reasons.
First, the narrative became a document. “AI data centers → power demand → Korean equipment makers benefit” was, until now, a thesis. The July 2 filing attaches a number and a delivery date to that thesis for the first time. It’s the structure our pillar piece mapped — now with a receipt.
🎩 Under the Gat — I don’t trust themes. I trust order slips. The value of the July 2 filing isn’t the number “₩1.12 trillion” — it’s that the number comes with a delivery schedule. You can’t manufacture a schedule out of a tweet. For an investor, that’s the difference between a story and a backlog.
Second, the timing is cruel. That same week, on July 2, the KOSPI plunged on a semiconductor rout — Samsung Electronics fell 9.1%, SK Hynix dropped 14.5%, and the index shed 7.9%. Power-equipment names were dragged down with it, entirely unrelated to their own fundamentals. A week when the contracts improved but the share price followed the index down: an entry for the bull, a warning for the bear that “it’s still an index stock.”
Reasons not to buy on this filing alone (fairly)
Drawing a buy conclusion from a single contract is dangerous. Here is the skeptic’s case, stated honestly.
- Customer concentration and secrecy. The bigger the deal, the more it leans on a few clients. With the customer undisclosed, outsiders can’t easily verify continuity or follow-on orders.
- It’s a cyclical business. Transformers are inherently boom-bust. If today’s peak-cycle margins (a record 27.6% in Q4 2025, 24.9% in Q1 2026) are the top, a two-year-plus backlog is visibility, not permanence.
- Won and access. For dollar investors, currency risk and access friction (no ADR) are real costs.
- The discount has reasons. Even if it trades below US pure-plays, a conglomerate discount and the absence of a pure-play structure are the grounds for that discount. (Verify consensus multiples at publication.)
🎩 Under the Gat — The Korea discount predates me. This one filing won’t close it overnight. But the rate at which order slips are printing as public disclosures is a new kind of data. Track that rate, not the narrative. A view, not advice.
What to watch
Not the headline size, but the follow-through: (1) additional disclosed contracts at similar scale, (2) whether margins hold as the backlog converts, and (3) big-tech capex guidance — the leading signal for whether these orders keep coming. Watch the delivery, not the drama.
This is not financial advice. Every figure here should be re-verified at the time of reading against primary sources (DART filings, company IR).
Related: Korea’s AI power supply chain — the full map · Korean transformer stocks · Data-center cooling & distribution.
Frequently Asked Questions
How big is the ₩1.12 trillion deal for HD Hyundai Electric?
It is a substantial single contract that will be recognized in stages through 2028, not booked all at once. Treat it as multi-year revenue visibility rather than a one-off sales spike. This is not financial advice.
Who is the big-tech customer?
Undisclosed. The filing names only a “global big tech” client, citing confidentiality. The pool of firms building data centers at this scale in North America is small, but any guess remains a guess.
Why can’t the US just make its own transformers?
Decades of shrinking domestic manufacturing pushed US high-voltage transformer lead times out to several years, so orders concentrate on a few Korean and European suppliers that kept their capacity. Verify latest lead-time data.
Why did the stock fall the same day the deal was announced?
A semiconductor-led KOSPI plunge on July 2 dragged power-equipment names down with the index. It’s a textbook case of the fundamentals (the contract) and the share price (index and flows) pulling in opposite directions.
What if this is the top of the cycle?
That is the core risk. Transformers are a cyclical business, and a three-year backlog is visibility, not permanence. Track big-tech capex guidance for signs of order slowdown. Not financial advice.
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.