Power Equipment Stocks — Same AI Supercycle, Opposite Moves in Korea and the US

Why have Korean and US power equipment stocks diverged so sharply within the same AI power supercycle, and what should investors check before betting on convergence

This is the English version of a post originally written in Korean for my algorithmic trading system devlog(new tab).

As AI data centers have exploded in number, "where does the power come from" has become a central market question this year. It's now widely understood that transformers, switchgear, and grid equipment are as much of a bottleneck as GPUs, and related companies led the market through the first half of the year. Right now, though, Korean and US power equipment stocks are telling opposite stories.

Where Korea's top four power equipment names stand

Looking at the leading names across Korea's power value chain, current prices are uniformly sitting around half of this year's peak.

One more thing stands out: all four stocks hit their yearly high on the exact same day (May 7). That's not company-specific bad news — it means the entire power equipment theme cooled off together, peaking on that one day. It reads as a sector-wide unwind of the extreme valuation premium the theme had accumulated during the first-half rally.

The other side of the world — the US keeps hitting new highs

Over the same stretch, the leading US name is telling the opposite story. GE Vernova (GEV), the top grid-and-generation-equipment name, set a new all-time high ($1,195.94) on July 6, up roughly 76% year to date. The reason is straightforward.

• GE Vernova's electrification segment booked $2.4 billion in data-center equipment orders in Q1 alone — already more than the entire prior year.
• Hyperscaler AI capex is projected to cross $750 billion in 2026 and $1 trillion in 2027.
• Executives at GE Vernova, Siemens Energy, and Eaton have all separately described the current phase as "a decades-long investment supercycle."

In other words, the theme's structural logic — grid investment plus data-center power demand — is still very much alive, and the US name most directly exposed to it is making new highs.

So this isn't "following along" — it's a bet on the gap closing

"If the US is rallying, won't Korea follow?" is a natural instinct. But the current setup isn't simple co-movement. The US is at all-time highs while Korea sits at roughly half its peak — an already-wide gap, and the question is whether it narrows. That's a much more conditional bet, and a completely different trade from just riding the theme.

Betting on convergence means first answering why the gap opened. Broadly, there are three possibilities.

1. Bubble unwind — if Korea simply ran up too fast in the first half and is now shedding that premium, it could still have further to fall.
2. Structural difference — if domestic companies have meaningfully less direct data-center order exposure than the US leader, there was less reason for them to rally as hard in the first place.
3. Temporary flows — if this is really just sentiment or capital rotation, it could be a genuine catch-up opportunity.

Distinguishing between these requires actual data — order backlogs and the share of revenue tied to data centers. That's why "it's down by half, so it must be cheap" isn't a conclusion you can draw from the chart alone.

What investors should actually check

• Valuation risk on the US side: GEV is at an all-time high but already trades around 59x earnings, and analyst price targets sit below the current price. "Following the US" could easily mean buying into the top.
• Falling-knife risk: some of the Korean names are close to their yearly lows. A convergence trade can keep getting pushed down before a bottom is confirmed, so it's worth setting an invalidation rule in advance (e.g., exit if the yearly low breaks).
• Catalysts to watch: quarterly earnings and order disclosures from Korean power-equipment companies, and whether the US leader can hold its elevated valuation or starts to wobble. If the US rolls over, the whole theme likely de-rates together.

Bottom line

The theme's big-picture logic — AI and grid demand — still holds. But the fact that Korea's leading names are "down by half from their peak" isn't itself a buy signal. This is a gap, not a co-movement, and whether it closes comes down to order-backlog data. Rather than chasing the US new-high headline, it's safer to first confirm why the gap opened and set an invalidation rule before stepping in.

This post is a personal analysis based on publicly available market data, not investment advice recommending the purchase or sale of any specific stock. Investment decisions and their outcomes are the sole responsibility of the investor.