SK Hynix -20% in Frankfurt: The Day Global Chips Cracked (June 5, 2026)

Global equities cracked all at once on June 5.

SK Hynix GDR -20.53%, Samsung KRX -6.40% (KRW 329,000), SK Hynix KRX -7.92% (KRW 2,116,000). KOSPI closed at 8,160 (-5.54%), and KOSPI200 futures fell as much as -8%, triggering a sell-side sidecar.

To understand the drop you have to look at both the chip sector and the US jobs print together.

First, why it fell. Then, what investors should actually think about.

The US May nonfarm payrolls released on June 5 came in at 172,000, double the 85,000 consensus. March and April were also revised higher.

Strong jobs mean a strong economy, but the market read it the other way.

If the economy doesn't cool, the Fed can't cut, and might even hike.

The US 10-year yield jumped to 4.54%. A cut at the June 16-17 FOMC all but vanished, and odds of a hike by year-end rose to about 70%.

When rates stay high, expensive tech gets hit first. A textbook 'good economic data is bad news for stocks' setup.

Broadcom's June 3 results were strong on the numbers: AI chip revenue of 10.8B for the quarter, total revenue 22.2B (+48% YoY).

The problem was next-quarter AI guidance of 16B, below the 17.2B the Street wanted.

Then CEO Hock Tan noted Google would likely use multiple chip suppliers and flagged that surging AI sales were pressuring margins, and the disappointment selling poured in.

Broadcom -13% led the Philadelphia Semiconductor Index (SOX) lower, dragging Samsung and Hynix with it.

US memory maker Micron fell 7.7%.

The key point is it dropped despite good news. Just before, Nvidia had approved Samsung, SK Hynix, and Micron all three as HBM4 suppliers for its next-gen Vera Rubin platform.

The stock fell anyway.

At Computex 2026 on June 2, SK Chairman Chey Tae-won announced a plan to double SK Hynix wafer capacity within five years.

That signals strong demand, but part of the market read it inversely.

'Great now because of AI, but won't oversupply hit in a few years?'

Add AI valuation strain (the top 5 S&P names are 30% of the index) and Iran-driven oil (Brent in the mid-90s), and the selling intensified.

Korea's drop had two more local drivers.

First, FX. On June 5 USD/KRW hit 1,559.95 (+1.77%), down about 8% in a month. A weaker won means foreigners also eat currency losses, so they sell Korean stocks harder.

Second, the National Pension Service (NPS). This one matters more than it looks.

With KOSPI up over 25% YTD, NPS's actual domestic-equity weight had pushed well past its 14.9% target to the 19% area, even spiking to 24.5% at end-February. By the old rule, it would have to mechanically sell the excess.

But on May 28 NPS raised the domestic-equity target from 14.9% to 20.8% (+5.9pp). The market called it dodging a 'roughly KRW 170 trillion forced-sell.'

For reference, the year-end target mix is domestic equity 20.8%, foreign equity 34.7%, domestic bonds 23.1%, foreign bonds 7.4%, alternatives 14.0%.

So on June 5, NPS was a buffer, not the culprit. Korea's flow picture is less dire than it looks.

This is the core.

1. This is a correction within a bull market, not a crisis

First, context: the market is in a fourth straight bull year and the S&P 500 sits near record highs. June 5 was profit-taking and rate repricing from an overheated state, not a crisis. In bull markets a 10-20% correction lasts about 17 days on average, a normal rhythm.

The short-term triggers (jobs, Broadcom) are days-to-weeks noise.

2. But this dip may last longer than usual

In the past, when markets wobbled the Fed cut rates to cushion them. Now, with inflation (April CPI 3.8%) and Iran-driven oil, the Fed can't cut even if it wants to.

So the 'a selloff brings a quick rate-cut backstop' card is weak this time. Valuations have to digest on their own, so the dip can run longer.

3. What really matters is whether peak-out shows up in results

The selloff is at the 'oversupply might come' worry stage. Actual memory prices and earnings have not rolled over.

If the worry shows up in numbers, it's a real correction; if not, it's overheating relief before another leg up. That makes DDR5 prices and HBM4 contract pace the swing factor.

4. Korea's flow is less bad than it looks

NPS dodged a forced-sell via the target hike. Foreign selling and FX are a drag, but with the pension as a backstop, a foreign-selling pause could give the rebound some lift.