Kim Yong-Beom is the presidential policy chief of South Korea. This morning he wrote something on Facebook about AI profits and Korean citizen dividends.
By the time Korean markets processed what he wrote, foreign investors had dumped $3.8 billion worth of Korean stocks. The Kospi, South Korea's benchmark index, plunged 5.1% in minutes, falling from a session high of 7,999 points to as low as 7,421. Samsung and SK Hynix, which together represent the backbone of my DRAM ETF position, reversed early gains and turned sharply lower.
Then Kim clarified what he actually meant. Markets partially recovered. The Kospi closed down 2.29%.
Total elapsed time from post to clarification: a few hours. Total market damage at the low: one of the largest single-session selloffs in Korean market history.
This is what fragile sentiment looks like at the top of a supercycle.
What He Actually Said
Kim's post argued that AI profits shouldn't remain concentrated in a handful of companies and asset holders. He proposed a "citizen dividend", a mechanism to distribute the gains of the AI infrastructure boom back to ordinary Koreans.
The market heard "windfall tax on Samsung and SK Hynix" and hit sell.
What Kim actually meant, as he later clarified, was using excess government tax revenue generated by the AI boom, not imposing a new corporate levy on earnings. The distinction is enormous. One directly compresses margins at your portfolio companies. The other is a government budget allocation decision that has nothing to do with Samsung's income statement.
Investors who read the clarification bought back. Investors who sold on the headline ate the loss.
Why the Market Was That Fragile
The Kospi has been the world's best performing major market this year, driven almost entirely by two stocks. When your entire index rally is concentrated in Samsung and SK Hynix and a senior presidential adviser starts writing about redistributing their profits, the exit math gets terrifying fast.
Foreign investors didn't wait to read the fine print. They sold $3.8 billion in a single session, one of the largest foreign outflow days in recent Korean market history. Institutional investors piled on with an additional 1.21 trillion won in net selling. Only retail investors were net buyers, absorbing 6.68 trillion won, not enough to stop the bleeding.
This is the concentration risk I've written about before. When two stocks drive an entire national index, any headline that threatens those two stocks becomes a systemic event. It doesn't matter if the headline turns out to be wrong.
The Number That Explains Why Politicians Are Paying Attention
Here's the part that gets buried in the political drama.
Samsung's chip division operating profit surged 48-fold in Q1 2026, from 1.1 trillion won to 53.7 trillion won in a single quarter. That semiconductor division now accounts for 94% of Samsung's entire company profit. Samsung's Q1 operating margin hit 70%, surpassing both Nvidia and TSMC. KB Securities projects Samsung will be the world's second most profitable company for all of 2026, behind only Nvidia.
Politicians don't write Facebook posts about industries that aren't making extraordinary amounts of money. The scale of profit concentration is precisely what's drawing political attention, and that profit concentration exists because the memory supercycle is real, not because of accounting tricks or financial engineering.
When a presidential adviser argues that memory chip profits should be redistributed to all citizens, he's inadvertently confirming that memory chip profits are unlike anything South Korea, or the world, has seen before.
That's my thesis, validated by the opposition.
The Risk That's Actually Real
I'm not going to pretend the political dimension doesn't exist. It does.
President Lee's administration ran on inclusive growth. South Korea has a long history of applying regulatory and political pressure on chaebols when profits become too visible. The Samsung workers threatening an 18-day strike are making the same argument as Kim Yong-beom, the profits are extraordinary, and we want our share. These pressures don't disappear just because one Facebook post got walked back.
The realistic risk progression:
Now: Policy adviser floats idea, market panics, clarification follows, partial recovery. Noise, but revealing noise that carry a hint of clarity.
Near term: Political pressure on Samsung and SK Hynix intensifies. Workers want profit-sharing. Politicians want tax revenue. Consumers are already paying higher prices for memory-dependent products. The pressure comes from every direction simultaneously.
Medium term: If memory profits remain this concentrated and visible for another 2-3 years, formal legislative proposals become more likely. Not inevitable. More likely than they were six months ago.
Worst case: A genuine windfall tax on semiconductor earnings. This would directly hurt your position. It's also what South Korea has historically avoided because they understand these companies need to remain globally competitive. But "historically avoided" is not the same as "impossible."
Low probability. Significant impact. Goes in the risk register, not the exit trigger.
The Signal I'm Watching
One Facebook post that gets clarified within hours is noise. A formal legislative proposal with cross-party support is a different conversation entirely.
I'm not watching Kim Yong-beom's Facebook page. I'm watching whether this morning's post gets picked up as serious policy by the opposition party, whether it appears in formal budget committee discussions, and whether Samsung or SK Hynix management addresses it on their next earnings call as a material risk.
None of those things have happened. Until they do, this is sentiment volatility, painful if you sold into it, irrelevant if you held.
Still Holding
A $3.8 billion selloff triggered by a Facebook post that got walked back within hours tells you two things simultaneously.
First, the supercycle is so profitable that senior government officials are publicly debating how to redistribute the gains. That's not a bubble narrative, that's real profit concentration driving real political response.
Second, the market is priced for perfection and deeply sensitive to anything that threatens the narrative. That fragility is its own risk, independent of whether the underlying thesis is right or wrong.
Both things are true. Neither changes my position.
Samsung's chip division just posted the most profitable quarter in its history. SK Hynix is sold out of HBM. Big Tech is offering to fund their factories and getting turned away. Workers are demanding annual profit-sharing because they believe this lasts a decade.
A Facebook post isn't going to change any of that.
The irony of it all? If a citizen dividend gets distributed to ordinary Koreans, most of them will probably deploy it straight back into Samsung and SK Hynix shares.
I didn't sell. I'm watching.
I hold a long position in DRAM ETF (ticker: DRAM). Not investment advice. Do your own research.
My current positions as of May 12, 2026.