Summary
- The Bank of Korea will almost certainly hold its 2.5% rate on May 28, with market chatter about Q3 hikes.
- That framing misses the real driver: KOSPI has surged past 7,800 - up 70%+ year-to-date - on an AI memory chip boom centered on Samsung and SK Hynix.
- SK Hynix posted a 72% operating margin in Q1 2026. Samsung's Q1 operating profit hit ₩57.2 trillion. These are not cyclical numbers.
- The won is weakening at 1,517 per dollar despite BoK intervention threats. A rate hike would strengthen it - and squeeze the very export margins fueling this rally.
- At these levels, the risk/reward on Korea exposure has shifted. The AI thesis is real, but the market has already done the work.
The consensus narrative is about to turn mechanical: the Bank of Korea holds rates on May 28, then hikes in Q3 to fight inflation ticking back above 2%. The logic is textbook emerging-market macro. I've been very surprised that anyone is actually framing this as a policy story when the single most important force shaping the Korean market has nothing to do with interest rates.
Let me be direct: the KOSPI is up more than 70% year-to-date because Samsung Electronics and SK Hynix are riding the biggest structural demand shock in their history, and the market doesn't yet realize how much of that it's already priced in.
Samsung and SK Hynix dominate the KOSPI index. They make up a large portion of the KOSPI index. When SK Hynix reported a 72% operating margin in the first quarter of 2026 - among the highest of any semiconductor company globally - and Samsung posted ₩57.2 trillion in operating profit on ₩133.9 trillion in revenue, the index didn't just move. It went ballistic. KOSPI broke 7,000 in early May, pushed past 7,800 by mid-month, with SK Hynix gaining more than 10% in a single session on May 10 alone.
This is AI cross-pollination in its purest form. Every dollar that NVIDIA, Google, Amazon, and Microsoft spend on HBM - high-bandwidth memory, the specialized memory module that makes AI chips actually function - flows through Samsung and SK Hynix. These companies aren't beneficiaries of the AI cycle; they're the plumbing. Both have warned that severe HBM shortages could persist through 2027. That's not guidance - that's a supply constraint telling you the earnings run hasn't peaked yet.
However, and this is where the false narrative does the most damage: if the BoK actually follows through on rate hikes in the second half of 2026, it will strengthen the won. The won is already under pressure, weakening 2.96% against the dollar over the past month to trade at 1,517 per dollar. South Korean authorities warned on May 22 that they would "take decisive action" against excessive won moves. A stronger won compresses the dollar-denominated margins of Korea's two largest companies. The monetary policy solution to a 2.6% inflation rate - the reading from April, a two-year high - would directly penalize the export engine driving the entire market.
That's the structural tension nobody is pricing in.
The inflation picture is genuinely concerning. Korea's consumer price index accelerated from 2.2% in March to 2.6% in April, driven by housing costs and energy prices. ING and other institutions expect the BoK to act in the second half of 2026. But rate hikes in an export economy where the top two companies are memory chip manufacturers are not a clean win. They're a trade between domestic price stability and the currency advantage that makes Samsung and SK Hynix so profitable.
Here's what I think the market is missing. South Korea ETFs like EWY have returned more than 140% over the last twelve months. A 3x leveraged South Korea ETF has been launched to chase the momentum. Foreign-held market cap rose nearly 87% year-to-date. At this point, the question isn't whether the AI thesis is real. It's whether there's any remaining asymmetry left in the trade.
The HBM supply shortage extending to 2027 gives me confidence that earnings power will hold. SK Hynix's 72% operating margin and Samsung's ₩57.2 trillion profit base are structural, not seasonal. But I also see a stretched market where a currency move, a US tariff shift, or a BoK pivot could unwind a meaningful chunk of the 140% return in a quarter rather than a year.
That being the case, I rate Korea equity exposure as a Hold at current levels. For investors who already have positions built from last year, the AI earnings runway through 2027 provides a cushion. But for new money, the entry has moved. The AI infrastructure thesis remains one of the strongest secular trends in global markets, but you don't need the leveraged won-risk of concentrated Korea exposure when the underlying theme is already available through broader semiconductor and technology positions.
The Bank of Korea's May 28 decision will be a hold. The Q3 hike chatter will grow. But the story isn't the rate. The story is whether the KOSPI can justify 7,800 when the entire rally rests on two companies whose margins depend on a weak currency - and whose central bank is being asked to strengthen it.
