Reading time: 6 min
Kioxia overtook Toyota last Friday to become the most valuable listed company in Japan, 18 months after its IPO. Every wire ran the same line: AI mania. The detail they buried is sharper. The Middle East shock weighing on Toyota is part of the same macro backdrop lifting Kioxia, the two pulling in opposite directions on one exchange, and the company now sitting at the top of Japan Inc. just refused to forecast its own next year.
That refusal is the story. Hold it for a second, because it changes how you read the milestone.
The Swap, In Numbers
Kioxia Holdings, the NAND flash memory maker spun out of Toshiba in 2018, saw its shares jump 7.6% on Friday, lifting its market value above ¥44 trillion, around $274 billion, per Bloomberg. Toyota closed at ¥43.8 trillion. The figures include treasury shares, and the gap narrowed by the bell, but the ranking flipped and the symbolism is hard to miss: a memory chipmaker that held its Tokyo IPO in December 2024 is now worth more than the automaker that anchored Japan’s corporate hierarchy for over two decades.
The run underneath it is close to absurd. Kioxia is up more than 660% since January, which makes it the single best performer on the entire MSCI World Index this year, per Startup Fortune’s reading of the data. That is not a sector rotation. That is a repricing of what investors think Japan’s economy is for.
And it isn’t just one name. SoftBank passed Toyota earlier this year on enthusiasm around OpenAI’s planned listing before slipping back, and Japan’s top-20 now carries AI-data-center suppliers like Murata and chip-tester Advantest. The hierarchy is being rewritten in favor of anything that feeds the AI buildout.
One Macro Backdrop, Two Opposite Trades
Here is the connection the wire coverage left on the table. Toyota has had a genuinely rough 2026: it guided to a 21% full-year operating profit decline for the year to March 2026, citing US tariffs, a stronger yen and higher material costs, per Reuters, and its global sales fell for a third straight month into April, with Middle East volumes down a steep 33.7% year on year as regional tensions and oil weighed on demand. The stock has slid well off its February high near $244 toward the low end of its 52-week range. So the picture isn’t one clean cause, but the Middle East is a real thread running through it.
Now hold that against Kioxia. The same week it took the crown, the broader Asian tape was rallying on reports of a temporary US-Iran deal to reopen the Strait of Hormuz, which softened oil and lifted sentiment, as Thailand’s Asia Plus Securities noted in its Monday outlook. The macro forces pressuring the old-economy champion and the ones rewarding the AI-infrastructure name are overlapping, not identical, and they’re pulling in opposite directions on the same exchange in the same quarter. That’s the part worth sitting with: a single regional shock doesn’t cleanly explain both, but the market is plainly using the same macro backdrop to sell the car company and buy the memory company. When that’s the mechanism, a durable Hormuz reopening that lifts the index broadly could also start unwinding the exact rotation that produced this milestone. Watch Toyota on any real de-escalation, not just Kioxia.
This is the regional version of a pattern we flagged last week, when Broadcom beat earnings and Asia’s AI trade sold off anyway. Positioning, not fundamentals, is increasingly setting the tape across the region’s chip complex.
The Earnings Are Real. That’s What Makes the Hedge Loud.
The bears’ easy line is that this is another valuation bubble detached from profit. It isn’t, and that’s precisely why the next part matters. Kioxia is forecasting net profit of ¥869 billion for the April-June quarter, a 48-fold jump year on year, on operating profit guidance of ¥1.298 trillion that blew past expectations, per BigGo Finance. The quarter before, US-dollar NAND selling prices doubled in a single three-month span while shipments actually fell about 10%, which is the textbook signature of pricing power, not volume growth.
The cause is a genuine supply-demand gap. AI data centers training and running large models need fast, high-capacity storage, and the handful of firms with enterprise-grade NAND capacity are extracting record prices from a shortage. Kioxia’s own management has said the market stays very tight through 2026 and 2027, with major new capacity not arriving until late 2027 at the earliest. Micron’s $24 billion Singapore fab, for scale, isn’t expected to output until the second half of 2028. The shortage has a runway.
So with profits compounding and a multi-year shortage in hand, what did the most valuable company in Japan do? It declined to give full-year guidance for the fiscal year ending March 2027, citing geopolitical uncertainty. Sit with that. A business projecting a 48-fold quarterly profit surge, sitting on pricing power its executives expect to last into 2027, will not put a number on the next twelve months. The earnings are the loudest argument for the stock, and the company’s own guidance is the loudest argument for caution. Both are true at once.
Why “Most Valuable” Is the Fragile Part
NAND is historically one of the most violently cyclical markets in all of semiconductors. Kioxia itself was caught in the last oversupply glut, from roughly 2022 into 2023, badly enough that it delayed this very IPO by years. Samsung, SK Hynix and Micron have all ridden the same boom-bust wheel. A 660% move since January has priced in a sustained shortage as the base case, which means the open question isn’t whether AI demand is real. It plainly is. The question is whether AI-driven demand is structurally different from the consumer and PC cycles that came before, and whether it holds long enough to outlast the new capacity landing in late 2027.
There’s a tell worth watching. Kioxia has disclosed preparations for a US ADS listing, a move that would open the stock to American investors and could pull its valuation closer to US-listed peers like Micron and SanDisk. That’s a real upside catalyst. It’s also exactly what you’d expect a company to pursue while its multiple is at a cyclical high. Both readings fit.
What To Actually Watch
Three things, into this week and beyond. First, the Toyota side of the trade: if the Hormuz reopening holds and oil keeps easing, the carmaker’s beaten-down shares are where a sentiment reversal would show up first, and that would tell you how much of this milestone was AI strength versus oil-driven rotation. Second, NAND contract prices into the second half: the entire thesis rests on the shortage outlasting 2027 capacity, so the pricing curve is the fundamental to track, not the share price. Third, the US listing timeline, which is the catalyst that could extend the run and the signal that the smart money wants liquidity near the top.
The headline is that Japan crowned a new king. The real event is subtler and more useful to understand: an overlapping set of macro forces, the Middle East among them, is bleeding the old economy while financing the new one, and the new king is quietly telling you it can’t see twelve months ahead. For the broader picture of how this AI-versus-everything-else split is playing out in markets, our running coverage of the equity story keeps tracking where the money is actually moving. Right now, in Japan, it moved from the thing you drive to the thing that remembers.
