NextFin News - Kioxia Holdings briefly surpassed Toyota Motor Corp. in market value on June 3, and by the middle of this month the memory-chip maker had become one of Japan’s most valuable listed companies. For a company that listed only 18 months ago, the move is extreme: the shares jumped as much as 7.2% in one session, and market capitalization topped ¥45 trillion, about $281 billion, at one point.
On the surface this looks like an AI stock rally; the real issue is a repricing of which parts of Japan Inc. investors think can capture the next round of capital spending. Kioxia, built from Toshiba’s NAND flash business, has become a direct way to buy the view that larger models and heavier inference workloads require more storage and bandwidth inside data centers. Its stock is up more than 660% this year, enough to displace the country’s old market-value order and turn a memory supplier into a flagship trade.
What really changed is not national symbolism but the market’s preferred profit engine. Toyota represented an earnings model tied to autos, industrial scale and export cycles; Kioxia represents a model tied to AI servers, memory intensity and component pricing. Kioxia is not about semiconductors in the abstract — it’s about whether scarce memory and storage can command stronger pricing power than traditional manufacturers can. SoftBank Group moving ahead of Toyota at times points to the same conclusion: investors are paying up for exposure to AI build-out rather than for domestic industrial leadership.
Jumpei Tanaka of Pictet Asset Management Japan argued that the reshuffling could change how overseas investors classify Japanese equities, from a market led by manufacturing-driven cyclicals to one increasingly led by AI-driven semiconductor growth stocks. The logic holds up, but only partly. Foreign investors do tend to reward markets that produce clear AI beneficiaries, and Kioxia fits that screen better than many Japanese blue chips. But the math doesn’t add up yet if the broader conclusion is that Japan has structurally changed because one memory name and SoftBank have rerated faster than autos. Whether that view lasts depends on whether earnings, not just positioning, can verify the new multiples.
The pressure point is clear. Kioxia’s valuation already assumes AI-related memory demand stays tight longer than in prior chip cycles, which would require hyperscale cloud spending to remain strong and supply discipline to hold. The real trade-off is between genuine scarcity today and the memory industry’s long history of destroying pricing when capacity comes on too fast. The risk nobody is talking about is not simply a pullback in sentiment; it is that a sector known for violent boom-bust swings is being priced as if this cycle will behave differently. If capital expenditure slows, if new supply arrives faster than expected, or if pricing pressure returns, a 660%-plus year-to-date gain leaves very little room for error.
Still, the market is not hallucinating demand. Kioxia sits in the part of the semiconductor value chain that AI has made newly scarce, which is why it could leapfrog older industrial giants so quickly. What remains to be verified is whether this is a durable shift in business quality or a peak-cycle valuation attached to the right narrative at the right moment. The answer will not come from headlines about overtaking Toyota; it will come from whether a ¥45 trillion company can turn AI enthusiasm into earnings support at something close to its current price.
Explore more exclusive insights at nextfin.ai.
