# Kioxia's Market Value Has Nearly Halved Since Its AI-Fueled Peak in June > Source: > Published: 2026-07-17 01:11:35+00:00 *Kioxia was Japan's hottest stock this year, briefly worth more than Toyota. Three weeks later, nearly forty percent of that value is gone.* You don't often see a company go from Toshiba castoff to Japan's most valuable firm and back to also-ran inside a single quarter. Kioxia just did it. The NAND flash maker's shares have fallen roughly 38.7% from their post-listing peak of ¥112,700, hit on June 22, according to Bloomberg's tracking of the stock. The trigger wasn't a bad earnings report or a product recall. It was a single word: resell. On July 1, CNBC and The Information reported that Meta is building an internal cloud unit called Meta Compute, designed to sell its excess AI computing capacity to outside customers. The next day, Kioxia shares slumped 13.47%, according to Reuters. The logic is blunt: if Meta has spare compute to rent out, maybe the AI buildout isn't as insatiable as Wall Street assumed. Every stock riding the infinite-demand story got hit at once. Micron, Intel and AMD all fell hard, and the Philadelphia Semiconductor Index dropped more than 10% over two sessions. Then came the second blow. The Information reported that OpenAI engineers had cut inference costs in half using pure software optimization, no new chips required. OpenAI is also working with Broadcom on a custom chip called Jalapeño, aimed at cutting inference costs by another 50% when it ships later this year. For a memory maker whose entire investment case rests on AI models needing ever more storage and bandwidth, that's an uncomfortable headline. Cheaper inference can mean less hardware, not more. ## How high Kioxia flew Context matters here. Kioxia didn't drift up. It went vertical. Shares gained more than 670% in 2026 through June, according to Bloomberg, making it the top performer on the entire MSCI World Index. On June 12, the stock jumped 7.6% in a single session, pushing its market cap past ¥44 trillion and knocking Toyota Motor off the top of Japan's corporate rankings for the first time in over two decades, Bloomberg reported. First-quarter fiscal 2026 revenue hit roughly ¥1 trillion, up 189% year over year, with operating profit up fifteenfold. Those aren't hype numbers. They're real. That's what makes the reversal notable. Toyota didn't actually reclaim the crown, though. Mitsubishi UFJ Financial Group did. On July 13, MUFG's market cap hit ¥41.9 trillion, edging past Toyota's ¥40.87 trillion and Kioxia's ¥36.58 trillion, according to Bloomberg and the Seoul Economic Daily. A Japanese bank hasn't topped the rankings since Sumitomo Bank in 1986. Rising rates put a bank on top. AI mania put a chipmaker there first. Neither Kioxia nor Toyota holds the position now. ## A sector-wide reckoning Kioxia has plenty of company in the pain. Micron fell 13% in a single session in early July, erasing about $138 billion in market value, after a South Korean brokerage cut its profit estimate for SK Hynix on weaker than expected HBM4 shipments, Yahoo Finance reported. Micron dropped again on July 15, this time on news that Chinese memory maker CXMT plans an $8.5 billion IPO to fund domestic DRAM expansion, plus reports that Washington is weighing tighter export restrictions on high-bandwidth memory. Samsung and SK Hynix haven't been spared either. This isn't a Kioxia problem. It's a memory sector repricing. Here's the thing: none of this means Kioxia's business is broken. Forward earnings still price the stock at roughly 11.5 times, according to GuruFocus, cheap for a company that just posted 189% revenue growth. Some analysts remain bullish on the medium term, betting that AI inference workloads will keep leaning on enterprise SSDs for KV cache storage regardless of how efficient the chips underneath get. That's a real structural driver, not a hypothetical one. What the last three weeks actually expose is how thin the margin for doubt has become in the AI bellwether trade. A stock that rallies 670% doesn't need bad news to fall. It just needs the market to stop being certain the good news will keep compounding at the same pace. Meta's cloud pivot and OpenAI's cost cuts didn't prove AI demand is slowing. They gave investors permission to ask the question out loud. For a stock priced for perfection, that was enough. **Also read:** [A Regulatory Filing Reveals DeepSeek Is Now Worth About $52 Billion](https://startupfortune.com/a-regulatory-filing-reveals-deepseek-is-now-worth-about-52-billion/) • [Moonshot's Kimi K3 Becomes the Largest Open-Weight Model Ever Built](https://startupfortune.com/moonshots-kimi-k3-becomes-the-largest-open-weight-model-ever-built/) • [Meta puts Muse Spark 1.1 on OpenRouter, then locks out the rest of the world](https://startupfortune.com/meta-puts-muse-spark-11-on-openrouter-then-locks-out-the-rest-of-the-world/)