Apple eyes AI memory fix that could dent Micron's boom

Apple has been raising device prices to cope with surging memory chip costs. The company increased computer prices by $200 or more and is considering similar hikes for iPhones, all because demand for memory chips has sharply risen as tech giants build out massive AI data centers. To avoid becoming permanently dependent on price increases—which risk alienating customers—Apple is exploring alternative solutions.

According to a recent CNBC report, Apple is in talks with PrismML, an AI startup that claims its technology can reduce memory usage for AI models by up to 15x while delivering responses up to 8x faster. The appeal is obvious: if the technology works as advertised, Apple could deliver advanced on-device AI features—such as improved Siri capabilities—without requiring expensive memory upgrades. However, no concrete deal or partnership between the two companies has been announced yet.

The implications for Micron Technology, a major memory chip supplier, are potentially significant but not imminent. Micron has ridden the AI data center boom to extraordinary profitability. Memory demand for AI infrastructure created a shortage of memory processors, allowing Micron to command premium prices. The company's non-GAAP earnings per share soared more than 1,200% to $25.11 per share in Q3 2026, and its profit margins have expanded to an enviable 74%. For Micron investors, this raises a question: can such growth last? If memory-efficiency technologies like PrismML's become widely adopted across the industry, demand for memory chips could moderate, threatening Micron's margins and growth rate.

However, several factors may cushion the blow. First, for such efficiency gains to meaningfully impact Micron's business, they would need to be implemented across a wide range of devices from many different tech companies—a slow process. Second, and more importantly, the overall memory processor market is still projected to expand dramatically, rising from $230 billion（約37兆円） in 2025 to more than $1 trillion（約160兆円） next year. Micron's management expects autonomous vehicles and robotics to drive a new wave of memory demand. In short, even if on-device AI models become more memory-efficient, the broader market tailwind could sustain high demand for years.

Apple has found itself caught between two pressures: the need to maintain healthy profit margins and the rising cost of memory chips driving up production expenses. Rather than absorb these costs, the company has chosen to pass them on to consumers—raising computer prices by $200 or more and signaling potential iPhone price increases. However, this strategy has an obvious limit: customers will eventually balk at large price jumps. This is why Apple's exploration of PrismML's memory-efficiency technology matters. If the startup's claims hold up—reducing memory demands by 15x while improving response speed—it could offer Apple a way to deliver advanced on-device AI features (like improved Siri) without expensive memory upgrades, ultimately preserving margins without further price hikes.

For Micron, the timing is bittersweet. The company has enjoyed an extraordinary windfall from the AI boom, with memory demand for data centers creating a genuine shortage that allowed it to raise prices and expand margins to an enviable 74%. That earnings growth—more than 1,200% year-on-year—looks spectacular on the surface. But the article's author correctly notes that such growth is unlikely to be sustainable. If major tech companies systematically adopt memory-efficiency technologies, it could eventually dampen the demand surge that has fueled Micron's recent gains. That said, the threat is not immediate: such technology would need to be implemented widely across many different companies and devices to materially impact Micron. Moreover, the overall memory market is still projected to expand dramatically—to more than $1 trillion（約160兆円） next year—and Micron's management expects autonomous vehicles and robotics to spark new demand waves that could keep the shortage alive for years.