Micron Technology just posted fiscal Q2 2026 revenue of $23.9 billion, nearly triple what it pulled in during the same period a year earlier. Record gross margins of 75%. A non-GAAP earnings per share of $12.20. And the company isn’t slowing down. Its Q3 guidance projects revenues of roughly $33.5 billion at an 81% gross margin.

The AI hunger games for memory chips

The driving force behind all of this is artificial intelligence, specifically the data centers powering it. Micron’s entire HBM capacity is sold out through 2026. The company currently meets only 50-67% of customer demand.

SK Hynix and Samsung, the other two pillars of the global memory industry, are in the same position. All three manufacturers have their HBM production fully booked through 2026. Each is pouring over $20 billion annually into capacity expansions.

The global memory market reflects this intensity. Projections estimate it will reach $551.6 billion in 2026, then swell to $843 billion in 2027. That’s a 53% jump in a single year. AI data centers alone are expected to consume roughly 70% of the world’s total memory chip supply, a historic shift away from traditional drivers like smartphones and PCs.

Why analysts think this time is different

D.A. Davidson put a $1,000 price target on Micron by 2026, a figure that captures the bullish sentiment around what the firm explicitly calls an “AI-driven memory supercycle.”

When three companies are each spending $20 billion-plus per year on new capacity, the math eventually catches up. Those fabs will produce chips eventually, and if AI spending cools, even modestly, the industry could find itself staring at oversupply faster than anyone expects.

What this means for investors

With Micron, SK Hynix, and Samsung all capacity-constrained, there’s limited risk of one player undercutting on price to grab share. When everyone is sold out, nobody needs to compete on price.

Investors should watch two things closely. First, the pace of new fab construction and when that capacity actually hits the market. Second, any signs of softening in AI capital expenditure from the hyperscalers — the Microsofts, Googles, and Amazons of the world — whose spending decisions will determine whether the supercycle label holds.