Micron Technology’s latest earnings were not just another strong AI-chip result, but a warning that the next big shortage in artificial intelligence may be hiding in memory.
The company reported record fiscal third-quarter revenue of $41.5 billion and adjusted profit of $25.11 per share, both comfortably ahead of Wall Street estimates.
Its shares jumped 12% in after-hours trading, extending a rally that has already pushed Micron’s market value above $1 trillion.
But the real story was the signal from customers: they are no longer just buying memory chips. They are trying to lock up future supply before everyone else does.
AI’s next shortage is hiding in memory chips
For years, the AI trade has revolved around Nvidia and the graphics processors needed to train and run large models.
Micron’s results suggest investors may now need to widen the frame.
AI models need more than processors. They need fast memory to move and store huge amounts of data.
That is where high-bandwidth memory, or HBM, comes in.
These chips sit alongside advanced AI processors and help feed them data at the speed modern AI workloads require.
Micron said that customers had committed $22 billion to secure memory-chip supply under agreements with 16 strategic customers across data centres, consumer devices and autos.
That number matters because it shows how memory is being increasingly treated as a strategic infrastructure.
Daniel Newman, CEO of Futurum Group, told Reuters that the scale of the AI buildout has been underestimated, adding that memory should continue to command “premium pricing” while supply remains constrained.
Micron is no longer just a cyclical chip story
Memory has historically been one of the most boom-and-bust corners of the chip industry.
As supply tightens, chipmakers raise prices and expand capacity, but once supply catches up, prices and margins typically come under pressure.
Micron is trying to change that story.
The company’s new customer agreements include take-or-pay commitments, cash deposits and pricing floors.
In plain English, customers are committing money and volume ahead of time, while Micron gets better visibility on demand and some protection if the market turns.
As per the company, the remaining performance obligations (RPOs) tied to these agreements are around $100 billion.
That gives investors a clearer view of future contracted revenue than they would normally expect from a memory company.
That is why investors are re-rating Micron: customers are not just buying for today’s demand, but securing future supply to avoid being caught short.
Art Hogan, chief market strategist at B. Riley Wealth, told Reuters that pure memory demand had risen rapidly and that Micron “sits at the center” of that shift.
He also described the company’s trillion-dollar valuation milestone as an “exclamation point” on the demand required to run AI data centres.
Wall Street’s new question: is the memory cycle still early?
The sharpest post-earnings reaction came from D.A. Davidson analyst Gil Luria, who raised his price target on Micron to $2,000 from $1,500, setting a new Street-high target while keeping a Buy rating.
His argument goes to the heart of the rerating debate.
Luria said Micron now has “some of the semi industry’s best visibility”, helped by long-term strategic customer agreements that give the company a clearer view of future demand than investors normally expect from a memory-chip maker.
The analyst pushed back against the idea that Micron is already near the top of the cycle.
He argued that “the memory cycle is far from over”, with tight supply-demand dynamics likely to last through at least calendar 2027, even as Micron spends heavily on capacity.
The post Micron’s record-breaking quarter reveals AI’s next trillion-dollar bottleneck appeared first on Invezz