Micron Technology guided next-quarter revenue to $61.5 billion, plus or minus $1.5 billion, the US memory chipmaker told investors this week. The midpoint runs $4.48 billion above the $57.02 billion consensus tracked by LSEG. Adjusted profit guidance came in at $38.15 per share, plus or minus $1. The chipmaker also topped expectations for the quarter it just reported.

Why did guidance run so far above consensus?

The upside reflects tightness in high-bandwidth memory (HBM), the DRAM variant that generative AI has turned into essential data-center hardware. HBM feeds data to AI chips fast enough to keep them working at full tilt. Reuters reported that orders are "far exceeding capacity." As one of a handful of producers, Micron faces buyers with limited substitutes.

How does HBM differ from conventional memory?

HBM packages memory in configurations designed for sustained high bandwidth, the workload profile generative AI training and inference demand. Conventional DDR cannot sustain these workloads at acceptable throughput, which is why hyperscalers treat HBM allocations as binding constraints on AI server rollouts. The product thus sits in a higher price tier than standardized DRAM.

What does the $250 billion US investment plan cover?

Micron lifted planned US investment to more than $250 billion through 2035, extending a multi-year envelope that funds domestic fab build-out, supplier ecosystem development, and advanced packaging capacity. Big cloud firms including Amazon and Alphabet continue spending heavily on AI infrastructure, anchoring demand visibility through the second half of the forecast window.

Can supply catch up?

Not quickly. Adding memory supply takes years: new factories, specialized equipment, and advanced packaging capacity each require extended lead times. When demand outruns supply, memory prices rise and profits can climb quickly because buyers have limited alternatives. Micron's capex commitment signals that the company expects tight conditions to persist into the latter half of the decade.

Will HBM escape the industry's boom-bust cycle?

Not entirely. The pattern is familiar: chipmakers invest, capacity catches up, and competition shifts from delivery to price — especially for standardized memory chips. HBM is harder to make, so tightness can last longer, but it still depends on adding wafers and packaging capacity. Earnings swing more than in less commodity-like semiconductor segments.

What does the outlook signal for buyers and investors?

The guide signals strength today while embedding the structural risk that punishes memory investors: capacity added during a boom arrives just as demand softens. The $250 billion US investment plan extends that risk on a longer timeline. For chip buyers, the implication is straightforward — HBM allocations, and the AI server schedules they gate, will remain tight through the rest of the decade.