Micron's Earnings Today Put a Trillion-Dollar AI Memory Bet to the Test
Micron Technology reports fiscal fourth-quarter results after markets close today, and the numbers Wall Street is already penciling in look less like a chip forecast than a rounding error from a different company. Analysts polled by Yahoo Finance and Benzinga put consensus revenue at roughly $50.75 billion to $51.47 billion, up from $11.31 billion a year earlier, with adjusted earnings per share near $31.6, up from $3.03. Options markets are pricing an approximate 6% move in the stock on the print, itself a sign of how much has already been baked into a share price that has risen 273% so far in 2026, according to Yahoo Finance.
The driver is high-bandwidth memory, the specialized DRAM stacked next to AI accelerators to feed them data fast enough to avoid starving expensive GPU capacity. According to 24/7 Wall St, Micron's HBM output is sold out through 2027, backed by roughly $100 billion in take-or-pay contracts, $22 billion in customer deposits and letters of credit, and 16 Strategic Customer Agreements locking in volume. That backlog helped push Micron's non-GAAP gross margin to 84.9% in the June quarter on $41.46 billion in revenue, up 345.72% year over year — and it is why the September quarter guide jumped to roughly $50 billion even before today's actual print.
"The memory industry has been structurally transformed by the proliferation of AI" — Sanjay Mehrotra, CEO, Micron Technology
What makes this earnings report more than a single-company event is the valuation math sitting underneath it. Despite the growth rates, Micron trades at roughly 24 times forward earnings, according to Benzinga's compilation of sell-side estimates — cheap relative to its own growth, but only if the shortage persists. JPMorgan's Harlan Sur and Bank of America's Vivek Arya carry price targets near $1,540 and $1,550 respectively, with the broader analyst consensus averaging about $1,541, per Benzinga's tally. Bank of America has floated fiscal 2027 EPS potential of $150 to $200 a share, a figure that only works if DRAM and NAND pricing keeps climbing at the roughly 20% sequential clip JPMorgan has modeled and Mehrotra has said the company cannot fully satisfy — filling, in his words relayed by 24/7 Wall St, only "half to two-thirds" of its largest customers' AI memory demand.
That gap between demand and supply is the entire investment case, and also its biggest vulnerability. Memory has historically been the most cyclical corner of semiconductors, prone to violent boom-bust swings whenever Samsung, SK Hynix or newer entrants like China's CXMT add capacity into a shortage. Wellington-Altus chief market strategist James Thorne framed the stakes bluntly, telling Yahoo Finance the real debate is "whether $180 in FY28 at 80%+ margins is a peak to fade or still early" — a question today's guidance, and specifically any commentary on fiscal 2027 capital spending, is likely to move meaningfully in one direction or the other.
For the AI industry more broadly, Micron's report is a useful counterweight to the GPU-centric narrative that has dominated the sector's financial coverage. It shows the AI capex boom flowing not just to accelerator makers but through the entire physical supply chain — memory, packaging, power — in ways that are only now becoming visible in hard numbers rather than announced spending plans. Whether that flow-through holds at current multiples depends less on today's headline beat, which most estimates already assume is coming, than on whether Micron's management signals the shortage extends meaningfully into calendar 2027 or shows early signs of the capacity race catching up to demand.