Micron's earnings are forecast to surge ~987% amid an AI-driven memory supercycle, with FY2025 revenue tracking at $37.4B (+48.9% YoY) and diluted EPS of $7.59. The massive growth setup is already partially priced in, creating a tension between whether the AI tailwind sustains or the print disappoints a high-expectation bar.
Micron's earnings are forecast to surge ~987% amid an AI-driven memory supercycle, with FY2025 revenue tracking at $37.4B (+48.9% YoY) and diluted EPS of $7.59.
MU is tracking toward a near-10x EPS surge on AI memory demand, but the question is whether that growth is already priced in or whether HBM execution and hyperscaler demand sustain a further re-rating into and through the print.
A hyperscaler capex pause, softer-than-expected HBM pricing commentary, or NAND oversupply signals on the earnings call would unwind the thesis quickly; 'sell the news' risk is elevated when consensus framing is this bullish pre-print.
CoverageSource: Yahoo Finance · Published here MON, JUN 22 · 4:44 PM ET · the only report in this recordHow this is decided →
Micron Technology is heading into its next earnings print with analyst consensus pointing to a ~987% year-over-year EPS surge, underpinned by explosive demand for HBM (High Bandwidth Memory) used in AI accelerators like Nvidia's H100 and Blackwell series. The enrichment data confirms the trajectory is real: FY2025 revenue is tracking at $37.4B, up nearly 49% YoY, with gross margins at 39.8% and net margins recovering to 22.8%.
The AI megatrend is the core driver — HBM3E content per GPU server is rising, Micron is ramping HBM3E production, and hyperscaler capex budgets remain robust. MU is the primary U.S.-listed pure-play on DRAM and NAND, making it the most direct expression of the memory upcycle thesis.
The bull case is straightforward: if AI server buildouts continue at pace, HBM pricing holds, and Micron executes on its HBM3E ramp, earnings revisions likely push higher from an already elevated base. A 987% EPS growth print, if confirmed, would validate the cycle thesis and could catalyze further multiple expansion.
The bear case is harder to dismiss: memory is a notoriously cyclical business, and 987% EPS growth implies the comparison period (FY2024) was a trough. Markets discount known cycles — MU already trades well off its lows, and any softness in NAND pricing, a hyperscaler capex pause, or HBM yield disappointment could reset expectations sharply. The headline itself is a consensus-framing piece, which often marks a late-stage pricing of good news.
Watch the upcoming earnings call for HBM allocation guidance, pricing commentary, and any signals on FY2026 demand visibility — those are the real catalysts that determine whether the stock re-rates further or stalls on 'sell the news.'
FY2025 revenue of $37.4B (+48.9% YoY) and a recovering 39.8% gross margin confirm the cycle is real and not just consensus hype; HBM3E ramp positions MU as the primary U.S. beneficiary of AI accelerator memory demand. However, the 987% EPS growth framing is consensus-known, so the trade is predicated on forward guidance upside — particularly HBM pricing and allocation — rather than the backward-looking print itself.
The read above, as written. kept as written · closes shown from JUN 23 on
Into earnings print, 2-4 weeks. Follow to be told when one lands.
Price context does not establish that the story caused the move.
With FY2025 revenue up ~49% YoY and HBM3E production ramping into a supply-constrained AI server market, there is a credible path to further earnings estimate revisions higher if Micron's HBM allocation guidance on the call exceeds Street models.
Memory is a commodity cycle business, and 987% EPS growth is largely a trough-comparison artifact — MU has already re-rated substantially off its lows, meaning the growth is consensus-known and any guidance in-line with expectations could trigger a 'sell the news' reaction, particularly if NAND pricing remains soft outside HBM.
Kept as written · your side, if you take one, is graded privately against licensed closes after 10 trading days · nothing here is advice · How the Wire is made →