Micron Q4 FY2026 Earnings, September 2026: Data Center Revenue Up 11-Fold as Legal and Competitive Pressures Mount

Micron's fiscal Q4 2026 results beat earnings and profit estimates with data center revenue rising eleven-fold year-over-year and a $22 billion order backlog, but two simultaneous patent actions and an accelerating Chinese challenger introduce material risk that the headline numbers alone do not capture.

Micron Technology reported its fiscal fourth-quarter 2026 results on September 30, delivering earnings and profit that beat analyst expectations alongside guidance that surpassed consensus. The standout figure was data center revenue, which rose eleven-fold year-over-year — a rate that, even discounted for a low comparable base, reflects a structural reorientation of Micron's revenue mix toward AI infrastructure demand. Heading into the print, the options market had priced in a sharp move and multiple analysts had flagged HBM4 certification progress and pricing guidance as the two variables the market would most closely parse. The results appear to have cleared the bar set by those expectations.

The backdrop is one of acute and sustained scarcity. Coverage reviewed by Slicast indicates that high-bandwidth memory capacity is sold out through 2027, with Micron entering its earnings call holding a $22 billion backlog and a fully subscribed HBM order book. JPMorgan, according to reporting cited in coverage, forecasts that the HBM shortage will persist for at least two more years, driven by strong accelerator demand. Micron is doubling its HBM capacity in response, a commitment already visible in the capital account: fiscal 2025 capital expenditure reached $15.86 billion, up 89.1% from $8.39 billion in fiscal 2024, representing 42% of that year's $37.38 billion in revenue. Slicast's compilation of SEC XBRL filings across chip-sector peers places Micron first by both absolute spend and capex intensity — a distinction that measures resolve as much as it does cash outflow.

In the HBM4 supply allocation for NVIDIA's next-generation accelerator program, SK Hynix holds an estimated 65–70% share of the 16-layer contract and has already commenced shipments for the Rubin architecture, according to September coverage. Micron and Samsung occupy the remaining allocation. One analyst perspective in coverage argues that Micron can create commercial value by holding a disciplined position in HBM while capturing disproportionate share in conventional DRAM, where its revenue grew 65.5% quarter-over-quarter in Q2 2026 — the fastest rate among the three memory incumbents, according to TrendForce data cited in coverage. Product decisions align with that reading: Micron discontinued its 2GB GDDR7 chip line in September to redirect wafer capacity toward higher-density 3GB configurations for AI and data-center applications, and introduced 512GB DDR5-9200 RDIMM modules claiming 60% better energy efficiency than competing configurations — density and power-consumption figures aimed directly at hyperscale infrastructure procurement.

Two simultaneous legal actions inject headline risk. Netlist filed a new International Trade Commission complaint against Micron in September alleging infringement of HBM patents covering components used in AI computing systems by Google and NVIDIA, and is seeking a U.S. import ban on the relevant chips. A report cited in coverage noted that such a ban, if granted, could disrupt AI accelerator supply chains for both hyperscalers. Separately, YMTC — the Chinese memory manufacturer — prevailed in a Munich court in a 3D NAND patent dispute against Micron, extending a cross-border litigation campaign now in its third year. Neither case has yet produced an enforceable outcome; but the combination places Micron in simultaneous defensive legal posture on two fronts precisely as its production ramp requires uninterrupted supply continuity.

The competitive threat from China's domestic memory industry operates on a longer timeline but is accelerating. CXMT's conventional DRAM revenue rose 99.3% quarter-over-quarter in Q2 2026 — faster than any incumbent — and one September report indicated CXMT had surpassed SK Hynix and Micron in profit margins. Futurum analysts cited in coverage maintain that Samsung's and Micron's technology lead over CXMT is unlikely to narrow soon, but acknowledge the gap is closing faster than previously modeled. On the demand side, DeepSeek's most recent inference architecture reportedly reduced KV-cache HBM requirements by 75% and SSD requirements by 87.5% — a reminder that memory intensity at the frontier is not fixed and that efficiency gains at the model layer can alter the supply-demand equation independently of macroeconomic conditions. A UBS analyst in September described Micron as a "battleground" stock caught between near-term AI momentum and the structural effect of long-term supply contracts potentially muting revenue upside on the way up.

Three signals will determine whether the eleven-fold data center revenue figure marks a durable new baseline or a cycle peak. First, the trajectory of the Netlist ITC complaint: whether it advances to a formal investigation sets the clock on supply-chain exposure for NVIDIA and Google customers, and Micron's legal response will indicate the company's own confidence in its patent position. Second, Micron's HBM4 qualification milestone and the share of NVIDIA's Rubin supply window it secures relative to SK Hynix's estimated 65–70% — that gap is the most concrete available measure of Micron's standing in the next product cycle. Third, whether the concurrent capacity expansions across all three memory incumbents begin to produce conventional DRAM price pressure before AI infrastructure spending absorbs the incremental supply. Memory markets have historically compressed margins rapidly when supply growth outpaces a demand cycle's turn; the distinguishing feature of this cycle is that the demand driver — accelerator memory content per generation — is itself still expanding, which is precisely the variable that makes the outcome genuinely uncertain.