Will Micron's average gross margin for FQ3-FQ4 FY2026 remain above 45%?

This market is now near-certain YES. FQ3 gross margin is guided at ~81%. For the average of FQ3+FQ4 to fall below 45%, FQ4 would need to collapse to below 9% — a margin decline of over 72 percentage points in a single quarter. The largest single-quarter gross margin decline in Micron's history was approximately 20pp during the FY2023 downturn, and that took multiple quarters of sustained pricing pressure. For FQ4 margins to collapse to 9%, DRAM prices would need to fall roughly 80% in a single quarter — there is no historical precedent for this in the DRAM market, even during the worst downturns. Supply tightness is expected to persist beyond CY2026, NAND has reversed from underutilized to supply-constrained, and every segment reported margins well above 45% in FQ2.

The math makes this nearly binary. FQ3 guided at 81%, so the two-quarter average needs FQ4 > 9% to clear 45%. Even in the most extreme Black Swan scenario — a sudden hyperscaler CapEx freeze, DRAM spot price collapse, and simultaneous inventory correction — the repricing mechanism (75% quarterly contracts) cannot compress margins fast enough. FQ4 would still have the benefit of contracted prices, HBM volume commitments, and 25% of revenue on longer-term arrangements. The only realistic question is whether margins stay above 60% or above 70%, not whether they stay above 45%.

I assign a slightly higher residual probability to failure than my colleagues because I consider the possibility of non-operational margin impacts — such as a large inventory writedown (Micron took $1.4B in FY2023), a one-time charge related to CapEx escalation, or an accounting reclassification. However, even a $3B inventory writedown on ~$24B quarterly revenue would reduce gross margin by ~12pp from a guided 81%, landing at ~69% for FQ3. FQ4 would need a similarly devastating quarter, and both quarters would still clear 45%. The non-GAAP nature of the guided 81% also adjusts for stock-based compensation, which provides additional buffer. This is a >95% probability event.

The question has been effectively answered by the Q2 results and FQ3 guidance. FQ2 actual: 74.9%. FQ3 guided: ~81%. For the FQ3+FQ4 average to miss 45%, FQ4 would need to print below 9%. This is a mathematical near-impossibility given: (1) supply tightness persisting beyond CY2026, (2) NAND reversing to supply-constrained from underutilized, (3) all four segments above 45% in FQ2, (4) pricing driven by structural shortage not temporary demand spike. The prior prediction of 74% was appropriate when we only had FQ1 data (56.8%) and FQ2 guidance (68%). With FQ2 actual at 74.9% and FQ3 guided at 81%, the 45% threshold is no longer a meaningful risk.

While I agree the 45% threshold is almost certainly cleared, I maintain slightly more residual uncertainty for extreme tail scenarios. The memory industry has surprised before — the FY2023 downturn saw margins go from positive to deeply negative in two quarters. However, the current setup is different: (1) starting point is 81% guided, not 45%, (2) supply constraints are structurally driven by HBM trade ratios, not demand-dependent, (3) the 75% quarterly repricing means even a sudden demand shock takes 2+ quarters to fully transmit. The 6% residual accounts for a true black swan (major geopolitical disruption, hyperscaler bankruptcy, fabricated earnings) rather than any normal business scenario.

The updated context makes this one of the highest-confidence predictions in our ensemble. Consider: FQ2 gross margin of 74.9% with FQ3 guided at 81% means the first of the two measured quarters alone would need to be averaged with a catastrophic FQ4 to miss the threshold. The 45% bar was set when the question was framed — at that time, it represented a meaningful test of margin sustainability. With margins now running at nearly 2x the threshold, the question has been overtaken by events. Supply tightness, NAND recovery, and continued HBM mix shift all support margins remaining in the 70-85% range through FQ4.

FQ3 guided at 81%, FQ2 actual at 74.9%. For average to miss 45%, FQ4 would need <9% — impossible without a catastrophic event with no historical precedent. Supply constraints persist beyond CY2026. All segments above 45% in FQ2. This is effectively resolved YES.

The 45% threshold is now a low bar given 81% FQ3 guidance. Even accounting for CapEx-related cost escalation (Idaho fab startup, rising depreciation from $25B+ annual CapEx), the margin buffer of 36pp above threshold absorbs any plausible cost headwind. Management guided continued margin expansion.

The mathematics eliminate normal business risk scenarios. The residual 6% accounts for truly extraordinary events — accounting fraud discovery, sanctions-driven production halt, or natural disaster destroying fabrication capacity. Under any plausible demand or pricing scenario, margins remain above 45% for the H2 FY2026 average.