Nvidia's Margin Pressure Is Micron's Revenue Boost

Nvidia guided Q4 gross margin to a 71%–72% trough (down from ~75% in Q2), before recovering to 72%–73% in fiscal 2028. Management attributed the decline to high-bandwidth memory costs rising to an estimated 30%–40% of accelerator build cost, with the supply constrained by Samsung, SK Hynix, and Micron, while already-negotiated price increases in fiscal 2028 are expected to land with a timing lag. Nvidia’s CFO expects next-year revenue growth around ~70% as AI accelerator demand outpaces near-term manufacturing capacity, and memory makers were reported up 3%–5% after the update.

Analysis

Near term, the market is likely to misread NVDA's margin guide as an earnings-quality problem when it is mostly a transfer of economics to a three-player HBM oligopoly. That makes MU the cleaner beneficiary over the next 1-3 quarters, with SKHY and SSNLF also capturing incremental bargaining power; the key is that capacity cannot respond fast enough to break pricing before the FY2028 reset.

The second-order risk is that rising accelerator bill-of-materials pushes hyperscalers toward custom silicon and tighter capex discipline. That is a 6-18 month issue: if each next-gen GPU requires materially more HBM dollars, the AI build-out becomes less unit-elastic, which could slow NVDA's mix expansion and pressure the broader AI hardware basket. The Nov. 17 print is the first catalyst to watch; if the trough lands below 71% or the recovery language softens, the "timing gap" thesis weakens.

Contrarianly, the consensus may be over-enthused about MU's leverage while underpricing how much of this is already in the stock after the post-earnings pop. Memory upside remains real, but it is still a cyclical business; if new capacity comes online faster than expected or HBM specification intensity flattens, the margin transfer can mean-revert. That argues for relative value rather than a naked beta bet.

Key Decisions for Investors

• Long MU / short NVDA in equal dollars for the next 1-3 months, targeting relative outperformance into the November NVDA update; thesis breaks if NVDA holds gross margin above 73% or if MU commentary stops implying HBM scarcity.
• Buy MU 6-9 month call spreads rather than outright stock if you want defined risk; the setup is asymmetric if HBM pricing stays tight into early FY2028, but upside is capped if memory supply ramps faster than expected.
• If already long NVDA, use any post-earnings strength to trim and re-enter only after the market digests whether the 71%-72% trough is real; otherwise the stock can de-rate on margin optics even with strong revenue growth.
• Watch SKHY/SSNLF as the economic winners, but prefer MU as the liquid proxy; if MU underperforms after NVDA's next print, it likely signals the market is discounting future capacity additions sooner than the article implies.