Micron's Taiwan Unions Are Asking for a Bonus Worth 83 Months of Pay. A Strike Vote Could Come This Month.

Micron labor unions in Taiwan are moving toward a strike unless the company overhauls bonuses—requesting a one-off fiscal 2026 bonus equivalent to ~83 months of salary and, from fiscal 2027, bonuses at 15% of operating profit paid quarterly. The company’s fiscal Q3 2026 operating income was $33.3B on $41.5B revenue (80% operating margin), implying the requested 15% pool could be roughly ~$5B in a quarter and potentially >$6B with fiscal Q4 guidance (about $50B revenue at ~86% gross margin). A strike risk is concentrated in Taiwan, which accounts for a majority of DRAM output, raising concerns about supply to the AI build-out even though mediation and a strike vote timing remain uncertain.

Analysis

The market should view this less as a labor-cost story and more as a test of how concentrated the AI memory supply chain really is. A settlement is the base case because the economics can absorb a one-time concession, but the path to settlement can still create a short, sharp supply shock if management miscalculates and workers force a vote. In the next 2-6 weeks, MU is exposed to headline-driven multiple compression even if the ultimate earnings impact is immaterial; the real downside is operational, not arithmetic.

Second-order, any Taiwan disruption would likely tighten DRAM and HBM lead times before it meaningfully changes reported revenue, which is why the first beneficiaries are probably competitors with cleaner non-Taiwan capacity, especially SSNLF and SKHYV. NVDA is the downstream risk: not because this would stop the AI build-out, but because it can temporarily raise memory input costs and lengthen qualification cycles. That is a months-not-days issue unless a strike actually lands, in which case spot memory pricing could gap higher within weeks.

The contrarian point is that a large bonus demand is not necessarily a red flag on margins; it is proof the cycle is strong enough to support a richer labor contract. What the consensus may miss is that an expensive settlement could be neutral-to-positive for MU if it removes a binary supply overhang, while a prolonged dispute would likely be the true earnings risk. Falsifier: no strike vote, no production interruption, and a clean record payout commentary on the next earnings call should unwind the premium quickly.

Key Decisions for Investors

• Do not fade MU outright on the headline; treat any 3-5% dip without a formal strike vote as a buy-the-dip setup for 1-3 months, since the base case is a negotiated settlement and the cost hit should be absorbed at current margins.
• If mediation fails and a strike vote is scheduled, buy short-dated MU call spreads only after an initial selloff; the risk/reward improves if the market overprices a lasting supply loss that is more likely to be temporary.
• For relative value, initiate a tactical long SSNLF / short MU pair on any escalation signal over the next 2-6 weeks to express the view that competitor capacity benefits more than MU from tighter memory pricing.
• Use NVDA as an indirect alert, not an outright trade: if the dispute extends beyond a few weeks, reassess AI hardware margins and lead times; if no vote materializes by month-end, remove the supply-chain risk premium.