How is Micron Technology planning to more than double its HBM production capacity to about 100,000 wafers per month by the end of 2026—particularly by expanding 12 layer HBM4 production for Nvidia’s Vera Rubin AI accelerator—and how does this effort, including investments in Taiw

Micron is pursuing a two-track strategy: rapidly convert existing DRAM output and packaging capacity to HBM4 now, while funding new Asian manufacturing and packaging capacity that supports growth from 2027 onward. This should make Micron a materially stronger third supplier in AI memory, but it does not yet displace SK hynix’s leadership—and Samsung is also closing the gap quickly.

• Near-term HBM expansion: Industry reporting says Micron plans to add up to 60,000 wafer starts per month of HBM capacity by year-end, taking total HBM wafer-input capacity to roughly 100,000 wafers per month—more than double its 2025 level. The immediate mechanism is aggressive equipment investment and a larger mix of 12-layer HBM4, rather than waiting for new greenfield fabs. 7

• Nvidia Vera Rubin product ramp: Micron began volume shipment of its 36GB, 12-high HBM4 in calendar Q1 2026. The product is designed for Nvidia’s Vera Rubin platform, with stated bandwidth above 2.8 TB/s and over 20% better power efficiency than Micron’s HBM3E. 9 That qualification is strategically important: HBM suppliers need both sufficient yield and platform validation, not merely installed wafer capacity.

• Revenue and customer commitments: HBM4’s early volume ramp moves Micron from a smaller HBM participant toward a supply-constrained, contracted supplier. Public reports support that Micron’s 2026 HBM output is sold out and that portions of 2027 capacity are already committed; evidence for a fully booked 2027 and firm bookings into 2028 is less transparent publicly, so that stronger formulation should be treated as company/industry-channel guidance rather than independently verified fact. 6

• Longer-term physical expansion:

  • Taiwan: Micron completed its acquisition of PSMC’s Tongluo P5 site, adding a future DRAM manufacturing base; wafer starts are expected from 2027. 811
  • Singapore: Micron is building a new advanced wafer fab with a planned $24 billion, 10-year investment and has separately announced an approximately $7 billion HBM advanced-packaging facility. These projects deepen both DRAM supply and the packaging needed to turn stacked dies into HBM products. 54
  • Japan: Micron has broken ground on an approximately ¥1.5 trillion ($9.3 billion) Hiroshima expansion intended to produce advanced memory, including HBM-relevant output. 15

• Why the shortage persists: HBM is capacity-intensive because it consumes leading-edge DRAM wafers and advanced stacking/packaging capacity. Supply expansion also has a long lead time—estimated at roughly 12–18 months—while AI demand is still outpacing supply through at least 2026. 4

• Competitive position:

  • SK hynix remains the leader, with roughly 50% HBM revenue share in Q2 2026, after an estimated 58% in Q1; it retains the deepest Nvidia-centric HBM production experience and scale. 105
  • Samsung has rebounded sharply: its Q2 HBM revenue share reportedly reached 33%, up from 21% in Q1, and it is targeting about 38% by year-end as HBM4 ramps. 103
  • Micron was estimated at about 21% of HBM revenue in Q1 2026, alongside Samsung, versus SK hynix’s 58%. 14 Its planned ~100,000-wafers-per-month HBM input capacity would narrow the scale gap, but market share will depend on HBM4 yields, Nvidia allocation, advanced-packaging throughput, and how much of the added capacity is actually realized on schedule.

In effect, Micron is changing from a capacity-constrained follower into a credible, strategically necessary second-source/third-source provider for AI accelerators. The sold-out supply and multiyear customer commitments should strengthen pricing power and revenue visibility, while the Taiwan, Singapore, and Japan investments aim to prevent that position from being temporary. Yet SK hynix remains the incumbent HBM leader, and Samsung’s accelerating HBM4 ramp means Micron faces a two-sided competitive race rather than a simple catch-up opportunity.