Two trillion-dollar milestones. One day apart. The memory chip sector just told AI infrastructure investors exactly where it thinks the next decade of AI value lives.
On May 26, Micron Technology surged more than 19% in a single session after UBS raised its price target from $535 to $1,525 per share, a tripling, not an upgrade. The move pushed Micron past $1 trillion in market capitalization. The next morning, SK Hynix climbed 12%, crossing the same threshold. Samsung had already been there since the start of the month. All three dominant memory chipmakers now trade in the trillion-dollar bracket, a category that, two years ago, belonged almost exclusively to hyperscalers and consumer platforms.
The UBS move is the real signal here. A price target jump from $535 to $1,525 isn’t a normal analyst revision, it’s a structural reassessment. UBS cited long-term agreement opportunities with major customers and partially fixed pricing as the rationale. That framing matters: it’s not a bet on a single product cycle. It’s a bet that memory chip demand from AI data centers is durable enough to command long-term contracts at premium pricing.
Memory Chip AI Hardware Milestones, May 2026
| Entity | Event | Date | Value | Source |
|---|---|---|---|---|
| Micron Technology | $1T Market Cap Crossed | 2026-05-26 | $1T+ | Silicon Republic |
| SK Hynix | $1T Market Cap Crossed | 2026-05-27 | $1T+ | Silicon Republic |
| Micron Technology | UBS Price Target (new) | 2026-05-26 | $1,525 | Silicon Republic |
| Micron Technology | UBS Price Target (prior) | 2026-05-26 | $535 | Silicon Republic |
| SK Hynix | HBM Market Share (Q4 2025) | 2025-Q4 | ~57% (Counterpoint Research, qualified) | Counterpoint Research |
SK Hynix’s milestone carries a different kind of weight. According to Counterpoint Research, SK Hynix holds approximately 57% of the global High Bandwidth Memory market. HBM isn’t interchangeable commodity DRAM, it’s the specialized memory architecture that Nvidia’s GPUs and AMD’s Instinct accelerators require to run large-scale AI inference at speed. When AI infrastructure spend grows, HBM demand grows with it, and SK Hynix captures most of that demand. The 12% single-session gain reflects the market pricing in that concentration advantage.
The third data point anchoring all of this: Nvidia’s reported Q1 revenue of $81.6 billion, per Nvidia’s Q1 earnings report. That figure confirms that demand for AI compute hardware isn’t softening. Memory is downstream of GPU orders, when Nvidia ships at that volume, HBM suppliers fill the orders. Micron and SK Hynix aren’t running on expectations. They’re running on actual purchase commitments from the data center buildout already underway.
This is the third consecutive month in which AI hardware infrastructure plays, not frontier lab valuations, not software SaaS rounds, have driven the market’s most visible valuation events. The pattern connecting Nvidia’s Q1 blowout, the HBM pricing premium documented in prior coverage, and now the simultaneous Micron/SK Hynix milestones is the same thesis repeating: AI value in 2026 is accruing at the infrastructure layer, not the application layer.
What to Watch
What to watch
Micron’s next earnings call for hard data on HBM contract pricing and whether fixed-price agreements are holding. SK Hynix’s HBM4 production ramp timeline, next-gen AI accelerators, including Nvidia’s Vera Rubin NVL72, will require it. And whether Samsung, which hit $1 trillion at the start of the month on similar dynamics, maintains that threshold as its HBM market share gap with SK Hynix widens.
The catch is that concentration cuts both ways. SK Hynix at 57% HBM share means AI infrastructure procurement is heavily dependent on a single supplier. Any disruption, geopolitical, operational, or competitive, lands directly on data center construction timelines. Investors pricing in the upside should price in that supply-chain risk simultaneously. Watch Q3 procurement announcements from the major hyperscalers for early signs of diversification pressure.