SMIC ≠ Memory. SMIC is a logic foundry. It manufactures processors, controllers, PMICs, and other logic chips on mature nodes (14nm and above). It does not manufacture DRAM, NAND, or HBM.
That market belongs to Samsung, SK Hynix, and Micron. The memory pricing windfall that gave Micron an 84.9% gross margin flows entirely to memory manufacturers. SMIC's gross margin? 20.1% in Q1 2026. Different business, different economics.
There IS a second-order benefit narrative that sounds plausible:
"TSMC and Samsung are using all their capacity for AI chips and HBM → mature-node logic supply tightens → orders redirect to SMIC."
This has some truth. SMIC guided Q2 revenue up 14-16% sequentially, citing AI-driven demand and overseas clients shifting orders to Chinese foundries. The memory shortage creates a halo effect across the entire semiconductor supply chain.
But here's the catch.
| Micron (MU) | SMIC (0981.HK) | |
|---|---|---|
| Product | DRAM, NAND, HBM | Logic chips (mature nodes) |
| Pricing power | Extreme (shortage until 2028) | Limited (competitive foundry market) |
| Gross margin | 84.9% | 20.1% |
| Moat | Oligopoly (3 players) | Sanctions-limited, TSMC competition |
| Revenue scale | $41.5B/quarter | $2.5B/quarter |
| Growth YoY | +346% | ~15-20% |
If you want China exposure to the memory shortage:
SMIC benefits from the general semiconductor shortage but gets none of the memory-specific pricing power that drove Micron's 84.9% margin. The trade thesis for MU (memory oligopoly extracting rent from customers) simply doesn't map to SMIC (mature-node foundry in a competitive, capex-heavy market with sanctions overhang).
If you're trading the memory theme, MU and Samsung are the direct plays. SMIC is a different thesis entirely — China semiconductor self-sufficiency, mature-node tightness, and domestic substitution. Related, but not the same trade.