Solidigm eyes US NAND fab as funding questions linger

This isn't a fab announcement—it's a structural fracture in NAND's geographic production logic. Technically, the US line will almost certainly launch at 128-176 layers, deliberately tiered below SK Hynix's 238+ layer Korean fabs. The US plant targets AI-inference enterprise SSDs and HBF prototypes; Korea keeps grinding high-layer consumer products. Upstream, Lam and AMAT gain shorter domestic lead times, but ASML's DUV export controls remain a quiet chokepoint. On compliance, NAND has been chronically under-allocated in remaining CHIPS Act funds. A $3-5B subsidy could compress the US cost premium (water, energy, labor running 15-22% above Korea) to near breakeven. The deeper risk: export-control lists may bifurcate US-made from non-US-made NAND, forcing downstream OEMs through full re-certification. Strategically, Micron is the only pure-play US NAND maker. Solidigm's entry pushes US domestic capacity share from roughly 25% to 40%+. Samsung's Texas packaging site now faces a binary: build a wafer fab or become a back-end contractor. Kioxia/WDC will likely accelerate Japan-US capacity coupling as a hedge. 12-24 month tail: AI inference clusters pulling QLC/PLC enterprise SSD demand will lift NAND ASPs 10-15% off the trough. The real variable is whether HBF displaces a slice of HBM capacity by 2026. If it does, NAND's strategic weight approaches DRAM's for the first time, and the memory sector's valuation anchor gets rewritten.