More than 40,000 Samsung Electronics workers rallied at the Pyeongtaek semiconductor complex on April 23, demanding a 7% wage increase and the removal of a cap on performance bonuses -- the union wants Samsung to allocate 15% of annual operating profit as bonus pool, which at current projections would be roughly 45 trillion won. Night shifts on April 23 and 24 saw partial no-shows, cutting chipset output by 58% and memory output by 18% during the affected period. The union set a deadline: if Samsung does not negotiate a revised package by May 21, an 18-day general strike begins.
The Pyeongtaek complex is not just a large factory. It is the largest memory fab on earth, producing a significant share of global DRAM and NAND. Samsung supplies most of the world's HBM3E and LPDDR5X that feeds AI training clusters and edge inference platforms. An 18-day work stoppage at that scale would not just affect Samsung's customers -- it would drain industry inventory buffers that are already under pressure from the surge in AI server demand. The 2024 NAND shortage, which was caused by a production curtailment, not a work stoppage, sent SSD prices up 40% in three months. A deliberate 18-day strike would be a harder stop with less warning.
For hardware teams, the relevant question is inventory position and order lead time. DRAM and NAND spot market pricing typically leads physical supply tightening by 6 to 8 weeks. If the May 21 deadline is missed and a strike begins, the pricing signal will arrive before the shortage does. Hardware teams with active designs or production ramps that depend on Samsung memory -- specifically LPDDR6X ramps, which Samsung is the primary supplier for -- should be watching May 21 as a decision gate. The company still has room to settle, and strikes of this kind often resolve under pressure, but the threat is credible enough that ignoring it is not the right call.