China has cut off Japan’s supply of high-purity tungsten powder, and the fallout is now reaching one of the semiconductor industry’s most sensitive choke points: tungsten hexafluoride, or WF6. Two Japanese suppliers, Kanto Denka and Central Glass, have told major customers including Samsung, SK Hynix, and TSMC that they will stop making WF6 from 1 July 2026 because they can no longer secure the raw material.
The immediate damage is not a finished-chip shortage, but the kind of upstream disruption that eventually becomes everyone’s problem. WF6 is a workhorse chemical in advanced chip fabrication, used to build interconnects in cutting-edge processors and heavily tied to NAND and HBM memory production. If that supply tightens, memory makers feel it first, then logic foundries start juggling inventories and purchase contracts.
Why WF6 matters for chip production
WF6 is made from high-purity tungsten powder, and the economics are brutal: 60% to 70% of its production cost comes from that single input. Japan’s dependence on China for the powder was already a vulnerability; once Beijing tightened export controls on tungsten earlier in 2026, the tap effectively dried up. The Japanese producers lasted about five months on stockpiles before admitting they could not replace the supply.
- Material hit: high-purity tungsten powder
- Industrial product at risk: tungsten hexafluoride (WF6)
- Customers named: Samsung, SK Hynix, and TSMC
- Production stop date announced: 1 July 2026
A supply chain built on a single choke point
China’s move looks less like a random trade hiccup and more like a reminder that critical materials are now part of industrial policy, not just commerce. That matters because Japan is not the only country exposed here: South Korea’s memory giants and Taiwan’s biggest foundry are downstream from the same chemistry. In other words, a powder shortage in one country can become a production headache for the companies that make the chips powering everything from phones to AI servers.
The bigger question is whether this spills into broader semiconductor pricing or stays contained as a niche materials problem. If substitute supply is unavailable for long enough, buyers will have to choose between paying more, redesigning processes, or delaying output. None of those options are especially elegant, which is exactly why raw materials have become such effective leverage.

