TSMC says it has no plans to move into 3D DRAM manufacturing, pushing back on a report that the foundry began laying groundwork for such an effort at the request of Nvidia. A TSMC spokesperson told Tom's Hardware that the company remains focused on its core business, dedicated IC foundry, and described its mission as serving as the trusted technology and capacity provider for the global logic IC industry. The denial follows reporting by The Information, which cited two people familiar with the initiative as saying TSMC had started preparing 3D DRAM manufacturing plans after a request from Nvidia, its largest customer.
According to those accounts, the effort is overseen by Hsu Kuo-chin, a TSMC vice president responsible for advanced packaging and a former head of Micron's Taiwanese operations. The reported initiative comes as widespread memory shortages have turned memory from a commodity into a strategic asset that can determine whether a company is able to ship its products to customers. TSMC already works closely with DRAM producers to package memory into system-in-packages for its customers, but it lacks the expertise and production capacity to make DRAM, let alone 3D DRAM.
Unlike conventional DRAM, which arranges memory cells in a single plane, 3D DRAM stacks multiple levels of memory cells vertically to increase capacity per wafer. Micron, Samsung, and SK hynix are all working on 3D DRAM projects, though Micron's DRAM strategy includes at least four more planar DRAM process technologies before 3D DRAM becomes a reality. Even for major memory makers, 3D DRAM is years away, meaning a TSMC strategy around it could not ease memory shortages for at least five years.
DRAM production requires manufacturing processes entirely different from the logic technologies TSMC is known for, and building multiple memory-cell layers while controlling defects and maintaining acceptable yields demands completely different fabrication techniques. A competitive planar DRAM process is hardly possible to skip, and TSMC does not have one. One route into the business would be acquiring a small specialty memory producer with its own fabrication capability, though such makers do not have leading-edge commodity DRAM nodes, so TSMC would still need billions of dollars and years of R&D to develop a leading-edge DRAM process.
Entering DRAM manufacturing would also carry substantial market risks for TSMC. Memory shortages frequently encourage capacity expansions that eventually create oversupply, depressing prices and profitability, and the move could complicate TSMC's relationships with large DRAM makers, particularly SK hynix, which partnered with the foundry in 2024 to manufacture logic base dies for next-generation HBM products. A potential entry could additionally raise antitrust questions, particularly if TSMC combined its dominant position in leading-edge logic manufacturing, its leading position in advanced chip packaging, and ramping memory production, with regulators examining whether such vertical integration would let TSMC favor its own memory products or disadvantage competitors that lack it.
TSMC does not necessarily need to manufacture DRAM to benefit from growing demand for advanced memory. It could capture a larger share of the value generated by DRAM by producing sophisticated logic base dies for HBM, integrating memory with processors using its SoIC hybrid bonding technology, and developing customized memory solutions with partners such as GUC and established DRAM manufacturers. That approach would let TSMC expand its presence in the advanced memory market without investing billions of dollars in developing competitive DRAM process technologies from scratch or building additional fabs for tens of billions of dollars.
More hardware news from TechManNews.
