TSMC weighs a move into advanced memory chips, The Information reports

TSMC, the company that builds the brains of the AI boom, is reportedly thinking about building its memory too.

According to a report from The Information, Taiwan Semiconductor Manufacturing Company is considering a push into advanced memory chips, with internal plans focused on 3D DRAM. The report ties the effort to demand from Nvidia, TSMC’s largest client.

Nothing has been announced. Investors in memory stocks reacted anyway.

What TSMC is reportedly weighing

The Information’s report, published November 6, 2026, says TSMC is drawing up internal plans to deepen its role in the AI hardware supply chain. The centerpiece is 3D DRAM production.

The research behind the report frames the move as being pursued at Nvidia’s request. That matters, because when your biggest customer asks for something, the request tends to carry a little extra weight.

The key word throughout is “considering.” TSMC is described as formulating plans, not breaking ground on memory fabs. No timeline, investment figure, or production target appears in the reporting.

Why memory became the bottleneck

The research describes a growing supply crunch in memory components, including DRAM and high-bandwidth memory, or HBM. HBM is a specialized stacked memory that sits right next to AI chips and feeds them data fast enough to keep them busy.

Shortages of HBM tied to AI demand have pushed prices higher, per the findings. Memory now sits alongside logic chips and advanced packaging as one of the chokepoints slowing the broader semiconductor industry.

Apple flagged rising memory costs earlier in 2026, a sign that the squeeze reaches consumer hardware makers too.

A departure from TSMC’s usual lane

TSMC built its reputation as a logic foundry. It manufactures processors designed by other companies, rather than selling its own branded chips.

Memory has historically been a separate business run by specialists. Moving into 3D DRAM would mark a notable shift away from that traditional focus, according to the research.

The findings suggest the change could point toward a more integrated manufacturing model. In that version, TSMC could offer more of the AI hardware stack from inside its own facilities.

The research also raises a second path: partnerships with existing memory specialists. That route would let TSMC expand its footprint without building an entire memory operation from scratch.

Memory stocks flinch

Markets did not wait for details. Shares of Micron (MU) and SK Hynix declined after the report surfaced, according to the research findings.

Memory makers have benefited from tight supply and rising prices tied to AI demand. A new high-volume supplier could ease those shortages, and investors moved to price in that possibility early.

What this means for the industry

For memory investors, the immediate takeaway is volatility. Positions in memory companies could see more swings as the market digests the prospect of added competition and potentially lower prices driven by higher supply, the research notes.

For AI chip buyers, including Nvidia, the stakes look different. More memory supply could relieve a constraint that has been capping how fast AI hardware ships. The research suggests that if TSMC succeeds, its 3D DRAM output could help stabilize pricing.

The research describes semiconductor companies increasingly chasing vertically integrated setups to cope with supply chain disruptions. There is also a strategic dimension worth watching: the research indicates a TSMC memory push could reshape alliances across the memory space and encourage investment in new technologies built for AI workloads.

Several signals will matter in the months ahead. Watch for any official confirmation from TSMC, any disclosed partnership with an established memory maker, and any comment from Nvidia on its memory sourcing.

For now, TSMC’s memory ambitions exist as internal plans described in a single report. That was still enough to move Micron and SK Hynix, which says a lot about how jittery the AI supply chain has become.