[News] TSMC Reportedly Eyes Up to 15% 3nm Price Hike in 2H26, Further 5%–10% Seen in 2027 Amid AI, ASIC Demand
AI continues to drive tightness in advanced process capacity at TSMC. According to Commercial Times, citing sources, TSMC is expected to raise 3nm pricing again in the second half of 2026, with increases of up to 15%, followed by a potential additional 5%–10% hike in 2027.
As the report notes, 3nm demand was previously driven mainly by smartphone SoCs. However, as AI server refresh cycles begin, companies including NVIDIA, AMD, Google, AWS, and multiple cloud service providers are accelerating 3nm adoption, sharply increasing wafer demand. The report adds that major cloud firms have increasingly invested in in-house ASICs to reduce reliance on general-purpose GPUs, further expanding demand.
Utilization at TSMC’s main 3nm production site, Fab 18, remains elevated, with customer queues showing little sign of easing. According to the report, monthly 3nm capacity rose from around 130,000 wafers in early 2026 to roughly 160,000–175,000 wafers in the second quarter. However, the report notes that even with continued expansion, AI demand growth still far exceeds market expectations.
Liberty Times further reported that production of NVIDIA GPUs, Broadcom ASICs, and Marvell custom chips remains heavily dependent on TSMC. According to TrendForce, TSMC held a 70.4% share of the global foundry market in 4Q25, maintaining a dominant lead in the industry.
Meanwhile, institutional investors cited by Commercial Times said 3nm has become the most stable mass-production node for AI chips, offering greater production readiness and cost advantages than 2nm, which remains in the early stages of yield ramp-up. The report also notes that amid rising costs from overseas fab expansion and growing depreciation pressure tied to advanced processes, higher 3nm pricing could also help support TSMC’s gross margins.
Against the backdrop of tightening 3nm capacity and potential price hikes, the report notes that TSMC’s annual shareholder meeting on June 4 is expected to draw close market attention, with Chairman C. C. Wei likely to provide updates on AI demand, advanced processes, and overseas expansion.
TSMC’s Role in AI Deepens Beyond Advanced Processes
As competition for TSMC capacity intensifies, a May 26 dinner between NVIDIA CEO Jensen Huang and TSMC executives also drew market attention. According to UDN News, when asked whether TSMC could meet NVIDIA’s capacity needs, Chairman C. C. Wei responded, “We’re already working very hard,” while making a zipper-across-the-mouth gesture, as Huang described TSMC as “my good friends.”
As Commercial Times notes, institutional investors said TSMC is not only NVIDIA’s key partner for advanced processes, but also plays a critical role in CoWoS packaging, SoIC stacking, and integration technologies such as CPO and silicon photonics. The report adds that as AI server architectures evolve toward rack-scale computing, CPO is emerging as a key technology for next-generation AI systems. With NVIDIA actively advancing silicon photonics platforms like Spectrum-X and Quantum-X, future optical interconnect adoption could increasingly rely on TSMC’s COUPE silicon photonics platform and advanced packaging capabilities.
Read more
- [News] TSMC Reportedly Upgrades Central Taiwan 28/22nm Fab to 4nm; Phase 2 1.4nm Trial Production May Start 3Q27
- [News] AMD Zen 7 Reportedly Built on TSMC A14 Node as Powertech’s FOPLP Packaging Said to Be Under Evaluation
(Photo credit: TSMC)
