Why TSMC Still Pared Gains Despite Dual Tailwinds From Nvidia AI Partnership and Price Hike Expectations?

TradingKey - During the Asian trading session on June 1, TSMC (2330.TW) rose as much as 2.7% intraday to NT$2,415, hitting a record high and pushing its market capitalization above NT$62 trillion. However, it eventually closed at NT$2,355, with gains narrowing to flat, giving back all earlier increases.

[Source: TradingView]

The early session surge was driven by two main factors: first, TSMC and Nvidia ( NVDA) announced an expansion of their strategic cooperation to fully integrate Nvidia's AI technology into the semiconductor manufacturing process; second, market rumors suggested that TSMC plans to raise quotes for its 3nm process by as much as 15% in the second half of 2026.

Reportedly, under the cooperation framework, TSMC will adopt Nvidia's CUDA-X acceleration libraries and AI models in areas such as computational lithography, electronic structure simulation, advanced process control, and fab operations optimization. TSMC stated that this move is expected to reduce lithography costs or cycle times by 20% to 50% and increase the speed of chemical simulations by an average of 50 times.

Regarding pricing, sources revealed that due to surging demand for AI chips, flagship smartphone chips, and high-performance computing (HPC), TSMC plans to raise 3nm process quotes by 15% in the second half of the year, with a potential further increase of 5% to 10% in 2027. The 1.6nm process, scheduled for launch in the second half of 2026, is expected to have wafer quotes as high as $45,000, approximately 50% higher than the current most advanced process.

At today's GTC Taipei conference, Nvidia unveiled the "RTX Spark" PC chip using TSMC's 3nm process, with multi-brand products expected to hit the market starting in the fall of 2026. Since the PC market volume is much smaller than TSMC's AI business, the product's impact on TSMC's short-term performance is limited, representing more of a medium-to-long-term positive.

Stimulated by multiple positive catalysts, TSMC rose over 2.5% in early trading but subsequently retreated, eventually closing flat. Market analysis suggests this was primarily due to profit-taking following excessive prior gains, while the Nvidia partnership and price hike rumors had already been gradually priced in ahead of GTC. So far this year, TSMC's Taiwan-listed shares have risen by more than 50%, leaving many investors with substantial floating profits. As the stock hit a record high, a concentration of take-profit selling emerged, causing the gains to narrow quickly.

Furthermore, the price hike rumors have not yet been officially confirmed, and some cautious investors opted to reduce their positions at high levels, waiting for clearer signals from the annual general meeting on June 4.

In terms of financial performance, TSMC's gross margin reached 66.2% and its net margin was 50.5% in the first quarter of 2026. The company expects second-quarter revenue to increase to $39.0 billion - $40.2 billion, with full-year USD revenue growth exceeding 30%.

Notably, the premium of TSMC's American Depositary Receipts (ADRs) ( TSM) over its Taipei-listed shares has dropped to a two-year low. The average premium in May was 13.7%, significantly lower than the 26% seen last December. So far this year, TSMC's Taiwan-listed shares have risen over 50%, while its ADRs have risen less than 40%, indicating that local investors' optimism regarding the AI cycle exceeds that of the U.S. market.

TSMC will hold its annual general meeting on June 4. The market is focused on issues such as the operational outlook for the second half of the year, the direction of capital expenditure, and the expansion of board seats.