Gross margin expanded by 150 basis points sequentially to 67.7%, above the guided range, while operating margin came in at 60.3%, exceeding the high-end guidance of 58.5% . Earnings per ADR share were $4.31, topping the consensus estimate of $3.82 by $0.49
.
The composition of TSMC's revenue tells a stark story about where the semiconductor industry is heading. Revenue from high-performance computing (HPC) — dominated by AI accelerators for customers like Nvidia, AMD, and cloud service providers — increased 20% quarter-over-quarter and accounted for 66% of Q2 revenue . In dollar terms, HPC alone generated over $26.5 billion in the quarter.
By contrast, smartphone revenue decreased 4% sequentially and now represents just 22% of total sales . The rotation is structural: TSMC is no longer a mobile-first foundry but an AI-first foundry.
TSMC raised its 2026 full-year revenue growth forecast to slightly above 40% in U.S. dollar terms, a significant upgrade from the prior expectation of mid-30s% growth . For the third quarter of 2026, the company guided revenue of $44.6–$45.8 billion, which would represent approximately 37% year-over-year growth
.
Reflecting the urgency of expanding capacity, TSMC increased its 2026 capital expenditure plan to $60–64 billion, an increase of up to ~14% versus the previous plan of $52–56 billion . In Q2 alone, the company spent TWD 496 billion (approximately $15.5 billion) on capex
.
The capex raise is a direct response to capacity constraints. CEO C.C. Wei stated that the company is "fully accelerating and procuring equipment in advance, yet supplies remain tight" .
Perhaps the most significant strategic announcement was TSMC's pledge of an additional $100 billion investment in its Arizona operations, bringing the company's total planned U.S. investment to $265 billion . The expansion includes four additional advanced semiconductor fabs (using 2nm and below technologies), bringing the Arizona complex to 10 fab modules plus two advanced packaging facilities and an R&D center
.
The U.S. Commerce Department confirmed that the deal was tied to the U.S.–Taiwan trade agreement announced in January 2026 . Two of the four new fabs are already planning 2nm-class production
.
The July 2026 updates reveal several critical insights about the nature of AI demand:
Demand far exceeds supply. CEO C.C. Wei stated bluntly that AI chip demand will "outpace supply for years" and that "it will be a long time before we can meet customer demand" . Cloud service providers continue to send "very strong signals and positive outlooks" for AI purchases
.
New AI use cases are emerging. Management called out agentic AI workloads as a newly visible demand layer on top of existing training and inference needs . CEO Wei described the transition from generative AI to agentic AI — where models not only respond to inquiries but also perform actions — as accelerating
.
Pricing power is shifting. In June 2026, Wei said he would "like" to hike chip prices, reflecting extreme tightness in leading-edge supply and TSMC's leverage with customers like Nvidia .
The AI megatrend remains unshaken. TSMC explicitly affirmed its "conviction in the multi-year AI megatrend" and did not revise its AI-related revenue growth forecast downward — if anything, management hinted it could accelerate .
TSMC's dominance as the sole manufacturer of Nvidia's most advanced AI GPUs and Apple's A/M-series chips means its capacity constraints directly cap AI infrastructure buildout globally . For hyperscalers like AWS, Microsoft, and Google, their AI capital expenditure plans are partly dependent on TSMC's ability to deliver enough advanced packaging (CoWoS) and 2nm/3nm wafers.
The $100 billion Arizona expansion serves both geopolitical and commercial purposes. It deepens TSMC's de-risking of Taiwan-concentrated production while aligning with U.S. government priorities for domestic chipmaking under the CHIPS and Science Act framework . The expansion positions TSMC as one of the largest foreign investors in U.S. history
.
Customers may face continued allocation, longer lead times, and potentially higher prices as TSMC navigates the supply-demand gap. The company's ability to meet customer demand will determine the pace of AI infrastructure deployment for years to come.
In summary, TSMC's July 2026 updates depict a company operating at the center of an AI-driven capacity crunch — demand far outstripping supply, record financial results, dramatically increased spending, and a bet-the-company expansion in the United States to try to close the gap.