Taiwan Semiconductor Manufacturing Company controls more than 95% of the global advanced AI chip foundry market, supplying major players in datacenter computing. In 2025 the company shifted its primary focus from smartphone logic nodes to high-performance computing wafers, leveraging its 3-nanometer (N3) and 2-nanometer (N2) process technologies. Its market share on the N3 node has risen to roughly 75%, while early N2 capacity is allocated to strategic hyperscaler contracts. This dominance ensures that TSMC remains the indispensable fabricator for next-generation GPUs, custom accelerators and AI training silicon.

TSMC reported full-year revenue of $122.4 billion for 2025, up 36% from 2024 and marking the first time annual sales exceeded $100 billion. Gross margin climbed from 56.1% in 2024 to 59.9% last year, while operating margin expanded from 45.7% to 50.8%. In the fourth quarter alone, gross margin reached 62.3% and operating margin 54.0%. These improvements reflect both strong volume growth—driven by new fab ramps in Taiwan and Arizona—and premium pricing power supported by constrained capacity on advanced nodes.

Despite a 69% share price increase since January 2025, TSMC trades at approximately 25 times projected 2026 earnings, below the forward multiples of peer semiconductor manufacturers. Analysts forecast an average annual revenue compound growth rate of 20–25% through 2029, underpinned by the AI super-cycle and continued migration of smartphone system-on-chips to N3 and N2. With net income margins near 50% and free cash flow in excess of $200 billion in Taiwanese dollars annually, the current valuation presents a compelling entry point for long-term investors.

TSMC has committed to a capital expenditure budget of $52–56 billion for 2026, up from $47 billion in 2025. Investments will prioritize capacity expansion for N3 and N2 nodes, advanced packaging facilities for CoWoS and CoWoP technologies, and completion of its second Arizona fab. Management expects wafer throughput on N3 to rise from 150,000 to 180,000 wafers per month, while N2 capacity could quadruple by year-end. These investments aim to alleviate the industry-wide bottleneck in AI compute supply and reinforce TSMC’s leadership for the next decade.