Taiwan Semiconductor Manufacturing reported fourth-quarter revenue of T$1,046.08 billion, representing a 20% year-on-year increase to approximately $33 billion. This performance surpassed consensus forecasts and underscored robust global demand for AI-optimized chips. Sales from its largest customer, Nvidia, accounted for nearly 25% of total revenue, while combined shipments to major hyperscale cloud providers rose 30%. The factory utilization rate remained above 100%, reflecting sold-out 3 nanometer capacity and the rapid ramp-up of its new 2 nanometer node.

According to third-party research, the company held a 72% share of the global foundry market in Q3 2025. Volume production of 2 nanometer chips commenced at its Kaohsiung Fab 22 facility in Q4, and management expects to double 2 nm capacity by year-end. Advanced packaging services, including CoWoS, are projected to deliver 1.2 million wafers in 2026—up from 800,000 in 2025—driven by demand for high-bandwidth memory integration in AI accelerators.

Goldman Sachs recently raised its earnings growth forecast by 35%, citing multiyear momentum in AI infrastructure spending. Consensus estimates now call for a 25% increase in net income for 2026, compared with 48% growth in 2025. With 2 nm premium pricing positioned 10–20% above the prior node, and margins expanding despite heavy capacity investments, the company could exceed Wall Street targets and sustain mid-20s percentage earnings gains.

While enthusiasm around generative AI has driven a 40% rally in the stock over the past twelve months, some strategists warn that capital expenditures on data centers may run ahead of software adoption. Deloitte projects AI data-center chip spending of $250–300 billion in 2026, up from $150 billion last year, but a moderation could expose valuation risks. Nonetheless, continued tight capacity at leading foundries and planned capacity expansions support a positive long-term outlook for a company poised to remain pivotal in AI supply chains.