Taiwan Semiconductor Manufacturing Co. raised its full-year revenue and capital spending outlook after reporting a steep increase in second-quarter profit, reinforcing its position at the center of the global artificial intelligence hardware boom while also signaling that the cost of staying ahead in advanced chipmaking is rising sharply.
The world’s largest contract chipmaker said second-quarter revenue reached $40.2 billion, up 33.7% from a year earlier. Net profit rose 77.4% to NT$706.6 billion, driven by strong demand for advanced semiconductors used in artificial intelligence accelerators, graphics processors, high-performance computing systems and premium electronics.
The company now expects annual capital spending of between $60 billion and $64 billion. It said 70% to 80% of that budget will be directed toward advanced manufacturing processes, about 10% toward specialized technologies, and 10% to 20% toward advanced packaging, testing and related equipment. On a broader basis, management said up to 80% of spending will support advanced process technologies, while up to 20% will go to advanced packaging and testing capacity.
TSMC also confirmed that its total commitment to U.S. expansion now stands at about $265 billion, including an additional $100 billion pledged for new facilities in Arizona. The program covers advanced logic fabrication plants as well as packaging operations, making it one of the largest foreign-backed manufacturing commitments in the U.S. technology sector.
The scale of the spending plan underlines how quickly computing demand has shifted toward AI infrastructure. It also shows the financial strain attached to that shift. Management said revenue growth remains strong, but margins and free cash flow will be watched closely as the company expands outside Taiwan, ramps up next-generation chip production and builds costly new facilities in the United States.
Second-quarter results show AI demand driving growth
TSMC’s latest results reflected strong orders for high-end chips built on the company’s most advanced nodes. Chips made using 7-nanometer technology and more advanced processes accounted for 77% of wafer revenue during the quarter.
The company also reported its first wafer revenue contribution from 2-nanometer technology, which accounted for 3% of the total. While still small, the contribution marks an important milestone as TSMC prepares for a broader transition to the next generation of chip production.
The company expects third-quarter revenue to reach between $44.6 billion and $45.8 billion, suggesting continued momentum from AI-related orders. Gross margin is expected to remain between 56% and 58%, a level that reflects strong pricing power and disciplined capacity allocation, even as costs rise.
Chairman C.C. Wei said planned investments are supported by multi-year demand from major U.S. customers. He also said the company intends to preserve healthy margins without relying on rapid price increases, stressing the importance of long-term customer relationships and stable supply.
That message is important because TSMC’s customers include some of the world’s largest technology companies, many of which depend on the company’s most advanced production lines to build AI accelerators, data center processors, smartphone chips and custom silicon. With demand concentrated at the leading edge, available capacity has become increasingly valuable.
U.S. expansion becomes a central focus
TSMC’s Arizona expansion has become a major part of its long-term strategy, but it is also expected to weigh on profitability. The company said the early phase of overseas operations, especially in the U.S., may reduce profitability by 2 to 3 percentage points. That impact could later widen to 3 to 4 percentage points as production expands and the company works to improve yields.
Higher construction costs, labor expenses and supply chain adjustments in the U.S. are expected to make American production less profitable than operations in Taiwan, at least during the early years. The company must also maintain high utilization rates once the plants begin operating, because underused advanced fabs can become a major drag on returns.
Management has said the long-term case for U.S. expansion remains solid because key American customers want greater geographic diversity in chip supply. The buildout is also aligned with U.S. policy efforts to localize more semiconductor manufacturing after years of concern about supply chain concentration in East Asia.
Still, the financial challenge is significant. TSMC’s overseas sites must eventually deliver acceptable yields, controlled costs and sufficient production volume. If those targets are not met, the company could face weaker returns on capital despite strong customer demand.
2-nanometer ramp-up may pressure margins
TSMC expects the ramp-up of 2-nanometer production in the second half of 2026 to reduce gross margin by roughly 3 to 4 percentage points. That is a common pattern during transitions to new manufacturing nodes, when production is more complex, yields are still improving and customers are gradually shifting designs to the new process.
The company’s 2-nanometer technology is expected to be used in future generations of high-performance processors, AI accelerators and flagship mobile chips. Its early contribution to wafer revenue shows that commercial adoption has begun, but the larger financial effect will come as production expands.
Advanced nodes require expensive tools, cleanroom space, engineering talent and materials. They also depend on tight coordination with customers, especially for custom chips designed for AI and high-performance computing. As the industry moves to smaller geometries and more complex packaging, the cost of each new technology cycle continues to rise.
TSMC’s margin guidance suggests that management believes the company can absorb these pressures without losing control of profitability. But the next several quarters will show how well the company balances higher depreciation, overseas costs and the early costs of 2-nanometer production against strong pricing and demand.
Cash flow remains under pressure from heavy spending
Finance chief Wendell Huang said operating cash flow remains strong and liquidity is sufficient. However, the company’s elevated capital outlay is expected to test free cash flow over the medium term.
That issue matters because TSMC is funding several major priorities at once. It is expanding leading-edge capacity, building advanced packaging operations, preparing for 2-nanometer volume production and constructing large overseas facilities. Each of these projects requires years of spending before full returns can be measured.
The company has also maintained distributions to shareholders, but rising capital needs could lead to closer scrutiny of cash allocation. Management must balance payouts, research and development, overseas expansion and the need to preserve financial flexibility.
For traders, TSMC’s cash flow profile has become one of the clearest indicators of how costly the AI infrastructure race has become. Strong earnings show that demand is real, but rising capital spending shows that meeting that demand requires vast upfront funding.
AI orders reshape capacity priorities
TSMC’s guidance points to a chip market increasingly shaped by artificial intelligence. Management has forecast dollar revenue growth slightly above 40% for 2026, driven by faster demand for AI chips used in GPUs and high-performance computing products.
That growth reflects the rapid buildout of data centers designed to train and operate large AI systems. These facilities require advanced processors and specialized accelerators, many of which depend on TSMC’s leading-edge technology.
As a result, capacity at advanced fabs is being allocated to the highest-priority customers and the most profitable product categories. Large technology companies with multi-year purchasing commitments are in a stronger position to secure supply, while smaller chip designers may face longer waits or higher costs.
This shift has implications beyond the largest AI companies. Firms that design custom processors for networking, security, telecommunications, cloud computing and decentralized infrastructure often rely on the same advanced manufacturing ecosystem. When the biggest buyers reserve capacity years in advance, other sectors can face tighter supply and rising production costs.
The effect is not limited to wafer fabrication. Advanced packaging is also becoming a bottleneck because many AI processors require complex packaging methods to connect logic chips with high-bandwidth memory. TSMC’s allocation of 10% to 20% of capital spending toward packaging and related equipment shows that back-end capacity is now a strategic priority rather than a support function.
Impact may extend to digital infrastructure and mining hardware
The heavy focus on AI and high-performance computing leaves less flexibility for some older or lower-margin technology segments. That includes certain types of custom hardware used in decentralized networks, security systems and crypto mining equipment.
Mining hardware producers and other custom chip designers depend on access to advanced manufacturing lines to improve energy efficiency and computing performance. If AI chip demand absorbs a large share of capacity, the cost of producing new equipment may rise, and delivery schedules may lengthen.
This does not mean all digital asset networks face the same pressure. Networks that rely less on specialized physical hardware will be less exposed to chip supply constraints. But operations that require constant hardware upgrades, especially energy-intensive mining operations, are more sensitive to semiconductor pricing, equipment availability and financing costs.
Some listed mining companies have already adjusted business models in response to changing economics. Several have expanded into high-performance computing and data center hosting, using power contracts, sites and infrastructure to serve AI workloads rather than relying only on token validation. That shift has become more visible as AI customers compete for data center capacity and as mining margins fluctuate.
For traders in digital assets, the connection between semiconductor supply and network economics is indirect but important. Hardware costs affect mining profitability, which can influence corporate treasury decisions, equipment purchases and network hash rate growth. Large sales of digital assets by mining firms can add pressure to short-term market liquidity, although such sales vary by company and market conditions.
Competitive position remains strong
Despite the rising costs, TSMC remains in a dominant position. Its advanced manufacturing capacity, yield performance and customer base give it a strong lead in the foundry market. For the largest chip designers, there are few alternatives at the most advanced nodes.
The company’s ability to keep gross margin near the high-50% range while spending heavily suggests that demand remains strong enough to support premium pricing. It also suggests that customers are willing to commit early to secure supply, especially for AI-related chips.
However, dominance does not eliminate risk. TSMC must execute major projects across several geographies, maintain technology leadership and avoid significant delays in new node deployment. It must also manage political, labor and cost differences across Taiwan, the United States and other production locations.
The company’s U.S. expansion gives it closer access to major American customers and supports supply chain resilience. But the economics of overseas manufacturing remain less favorable than in Taiwan. How quickly TSMC can narrow that gap will influence long-term profitability.
Margins and returns become the key test
The central question is no longer whether AI demand is strong. TSMC’s results and guidance make that clear. The bigger question is whether record capital spending can generate efficient returns.
Gross margin compression from 2-nanometer production, early losses from overseas facilities and the burden of high depreciation will all test the company’s financial model. Free cash flow will be especially important because it shows how much cash remains after the company funds its aggressive expansion plans.
If TSMC can maintain strong margins while scaling new technologies and overseas sites, it will strengthen its role as the key manufacturing partner for the AI era. If costs rise faster than expected or yields improve more slowly, profit growth could become harder to sustain even with strong revenue.
For the broader technology sector, TSMC’s performance will act as a benchmark. The company sits at the point where AI demand, government industrial policy, supply chain security and semiconductor economics meet. Its results offer one of the clearest views into whether the global chip industry can expand fast enough to support AI growth without damaging profitability.
The latest quarter shows a company benefiting from extraordinary demand, but also committing to extraordinary spending. That balance will define TSMC’s next phase: technological leadership backed by massive capital commitments, with long-term returns still dependent on execution, yields and disciplined capacity management.
Want to navigate AI-driven volatility like TSMC’s investors? Master technical analysis to time entries, exits, and manage risk.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
