TL;DR

• TSMC's second-quarter performance and full-year guidance reinforce AI chip demand, with the full-year capital budget raised to $60 billion to $64 billion.
• Market divergence focuses on U.S. manufacturing costs, the ramp-up of 2nm technology, and long-term capital returns; strong demand does not equate to pressure-free margins.
• Related stocks: TSM, NVDA, AMD, AVGO, ASML.

After TSMC announced its second-quarter financial results on July 16, it raised its full-year revenue and capital budget guidance and continued to advance its U.S. expansion plans.

This should have been standard good news for the AI supply chain. High revenue growth, an increased share of advanced processes, the 2nm technology beginning to contribute to revenue, and demand from U.S. customers described by management as structural demand over several years. However, the market's focus shifted from "Are there orders?" to "Can the orders turn into sufficiently good returns?"

The reason is straightforward. The stronger the AI demand, the more capital TSMC needs to invest. Building factories, purchasing equipment, and expanding advanced packaging all require upfront expenditures. U.S. manufacturing corresponds to higher costs, longer ramp-up times, and more complex execution environments.

TSMC Chairman Wei Che-Chia provided a somewhat optimistic answer during the conference call. He emphasized that the company's investments come from long-term demand from leading U.S. customers and that they will maintain a reasonable profit margin necessary for long-term expansion, without suddenly raising prices significantly due to supply constraints.

Strong financial results do not eliminate return questions

The second-quarter data itself is very strong. TSMC officially disclosed revenue of $40.20 billion, a year-on-year increase of 33.7%. In New Taiwan dollars, revenue grew by 36.0% year-on-year. Net profit was NT$706.56 billion, up 77.4% year-on-year.

The revenue structure also points to AI. Advanced processes, which include nodes of 7nm and below, already account for 77% of wafer revenue. The 2nm technology contributed 3% in its first quarter. High-end GPUs, custom AI chips, and high-performance computing chips all depend on this type of capacity.

The information conveyed by the financial report is clear. AI demand has entered TSMC's revenue structure, rather than being confined to the narrative of orders. The company also raised its 2026 revenue growth expectation to slightly above 40% and increased its full-year capital budget to $60 billion to $64 billion.

However, revenue growth rate is not the only variable. The higher the capital budget, the more questions investors will have: When will this money turn into capacity? When will capacity turn into revenue? Can revenue maintain past high-profit margins?

AI demand makes the expansion cycle heavier

TSMC's current expansion is not just a regular inventory cycle to replenish capacity; it is more like rearranging for AI supply from 2026 to 2028.

AI chips require not only advanced processes but also advanced packaging. Advanced packaging can be understood as efficiently assembling multiple chips together, allowing computation, memory, and data transmission to cooperate faster. For AI servers, packaging also determines the final supply.

The company disclosed that 70% to 80% of the annual capital budget will be allocated to advanced processes, about 10% to special processes, and 10% to 20% to advanced packaging, testing, and photomask segments.

The expansion in the U.S. is also progressing under the same logic. According to company announcements and media summaries, TSMC has committed about $265 billion in total investment in the U.S. after adding $100 billion investment in Arizona. The additional investment will be used for several advanced logic wafer fabs and advanced packaging facilities.

Wei Che-Chia emphasized that not raising prices suddenly and significantly should also be viewed within this framework. This does not mean TSMC lacks pricing power; rather, the company places more importance on long-term customer relationships and capacity binding. For Nvidia, AMD, and Broadcom, securing stable capacity is more important than short-term pricing.

U.S. expansion raises cost benchmarks

The expansion of investment in the U.S. is the most sensitive part of this revaluation. It reduces market concerns about the concentration of capacity in a single region, and strengthens the narrative of domestic manufacturing in the U.S. AI supply chain.

However, building factories in the U.S. is not a free safety net. Labor, construction, supply chain support, and management costs are usually higher than in Taiwan, and new factories must also go through yield and utilization ramp-up phases. The lower the yield, the higher the unit cost.

Thus, U.S. expansion has two implications. It helps alleviate geopolitical risks, making customers and the U.S. government more willing to see TSMC as part of AI infrastructure, but it may also compress overall profit margins, diluting the past cost advantages of "Taiwan's ecosystem plus ultra-high efficiency."

This will change TSMC's valuation anchor. In the past, the market was willing to pay for its technological scarcity and high profit margins. Now, it must also assess whether the post-Americanization capacity can maintain a sufficiently high return on invested capital.

Gross margin and cash flow score the expansion

What TSMC can currently demonstrate is that AI demand is strong, U.S. expansion is accelerating, and the company is entering a phase of higher capital expenditure. However, these conditions do not directly imply better long-term returns.

Short-term pressure is on gross margins. The company's conference call statements indicate that the steep ramp-up of 2nm in the second half of 2026 is expected to dilute gross margins by about 3 to 4 percentage points. The dilution of gross margins from overseas fabs is initially expected to be 2 to 3 percentage points but may later expand to 3 to 4 percentage points.

This indicates that margin pressure has already been incorporated into management's guidance. Even with fast revenue growth, the ramp-up of 2nm and the startup costs of overseas fabs may put phase pressures on margins. If the market values it based on high-margin leading advanced contract manufacturers, a downward revision of margins will amplify stock price volatility.

The mid-term variable is free cash flow. Financial officer Wendell Huang emphasized strong operating cash flow, indicating that the company is not expanding due to cash tightness. However, the $60 billion to $64 billion capital budget will still consume a significant amount of cash, and the space for dividends and shareholder returns will need to continue to be tested.

Long-term depends on the yields and utilization rates of U.S. fabs. If the advanced process and packaging capacity in the U.S. stabilize and ramp up while being continuously absorbed by AI customers, high capital expenditures will translate into stronger barriers. If AI capital expenditures slow after 2027, or if the costs at U.S. fabs remain high over the long term, TSMC's growth story will be repriced with a return discount.

What is currently most worth tracking for TSMC is not whether single-quarter revenue continues to reach new highs but rather the quality of profits after high investments. The AI supercycle has provided it with a reason to expand capacity, and the market will reassess whether this company can still generate the kind of high-quality returns it has in the past using gross margins, free cash flow, and the ramp-up progress at U.S. fabs.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。