TSMC to Expand in the US with $265 Billion Investment, AI Valuations Increasingly Focus on Cash Flow

TL;DR

• TSMC's Q2 results and full-year guidance reinforced AI chip demand, with the annual capital budget raised to between $60 billion and $64 billion.
• Market divergence centers on US fab construction costs, the 2nm ramp-up, and long-term capital returns; strong demand does not guarantee margin stability.
• Related tickers: TSM, NVDA, AMD, AVGO, ASML.

After releasing its Q2 earnings on July 16, TSMC raised its full-year revenue and capital budget guidance and continued advancing its US expansion plans.

This should have been a standard positive catalyst for the AI supply chain: high revenue growth, an increasing share of advanced nodes, initial revenue contributions from 2nm, and management framing US customer demand as multi-year structural demand. Yet the market's focus has shifted from "do orders exist" to "can orders translate into sufficient returns."

The reason is straightforward. Stronger AI demand necessitates heavier capital investment from TSMC. Building fabs, purchasing equipment, and expanding advanced packaging all require upfront spending. US fab construction also entails higher costs, longer ramp-up periods, and a more complex execution environment.

TSMC Chairman C.C. Wei offered a relatively optimistic outlook during the earnings call. He emphasized that the company's investments are driven by the multi-year demand of leading US customers and will maintain the reasonable gross margins required for long-term expansion, without sudden, sharp price increases due to supply constraints.

Strong Earnings Don't Eliminate Return Questions

The Q2 figures themselves were robust. TSMC officially reported revenue of $40.20 billion, up 33.7% year-over-year. In New Taiwan Dollars, revenue grew 36.0% year-over-year. Net profit reached NT$706.56 billion, a year-over-year increase of 77.4%.

The revenue mix also pointed firmly towards AI. Advanced process technologies (7nm and below) accounted for 77% of wafer revenue, while 2nm contributed 3% in its first quarter producing revenue. High-end GPUs, custom AI chips, and high-performance computing chips all rely heavily on this type of capacity.

The message from the earnings report was clear. AI demand has already permeated TSMC's revenue structure, not merely remaining a narrative in the order book. The company also raised its 2026 US dollar revenue growth expectation to slightly above 40% and increased its full-year capital budget to between $60 billion and $64 billion.

However, revenue growth is not the only variable. The higher the capital budget, the more investors will question: when will this spending translate into capacity, when will that capacity generate revenue, and can revenue sustain the high profit margins of the past?

AI Demand Intensifies the Expansion Cycle

TSMC's current expansion cycle is not a typical inventory-driven capacity replenishment; it's more akin to reshuffling the deck for the AI supply landscape of 2026 to 2028.

AI chips require not only advanced process nodes but also advanced packaging. Advanced packaging can be understood as efficiently integrating multiple chips to enable faster coordination between computation, memory, and data transmission. For AI servers, packaging is equally decisive for final supply.

The company disclosed that 70% to 80% of the full-year capital budget will be allocated to advanced process technologies, approximately 10% to specialty processes, and 10% to 20% to advanced packaging, testing, and mask-making operations.

The US expansion follows the same logic. According to company announcements and media summaries, after an additional $100 billion investment in Arizona, TSMC's cumulative committed investment in the US stands at approximately $265 billion. The new investment will fund several advanced logic wafer fabs and advanced packaging facilities.

Wei's emphasis on avoiding sudden significant price increases should also be viewed within this framework. It doesn't mean TSMC lacks pricing power, but rather that the company prioritizes long-term customer relationships and capacity lock-in. For Nvidia, AMD, and Broadcom, securing stable capacity is more crucial than short-term pricing.

US Expansion Raises the Cost Baseline

The scale of the US investment is the most sensitive aspect of this re-evaluation. It reduces market concerns about geographic concentration of capacity while providing a stronger onshore manufacturing narrative for the US AI supply chain.

However, building fabs in the US is not a cost-free safety net. Labor, construction, supply chain, and management costs are typically higher than in Taiwan, and new fabs must undergo yield and utilization rate ramp-ups during their startup phase. Lower yields translate directly to higher unit costs.

Thus, US expansion carries two implications. It helps mitigate geopolitical risk, making customers and the US government more willing to view TSMC as part of the AI infrastructure. Conversely, it may compress overall profit margins, diluting the cost advantages derived from the "Taiwan ecosystem plus ultra-high efficiency" model.

This will change TSMC's valuation anchor. In the past, the market rewarded its technological scarcity and high margins. Now, investors must also assess whether US-based capacity can maintain a sufficiently high return on invested capital.

Gross Margin and Cash Flow Score the Expansion

What TSMC can currently prove is that AI demand is robust, the US expansion is accelerating, and the company has entered a phase of higher capital intensity. However, these conditions alone do not guarantee better long-term returns.

The near-term pressure is on gross margin. Company guidance from the earnings call indicated that the steep ramp-up of 2nm in the second half of 2026 is expected to dilute gross margin by approximately 3 to 4 percentage points. The dilution from overseas fab ramp-ups is initially estimated at 2 to 3 percentage points, potentially widening to 3 to 4 percentage points later on.

This shows that margin pressure has already been incorporated into management's guidance. Even with rapid revenue growth, the costs associated with the 2nm ramp-up and overseas fab startup may put temporary pressure on profit margins. If the market values TSMC as a high-margin advanced foundry leader, a downward revision in gross margins could amplify stock price volatility.

The medium-term variable is free cash flow. CFO Wendell Huang emphasized strong operating cash flow, indicating the expansion is not driven by cash constraints. Nevertheless, a capital budget of $60 billion to $64 billion will consume significant cash, meaning dividends and shareholder returns will continue to face scrutiny.

Long-term, the outcome hinges on the yields and utilization rates of the US fabs. If advanced US process and packaging capacity ramps up steadily and is consistently absorbed by AI customers, high capital expenditure will translate into stronger competitive moats. Conversely, if AI capital spending slows after 2027, or if US fab costs remain persistently higher than expected, TSMC's growth story could be repriced lower due to a discount on returns.

For now, the most critical metric for TSMC is not whether single-quarter revenue can hit new highs, but the quality of profits generated after this heavy investment. The AI super-cycle has provided the rationale for expansion; the market will use gross margins, free cash flow, and the progress of US fab ramp-ups to reassess whether this company can still produce the high-quality returns of the past.