Is TSMC still the biggest winner in the second half of the AI era?

```

As tech giants scramble to buy AI chips and global advanced manufacturing capacity becomes increasingly tight, TSMC’s core beneficiary status in this arms race is becoming more solid—while its stock price valuation remains inexpensive.

Microsoft, Meta, Alphabet, and Amazon have a combined capital expenditure plan of up to $725 billion this year, with most of it flowing into AI chips. Such enormous demand directly benefits chip manufacturers, especially the world's largest foundry, TSMC. TSMC’s gross margin in the first quarter has climbed from about 59% a year ago to about 66%. This year’s capital expenditure is expected to reach the upper limit of the forecast range of $52–56 billion.

TSMC CEO C.C. Wei said last month he was "full of confidence" that this year's revenue growth will exceed 30%. Meanwhile, Nvidia’s procurement commitments to TSMC have surged from about $16 billion two years ago to over $95 billion in the latest fiscal quarter ending this January, reflecting customers’ strong demand for capacity.

Although N2 process mass production ramp-up and U.S. factory costs will temporarily put pressure on gross margins, this is a proactive layout rather than a structural disadvantage. TSMC’s current stock price corresponds to a forward price-to-earnings ratio of about 21x, below the Philadelphia Semiconductor Index component average of about 26x. With multiple strengths, the valuation is highly attractive.

Gross Margin Expansion, Full-Capacity Operation Is the Core Driver

TSMC’s most direct competitive advantage is reflected in its rapidly expanding gross margin. When sales growth exceeds expense growth, gross margin rises—this is especially evident when TSMC’s factories are running near full capacity, since high capacity utilization can effectively dilute the high fixed costs of chip plants.

TSMC does not produce memory chips, but it almost monopolizes the manufacturing of all other types of advanced chips, covering Nvidia’s market-leading AI chips and Apple’s smartphone chips. Strong demand has led some customers to lock in capacity ahead of time, even prepaying billions of dollars to secure chip supply. Nvidia’s procurement commitment in the latest fiscal quarter ending this January has exceeded $95 billion, with a significant proportion going to TSMC, compared to about $16 billion two years ago.

TSMC CFO Wendell Huang told analysts last month that the company’s gross margin will narrow in the second half of this year—but the cause is not worrisome. TSMC is advancing mass production of the latest generation N2 process, with natural cost increases during the ramp-up phase that will recede once production stabilizes.

Another source of pressure is the U.S. factory. Operating a chip fab in the U.S. costs more than in Taiwan, inevitably dragging down overall profitability. Yet TSMC’s strategy is clear: expanding geographically both hedges against Chinese geopolitical risk and brings it closer to U.S. clients like Apple and Nvidia who wish to increase local suppliers.

In the medium to long term, as advanced processes mature, gross margin is expected to return to a growth track. TSMC’s plans to expand N3 process capacity in Japan and the U.S. will also provide extra support—N3 uses older, lower-cost manufacturing equipment, can handle AI chip demand, and helps improve capital expenditure efficiency.

Huge Capital Expenditure, but Faster Revenue Growth

This year’s capital expenditure approach the upper limit of $56 billion, which may worry some investors about TSMC over-expanding. In the semiconductor industry, excessive capacity expansion often brings backlash—once demand cools, manufacturers are forced to maintain large idle capacity, increasing costs.

But for now, TSMC has not exceeded a reasonable expansion boundary. C.C. Wei made clear last month that this year’s revenue growth will exceed 30%, higher than the capital expenditure growth rate, meaning expansion remains controllable. Customers locking in capacity and prepayments also strongly endorse future revenue, making TSMC's revenue growth a highly certain expectation.

In advanced process, TSMC has virtually no substantive competition. Samsung’s foundry business ranks second globally, but its revenue scale is far behind TSMC. Intel and Japanese startup Rapidus are striving to enter the market, but progress is limited. Elon Musk’s recent Terafab advanced process project plans to use Intel to produce advanced chips, but at best that is a distant vision.

High demand combined with nearly monopolistic market position give TSMC considerable pricing power. As customers shift to more advanced processes, higher average prices will continue to boost revenue in coming years. However, TSMC executives remain cautious. C.C. Wei said last month: "We will not sharply adjust pricing, just ensure that customers can succeed in their respective markets."

Despite owning many advantages, TSMC’s valuation remains relatively low-key. Its current stock price corresponds to a forward P/E ratio of about 21x, below the Philadelphia Semiconductor Index component average of about 26x and much lower than Intel and AMD.

Risk Warning and DisclaimerThe market carries risk; investments require caution. This article does not constitute personal investment advice, nor does it take into account individual users’ specific investment goals, financial situation, or needs. Users should consider whether any opinions, views, or conclusions in this article suit their particular circumstances. Investing based on this article is at your own risk. ```