Skip to the content

TSMC says 'insane' AI demand is 'real'

 

This week's Fiery Flash is Taiwan Semiconductor Manufacturing Company, or TSMC. It is the world's largest independent semiconductor foundry, or computer chip bakery, producing integrated circuits for clients worldwide. As a pure-play foundry, it doesn't design its own chips but manufactures them for companies that do. Their clients include industry giants like Apple, NVIDIA, and AMD, whose products rely heavily on TSMC's advanced chip-making capabilities. This makes TSMC a critical player in the global tech industry, as many devices and systems we use daily depend on the chips they produce.

The difference between the company's lowest and highest share price this week amounted to 15%, with a weekly gain of approximately 10%. On Tuesday, TSMC's share price experienced some weakness, caught up in the downdraft from the surprise profit warning issued by the Dutch company ASML. However, it quickly became clear that ASML's weak 2025 outlook was more related to TSMC's competitors, Intel and Samsung, rather than TSMC itself. As a result, TSMC's share price regained some ground. The biggest uplift however came from the stronger than expected third-quarter results and positive outlook on Thursday morning.

The company reported third-quarter revenues of $23.5 billion, an increase of 36% compared to the same period last year and 13% over the previous quarter. It attributed this strong performance to robust demand in smartphones and AI applications for its industry-leading 3-nanometer and 5-nanometer technologies. These technologies refer to the size of the transistors on the chips—the smaller the nanometer measurement, the more transistors can fit on a chip, resulting in faster, more efficient, and more powerful semiconductors. The surge in investment in artificial intelligence (AI) is significantly boosting TSMC's business, even though it's starting from a relatively small base. TSMC defines "server AI processors" as the chips used in powerful computers to train and run AI programs. This includes graphics processing units (GPUs), AI accelerators, and central processing units (CPUs) that handle AI tasks like learning from data (training) and making predictions (inference). They exclude other areas like networking equipment, edge computing devices, or AI chips built directly into gadgets. The company now forecasts that revenue from these server AI processors will more than triple this year and is expected to account for a mid-teens percentage (ca. 15%) of their total 2024 revenue.

Further down the P&L the company achieved a gross margin of 57.8%, operating margin of 47.5%, and net profit margin of 42.8%. Gross profit grew by 45%, operating profit by 55%, and net profit by 51% compared to the same period last year. This substantial profit growth demonstrates an operational leverage effect, where profits increase at a faster rate than sales because fixed costs are spread over higher revenue.

The company's outlook exceeded market expectations, and perhaps most notably, C.C. Wei, the Chairman and CEO of TSMC, emphasized that the current strong demand for their products is genuine and just the beginning. He mentioned that one of his key customers, Nvidia, described the demand as "insane". Wei believes this high demand will continue for many years.

About the author

Siddy Jobe

Siddy Jobe

Siddy holds a Master’s degree in Economics from the University of Antwerp and a Master's degree in Financial Management from the Vlerick Business School. Passionate by innovation and entrepreneurship, he also participated to an Executive Master in Venture Capital at the Berkeley Haas School of Business. Prior to joining Econopolis, he managed the Investor Relations & Treasury office at Orange Belgium, a telecom company. Siddy also held the position of Telecom, Media & Technology analyst at a large Belgian Asset Management firm. Further, he is also active in the advisory board of StartupVillage and The Beacon, a business and innovation hub in the center of Antwerp focused on Internet of Things and Artificial Intelligence in the domains of industry, logistics and smart city. At Econopolis, he is Portfolio Manager of the Econopolis Exponential Technologies Fund.

comments powered by Disqus