A recent report from TSMC, the chip foundry giant, indicates that the company is on track to record better-than-expected results in the fourth quarter of last year, reinforcing market confidence that global spending on artificial intelligence (AI) will remain strong in 2026.
On Friday (January 9) local time, TSMC announced on its official website that the company's revenue in December 2025 was NT$335 billion, a decrease of 2.5% compared to November last year, but an increase of 20.4% compared to the same period in 2024.
TSMC also stated that its total revenue for January to December 2025 will be NT$3,809.05 billion, representing a 31.6% increase compared to the same period in 2024.
Based on the latest monthly data, TSMC's revenue in the fourth quarter of last year reached NT$1.05 trillion, an increase of about 20% year-on-year , higher than the market's average expectation of NT$1.02 trillion.
TSMC is scheduled to release its full quarterly earnings report next Thursday (January 15th) and will provide capital expenditure guidance for 2026. Last year, the company benefited significantly from customers stockpiling chips before the US tariffs took effect.
Since the beginning of the year, several brokerages, including JPMorgan Chase, have raised their target price for TSMC, citing its strong revenue growth prospects and continued improvement in profitability.
Currently, advancements in AI technology have completely offset the decline in demand for consumer electronics following the COVID-19 pandemic, and TSMC, with its central role in the manufacturing of advanced AI accelerators, has become one of the biggest beneficiaries.
As a major chip supplier for Apple, TSMC may also benefit from the strong sales of the iPhone 17 released last September.
Driven by demand for data center chips, executives at Nvidia, a major customer of TSMC, issued positive signals this week, expressing optimism that the company's revenue is more optimistic than predicted, partially alleviating concerns that infrastructure construction is outpacing AI applications.
Earlier this week, Jensen Huang showcased NVIDIA's next-generation AI platform, Rubin, at CES 2026 in Las Vegas, saying that products based on the platform will be available through partners in the second half of 2026.
Global tech giants, including Microsoft and Meta, have collectively invested more than $1 trillion in data center projects, betting on continued growth in AI applications, but investors are worried that the capacity under construction will eventually exceed actual demand.
In addition, many data center investment arrangements exhibit a “circular structure,” where OpenAI and a few publicly traded tech giants invest in and spend on each other, a model that has unsettled Wall Street.
