TSMC's Q4 earnings reignite AI bubble debate

Taiwan Semiconductor Manufacturing Company's shares soared 5% intraday on Thursday, reaching a new 52-week high. This followed its fourth-quarter earnings report on January 15, which fueled optimism about artificial intelligence.

As it beat Wall Street expectations, buoyed by AI investment from hyperscalers, investors are also piling into the stock because of its future positive guidance. The semiconductor manufacturer has seen its stock price rise 69% over the year.

In its latest Q4 report, it noted diluted EPS of $3.14, beating the $2.94 consensus. Revenue for the quarter came at $33.73 billion, representing a 25.5% increase year-over-year. 

The growth underscores rapid adoption of TSMC’s most advanced manufacturing nodes. The combined sales of its 3-nanometer, 5-nanometer, and 7-nanometer products accounted for 77% of the company’s total wafer revenue. 

• 3-nanometer technology accounted for 28% of total revenue
• 5-nanometer technology accounted for 35%
• 7-nanometer technology accounted for 14%

And Wendell Huang, Senior VP and CFO of TSMC, expects this “strong demand for our leading-edge process technologies” to continue into the first quarter of 2026.

Veteran analyst Chris Versace noted in TheStreetPro that TSM’s AI-related revenue compound annual growth rate, up mid-to-high 50% between 2024 and 2025, “is a key driver in TSM’s overall revenue CAGR of mid-to-high 30% over the same time frame. That level of AI growth tells us AI-related revenue at TSM will ramp significantly over the coming years and will represent far more than the high-teens total revenue percentage it did in 2025.”

Here is how revenue generation looked geographically for TSMC:

• North America accounted for 74% of total revenue in Q4
• Asia Pacific accounted for 9%
• China also accounted for 9%
• Japan and EMEA(Europe, Middle East and Africa) both accounted for 4% each of the total revenue.

Analysts react to TSMC's 2026 outlook guidance

The gross margin of 62.3% and the operating margin of 54% indicate significant profitability for the company, a key focus of analysts. 

However, to maintain these gains, TSMC is also planning an unprecedented level of investment. In its 2026 guidance, the company forecast Capital Expenditures of $52 billion and $56 billion, with a further leg up to $58 billion in 2027. For context, its CapEx for 2025 was $40.90 billion.

More Tech Stocks:

• Morgan Stanley sets jaw-dropping Micron price target after event
• Nvidia’s China chip problem isn’t what most investors think
• Quantum Computing makes $110 million move nobody saw coming
• Morgan Stanley drops eye-popping Broadcom price target
• Apple analyst sets bold stock target for 2026

The robust reports and the company’s outlook, which also drove the stock's movement on Thursday, prompted analysts to increase their bets on this semiconductor manufacturer. 

Bank of America raised its price target from $430 to $470, keeping a Buy rating.

Needham also raised its price target to $410 from $360, while maintaining a Buy rating, after TSMC reported strong earnings and guided key 2026 metrics above Street estimates. Needham further notes that the company’s Q1 2026 revenue is expected to grow 4% sequentially, which defies historical seasonality as Q1 is usually considered a lull period for the smartphone industry. 

Analysts attribute this to TSMC’s meticulous capacity planning, which enabled it to ship more wafers to High Performance Computing (HPC) customers in the first half of the year, as noted by TheFly.

AI Bubble or AI Megatrend?

TSMC’s profitable quarter and positive outlook have once again raised questions about the AI bubble, AI expansion, investment, and the exploding AI market. Especially considering that the TSM management refers to AI as a “megatrend”, as noted by Versace in TheStreetPro.

TSMC leadership further noted that major cloud service providers are proactively securing manufacturing capacity to support business, leading to “conviction in the multi-year AI megatrend,” which remains strong, and that “demand for semiconductors will continue to be very fundamental” to the global economy.

Supporting this outlook is BofA’s global research, which estimates AI-related capex to surpass $1.2 trillion by 2030, a threefold increase from 2025.

The AI buildout currently lasts only three years, while prior platform cycles historically lasted 5-10 years. Thus, hyperscalers are seeing this AI spending as mission-critical, as evidenced by near-zero data center vacancy and full construction pipelines. Further cementing the industry’s conviction in a long-term AI-driven expansion.

Related: Analyst says 'buy the dip' in this bank stock