TSMC Valuation: Hidden in Plain Sight (Part 3)
A look into TSMC 4Q25 blowout results and current valuation
Disclaimer: This is not investment advice. As usual, please do your own due diligence before investing. If you lose money, I will offer sympathy, but no refunds.
Introduction
In Part 1: The Physics of Money, we explored the almost immutable laws of physics and economics that gave TSMC its initial moat. In Part 2: Why I Broke Character, I walked through the qualitative reasons that forced me to deviate from my usual state of mind to own this company.
Now, in Part 3, we dig into the TSMC current valuation. But before doing that, it is useful to present the case for why the future is likely to be significantly better than the past (barring a Black Swan event like a Chinese invasion of Taiwan).
Key Takeaways from the Q4 2025 Earnings Call
The market often views semiconductor manufacturing as a cyclical, capital-intensive grind (and it is!). However, the Q4 2025 earnings call reaffirms an additional layer: TSMC is no longer just a manufacturer; it is a key enabler of a technological revolution. Management’s commentary from the recent call was unusually bullish, effectively declaring the start of a new, sustained growth phase.
1. Expecting a long AI Cycle
While the market fears an AI bubble, TSMC sees a structural megatrend. CEO C.C. Wei provided a rare glimpse into his confidence level, noting that internal productivity gains from AI are already visible. During yesterday’s conference call, he said:
You are asking if AI demand is real. I am also very nervous about it because we have to invest USD 52 billion to 56 billion. If I didn’t do it carefully, that would be a disaster for TSMC.
Internally, we also see productivity improvements of 1%–2% using AI. In my view, AI is real and growing into daily life. It is a megatrend. As for how long the cycle lasts, I don’t know if it will be 3, 4, or 5 years, but it looks endless. No matter what, TSMC relies on fundamental technology leadership. We project a 25% CAGR, and we used to be conservative. [emphasis added]
2. Financial Firepower & Guidance
The numbers support the narrative. TSMC raised its long-term forecasts significantly:
AI Accelerator Revenue CAGR (2024–2029): Raised to mid-to-high 30%.
Overall Revenue CAGR (2024–2029): Now expected to approach 25% (USD terms).
2026 Outlook: The company forecasts full-year revenue growth close to 30%.
Despite the capital intensity, profitability remains elite. Management reiterated that 56%+ Gross Margins and High-20s ROE are achievable through the cycle. For Q1 2026 alone, gross margins are guided to ~64%, driven by manufacturing excellence that is outpacing inflation.
3. The Moat is Widening (Capex as a Barrier to Entry)
The sheer scale of investment required to compete is becoming prohibitive for peers.
Capex: TSMC plans to spend $52–$56 billion in 2026.
Historical Context: In the last 5 years, Capex totaled $167 billion and R&D $30 billion.
Complexity: Management noted that the Capex required for 100 wafers-per-month of N2 (2nm) is substantially higher than N3.
This escalating cost structure is a feature, not a bug—it locks out competitors and cements TSMC’s near-monopoly on the leading edge.
The Dual Engine: A New Anchor Client
For the last decade, TSMC’s growth narrative was dominated by the “smartphone cycle,” primarily driven by Apple. We are now entering a “Dual Engine” era.
TSMC and Apple share a symbiotic relationship. Apple provided the massive volume needed to ramp new nodes (N7, N5, N3) and pay for the initial yield learning. In return, TSMC gave Apple a power/efficiency lead over Android. This relationship remains the bedrock of TSMC’s scale.
Now Nvidia has emerged as the new anchor client, but with a different profile. While Apple drives wafer volume, Nvidia drives complexity and value.
Value Over Volume: AI chips are much larger and command higher pricing power.
Packaging Dependency: Unlike smartphone chips, AI accelerators are useless without advanced packaging (CoWoS). This locks Nvidia into the TSMC ecosystem even more tightly than Apple.
Insatiable Demand: C.C. Wei noted that “TSMC silicon is the bottleneck, not power.” This indicates that demand is uncapped by consumer wallets, but rather by infrastructure build-out speeds.
More Than Chips: The Packaging Moat
TSMC’s dominance in advanced packaging is an anomaly. Historically, “packaging” was a low-margin, afterthought done by third-party companies like ASE or Amkor. Fab companies like TSMC considered it “dirty work” beneath their high-margin business models. TSMC became dominant because they were the first to realize that packaging was no longer a soldering problem; it was a lithography problem.
The company’s bet on advanced packaging (CoWoS, SoIC) has been in the making for 10 years, and it is finally paying off:
Revenue contribution from advanced packaging last year was close to 10% (about 8%). For this year, we expect it to be slightly over 10%, growing faster than the corporate average over the next 5 years. regarding CapEx, yes, it is now 10%–20%.
Wendell Huang (TSMC CFO) during 4Q25 earnings call
This is a critical defensive moat. Even if a competitor could match TSMC’s silicon lithography (which they will struggle to), they may still fail to deliver a working AI system without the accompanying advanced packaging capacity.
Valuation: The “Taiwan Discount”
Note: For this valuation, I am using the Taiwan-listed shares (Ticker: 2330), which historically trade at a ~20% discount to the US-listed ADRs (Ticker: TSM).
We just discussed the qualitative data suggests we are nowhere near the peak of this cycle. But does the price reflect this?
Let’s do some napkin math:
Price (2330.TW): NT$ 1,690 (at time of writing)
2025 EPS: NT$ 66.25
2026e EPS: >NT$ 86 (baseline; assumes that 30% revenue growth with no margin expansion)
If we apply a conservative NT$ 86 EPS estimate for 2026—assuming no operating leverage despite the guidance for 30% top-line growth1 —the stock trades at just ~20x Forward Earnings.
Consider what you get for that multiple:
Hyper-Growth: Revenue compounding at 25-30% annually.
Monopoly Power: The main supplier of the 21st century’s oil (compute).
Elite Profitability: Gross margins structurally above 56%.
Fortress Balance Sheet: A net cash position that further de-risks the entry price.
The ‘War Tax’ Reality Check: For comparison, ASML—which holds the ‘Physics Monopoly’ on lithography—trades at roughly 40x earnings. The market is effectively applying a 50% discount to TSMC mostly for geopolitical risk. You are getting the same growth engine at half the price.
Ultimately, I view TSMC not as a technology bet, but as an infrastructure play. It is the toll road for the digital age. Whether the future belongs to Nvidia, AMD, or a startup we haven’t heard of yet, they all have to pay the toll to pass through Hsinchu.
The market is currently distracted by the geopolitical risks of the toll booth’s location. That fear has created a dislocation between price and value. As long as that gap exists, I am happy to keep collecting the toll.
TSMC management is usually conservative and that shows on FY26 guidance. Note that 1Q26 revenue guidance represents a 4% sequential increase or a 38% year-over-year increase at the midpoint. Management also guided for margin expansion on 1Q26.
Great analysis!