Taiwan Semiconductor (TSM): fundamental analysis

2026-06-22 · By Lubin Danilo, founder of Lubin Investment

TSMC is the company our method rates highest: perfect quality, monopoly on advanced chips, structural AI growth. But it trades at 2.2 times its free cash flow because the market prices in a major geopolitical risk: military tension with China could halt its production. This trade-off is the core of the thesis.

• TSMC scores 10/10 in our fundamental rating model, the highest possible score.
• It manufactures 85% of the world's advanced chips (3nm, 2nm): Apple, Nvidia, AMD, and Qualcomm all depend on it.
• Intel Foundry and Samsung are 2 to 3 years behind technologically: there is no real substitute.
• The valuation of 2.2 times FCF is exceptionally low for a company of this quality.
• The reason: the Taiwan/China geopolitical risk is directly priced into the stock.
• CAPEX is massive to build new fabs in Arizona, Japan, and Germany (geographic diversification underway).

Some companies you analyze where everything holds together: the numbers are strong, the competitive position is solid, the growth is visible. And then there is TSMC. Taiwan Semiconductor Manufacturing Company sits in a category of its own. Our rating methodology gives it a perfect score. And yet the stock trades at a level that would suggest it is struggling. This paradox is not a market inefficiency. It is a deliberate decision by investors facing a well-identified risk.

In this article, I will break down this thesis starting from the fundamental question: why does a company with perfect quality trade so cheaply? The answer is honest, complete, and touches on something that few financial analyses dare to put front and center.

TSMC: the invisible monopoly that runs the world

Taiwan Semiconductor Manufacturing Company Limited (ticker: TSM on NYSE, 2330.TW in Taiwan) is the world's largest contract semiconductor manufacturer. Its market capitalization exceeds $900 billion, making it one of the ten largest companies in the world. But what makes TSMC truly unique is not its size. It is its technological monopoly.

TSMC operates as a foundry. It does not design its own chips. It manufactures chips for other companies. Its clients include Apple for iPhone processors, Nvidia for AI GPUs, AMD for datacenter processors, Qualcomm for mobile chips, and ARM for embedded architectures. In other words, virtually all the chips powering the global digital economy flow through TSMC's factories in Taiwan.

What no one else can do: TSMC masters the most advanced manufacturing nodes at 3nm and 2nm. These figures describe the fineness of transistors etched onto chips. The smaller, the more powerful and energy-efficient the chip. At these leading-edge technologies, TSMC accounts for roughly 85% of global production. Its two potential rivals, Intel Foundry and Samsung, are 2 to 3 years behind. In the semiconductor industry, a 2 to 3 year technology gap is an eternity.

AI drives demand: why TSMC is at the heart of the revolution

The artificial intelligence boom is not a passing trend for TSMC. It is a structural growth driver. Every large AI model trained by OpenAI, Google, Meta or Anthropic requires thousands of high-performance GPU chips. These chips are built by Nvidia. And Nvidia has them built by TSMC.

Demand for AI servers is exploding. Datacenters being built urgently worldwide need advanced chips. TSMC is the only bottleneck in this supply chain. Not because it is a transient constraint, but because it is structurally impossible to build equivalent manufacturing capacity in a few years. A state-of-the-art semiconductor fab takes 3 to 5 years to build and costs between $15 and $30 billion. The required technology (ASML's EUV lithography machines, for example) is itself produced in limited quantities.

The free cash flow generated by TSMC is structurally massive. Margins are solid. Growth is visible for several years ahead thanks to AI investment cycles, electric vehicles, and cloud infrastructure. On our ten rating criteria, TSMC validates them all. It is one of the rare companies to reach this level in our database.

What our method says: 10/10 across the board

Our analysis framework evaluates each company on ten fundamental criteria: revenue growth, operating profitability, free cash flow generation, balance sheet quality, return on capital, consistency of results, resilience during downturns, capital allocation policy, relative valuation, and business model quality. Each criterion is worth one point.

TSMC validates all ten. Growth is structural and driven by multiple megatrends. Margins are high and defensible through its technological monopoly. Free cash flow generation is massive. The balance sheet is sound. Return on invested capital is well above our thresholds. Results are consistent and predictable. Capital allocation policy (reinvestment in fabs, share buybacks) is rigorous. And the business model, that of a mission-critical foundry for the world's largest tech companies, is among the most defensible that exists.

Ten out of ten. It is not a score I give lightly. Across several hundred analyzed companies, only a handful reach this level. A score of 10/10 does not mean the stock is risk-free. It means the fundamental quality of the business is at the highest level our method can measure.

The Taiwan risk: the real reason for the low valuation

Here is the heart of the paradox. If TSMC scores perfectly and demand for its chips is only growing, why does the stock trade at 2.2 times its free cash flow? For comparison, an average-quality company in an ordinary sector often trades at 15 to 25 times FCF. TSMC is at 2.2 times. That is a discount of over 90% relative to what its fundamental quality alone would justify.

The answer is simple and blunt: the Taiwan risk. Almost all of TSMC's production sits on the island of Taiwan. And Taiwan is at the center of one of the most closely watched geopolitical tensions in the world, the one between China and the United States. Beijing considers Taiwan a rogue province it intends to reunify, by diplomacy if possible, by force if necessary. A military invasion of Taiwan, even partial, or a simple naval blockade, would be enough to halt TSMC's production.

The consequences of such an interruption would be catastrophic for the global economy. Not just for TSMC and its shareholders. For the entire global technology industry. For automakers dependent on chips. For telecom operators. For cloud providers. Everything containing an advanced chip depends, directly or indirectly, on Taiwan.

The market prices this risk rationally. It is not saying the invasion will happen. It is saying it is possible, and that if it does, the stock's value would be wiped out or severely impacted. This tail risk, even if its probability is low, justifies a meaningful structural discount. That is exactly what the 2.2 times FCF valuation reflects.

The diversification strategy: Arizona, Japan, Germany

TSMC is not ignoring this risk. For several years, the company has been investing heavily to build fabs outside Taiwan. In Arizona, TSMC is building multiple plants with financial backing from the US government under the CHIPS Act. In Japan, a fab is already in production. In Germany, a project is under development with support from the European Union.

These investments are positive for reducing geopolitical risk over the long term. But they come at a cost. The committed CAPEX is massive, tens of billions of dollars over the coming years. And these new fabs will not produce with the same yield as those in Taiwan, at least not initially. There is an operational cost to geographic diversification that TSMC's margins will need to absorb.

For investors, this is an ambiguous signal. Diversification reduces long-term risk, which should support valuation. But the high CAPEX weighs on FCF in the short and medium term. A multi-year horizon is therefore needed to invest in this thesis.

The 2.2× valuation: opportunity or trap?

Let us come back to the central figure: a valuation of 2.2 times free cash flow. For a company of this quality, that is historically very low. If geopolitical risk decreases (diplomatic agreement, de-escalation, accelerated geographic diversification), the market could quickly reprice the stock toward far higher multiples. The potential upside is considerable.

But if the geopolitical risk materializes, even partially, the current valuation does not protect the investor. A stock at 2.2 times FCF can still lose 50, 70, or 90% of its value if Taiwan's factories become inaccessible or inoperable. That is the fundamental trade-off of this thesis: asymmetric upside potential in the favorable scenario, extreme loss risk in the unfavorable one.

This is not a thesis for every investor profile. If you manage a portfolio that cannot afford a 50% loss on a single position, TSMC is not suitable. If you understand the risk, have a long-term view, and believe that geopolitical de-escalation is the most likely scenario, then the current valuation is one of the most interesting asymmetries in global markets.

My take on the TSMC thesis

TSMC is the kind of company you would buy with your eyes closed if it were based in Zurich or Singapore. Perfect quality, unassailable monopoly, structural growth driven by AI, valuation that leaves enormous room for upside. But it is based in Taiwan, and that detail changes everything.

In our method, the quality score is separate from the investment decision. A quality score of 10/10 does not automatically mean "buy." It means that if you take this geopolitical risk, you are taking it on the best possible company in its sector. The counterpart to this risk is a valuation that faithfully reflects the uncertainty. The market is not inefficient here. It is honest.

My personal approach: TSMC deserves a place in a diversified portfolio, with a weighting that reflects the investor's tolerance for geopolitical risk. It is not a position to overweight. It is a position to understand deeply before entering. The valuation of 2.2 times FCF is a conditional gift: you need to know what you are signing up for to accept it.

FAQ

Why does TSMC trade so cheaply despite a perfect quality score?

The market prices in the geopolitical risk linked to Taiwan. If military tension with China were to halt production, the stock's value would be severely impacted. This probability, even if low, justifies a structural discount. The 2.2 times FCF valuation reflects this risk rationally, not a market inefficiency.

What does the 2.2× P/FCF valuation mean in practice?

P/FCF (Price-to-Free-Cash-Flow) measures how many times the market values the company's annual free cash flow. At 2.2 times, you are paying less than two and a half years of free cash flow to own the stock. For a high-quality company, this level is historically very low.

Does TSMC have real competitors capable of replacing it?

No, not on the most advanced chips. Intel Foundry and Samsung are the two closest rivals, but they are 2 to 3 years behind on 3nm and 2nm nodes. In the semiconductor industry, this lag is structurally very difficult to close, as it requires massive investment, time, and expertise built over decades.

Why can't Apple, Nvidia, and AMD have their chips made elsewhere?

Because there is no alternative at the same performance level. The most advanced chips require the finest manufacturing nodes (3nm, 2nm) that only TSMC masters at scale. Switching foundries would mean either less-performant chips or several years of delay. For products like iPhones or AI GPUs, that is not acceptable.

Does TSMC's geographic diversification actually reduce the risk?

Over time, yes. Fabs in Arizona, Japan, and Germany reduce dependence on Taiwan. But this diversification takes time (3 to 5 years per plant) and costs a great deal in CAPEX. In the short term, more than 90% of advanced production remains in Taiwan. The geopolitical risk will not disappear for several years at minimum.

Voir l'analyse TSM sur Lubin Investment

About the author

Written by Lubin Danilo, founder of Lubin Investment. A self-taught individual investor, I have analyzed stocks through their fundamentals for several years and invest my own money with this method. I codified it into a tool that judges a company's quality and its price separately, using criteria drawn from the financial literature (Warren Buffett, Michael Mauboussin, Aswath Damodaran).