Capital-intensive winners can look “expensive” at the exact moment they’re creating the most value. Free cash flow compresses as capex surges, even while ROIC stays far above WACC. Today’s artilce shows you how to value those businesses anyway—using a quality lens and a mini case on TSMC, the world’s leading foundry. TL;DR
Why it mattersWhy it mattersQuality investing is about owning businesses that compound value because their returns on invested capital consistently exceed their cost of capital. When a company can reinvest large amounts at attractive returns, time becomes your ally. TSMC is a prime example in today’s market: it operates at the critical junction of AI and high‑performance computing, it benefits from scale and process leadership, and it has customer relationships that are hard to displace. The catch is that this advantage is capital‑hungry. Periods of heavy investment can temporarily mask underlying economics, especially if you focus only on near‑term free cash flow. That is exactly where investors can misread the story. For 2025, management has signaled another big build year with capital expenditures guided to roughly $38–$42 billion and USD revenue growth around 30%. On the surface, the combination sounds contradictory—why would free cash flow not surge with that growth? Because the fab buildout and node transitions front‑load cash outlays before they show up as depreciation, which means free cash flow can lag the income statement for a while. The key to quality investing here is distinguishing between structurally weak unit economics and a cyclical cash trough created by reinvestment. If you believe TSMC’s returns remain comfortably above its cost of capital and demand visibility is durable, then heavy capex is not a red flag; it is the engine that fuels the next leg of compounding. Core conceptsROIC vs WACC and value creation
Reinvestment runway and capital intensity
Unit economics and per-share metrics
Moat durability and evidence
Valuation Monday: frameworks and scenariosOn Valuation Monday, the goal is to translate business quality into numbers you can underwrite. With TSMC, the starting point is to hold two ideas in mind: near-term free cash flow will be dampened by the 2025 capex plan of $38–$42 billion, while underlying earnings power benefits from roughly 30% USD revenue growth as AI/HPC ramps up. A sensible way to reconcile those is a reverse DCF. Instead of guessing today’s “fair value,” you ask what set of growth, margin durability, and reinvestment assumptions would justify the current price. The exercise works like this. First, anchor the near term in what management has already guided: strong top-line momentum into 2025, but cash conversion suppressed by the capital expenditure cycle. Next, extend that into a mid-cycle view where capital expenditure intensity normalizes as the large fabs and advanced nodes reach scale, and depreciation begins to catch up with spending. That’s typically when free cash flow inflects.
Finally, check that the implied long‑run economics are consistent with a business whose returns sit above its cost of capital—i.e., reinvestment still creates value rather than merely sustaining the base. If the reverse DCF implies that investors must assume permanently elevated capex with no relief, or that growth fades faster than the order book suggests, the bar may be too high. If, instead, it assumes a reasonable glide path—strong 2025 growth, followed by a couple of years of heavy spending, then normalization, you’re evaluating the business on its true economic engine rather than a single year’s cash flow. Worked example: TSMCStart from the knowns. Use 2025 as a bridge year in your model with USD revenue growth around 30% and capex at the midpoint of guidance, about $40 billion. That setup produces healthy operating results but a free‑cash‑flow picture that looks tight because investment is running ahead of depreciation. Treat that as a feature of the cycle, not a flaw in the business. Roll the model forward and let capital expenditure intensity step down as the current wave of fabs and advanced nodes stabilizes. As utilization improves and yields mature, operating cash flow benefits from scale while cash outlays begin to fall toward maintenance plus targeted growth. That’s when free cash flow expands, not because the business suddenly changed, but because timing differences between spend and accounting catch up. In this mid‑cycle state, sanity‑check that returns remain above the cost of capital; if they do, reinvestment continues to add value, and a non‑zero terminal growth rate is defensible. Now pressure‑test the path. In the base case, 2025 growth is projected to be near 30%, capex is expected to be around $40 billion, and capex intensity is anticipated to ease as planned, allowing free cash flow to inflect as fabs ramp up. In a downside scenario, utilization lags or tool delivery schedules push capital expenditures higher for longer, which keeps free cash flow subdued and pushes out the inflection point. In an upside case, AI/HPC demand stays hotter for longer, the ramp is smoother, and process leadership translates into faster mix shift to advanced nodes, accelerating cash conversion once the peak spend is behind you. Across all cases, the signal you’re watching is the relationship between returns and reinvestment. As long as ROIC remains comfortably above the cost of capital, heavy spending is a bridge to more earnings power rather than a hole in the balance sheet. The practical takeaway is to judge TSMC on normalized free cash flow and mid-cycle economics, rather than on a single year when capital expenditures peak. Using the same financial anchors—capex of $38–$42 billion for 2025 and USD revenue growth of around 30%, the story remains consistent: this is a high-quality, capital-intensive compounder where the cash flow timing can obscure, but does not erase, the underlying value creation. How to do this yourselfSteps and sources:
Pitfalls and false positives
Checklist / one-pager
Glossary
Further reading
References
DisclaimerEducational content, not investment advice. You're currently a free subscriber to School of Investing. For the full experience, upgrade your subscription.
|
Monday, October 6, 2025
TSMC’s AI-Driven Buildout: Translating Capital Intensity Into Intrinsic Value
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment