People

The People Running TSMC

Governance grade: A−. The board is genuinely independent, the audit and compensation committees are run by senior global operators (former CEOs of NXP-era BT, Applied Materials, Xilinx, Xerox, Sunoco, plus a former MIT president), and the CEO's pay is heavily variable. The one structural tension is the inverse of most concerns: the people running the world's most strategic factory own almost none of it — combined insider holdings outside the Taiwan government's National Development Fund are 0.23% of shares outstanding.

1. The People Running This Company

TSMC is run by a small group of long-tenure technologists. Founder Morris Chang retired in 2018; Mark Liu retired as Chairman in 2024. C.C. Wei now holds both Chairman and CEO seats, supported by two Co-COOs and a CFO who have each spent their entire careers inside the company.

What matters here:

• C.C. Wei is the consequential figure. Ph.D. EE from Yale; joined 1998 from Chartered/STMicro; ran technology, then operations, then sales/marketing before becoming Co-CEO in 2013, CEO in 2018, and Chairman/CEO in June 2024. Capability is not in doubt: every advanced-node ramp (N7 → N3) since 2018 happened under his operational ownership.
• Co-COO bench is real. Y.P. Chyn (Operations, 39 years) and Y.J. Mii (R&D, 32 years) are full operating partners, not figureheads. This is the de-facto succession structure — Mark Liu was promoted from Co-COO to Chairman, and the same pattern is now pre-positioned.
• F.C. Tseng (39 years; former Vice Chair until 2018) remains on the board and holds 0.11% — the closest thing to a founder-stake on the board.
• CFO Wendell Huang has held the role since 2019; his FY25 cash + stock package of NT$339.5M is roughly 1/7 of Wei's, which is normal for a foundry of this scale.

2. What They Get Paid

Total FY2025 executive compensation across 27 named officers was NT$9,060M (US$288.8M). The CEO took NT$2,422.7M (US$77.2M) — a single-individual record among Taiwan-listed companies, but driven almost entirely by the bonus and LTI pool, which is tied to profit and total-shareholder-return metrics under the 2025 LTI plan.

Is the pay sensible?

• Variable share is high. Of CEO compensation, base salary is 0.7%; the rest is bonus + stock + LTI, decided annually by an all-independent Compensation Committee. This is the right structure for a CEO at a company of this scale.
• Pay-for-performance is plausible. TSMC's articles cap director compensation at 0.3% of annual profit and require ≥1% of profit be distributed to employees. Pay scales with the P&L, not in spite of it.
• The absolute level is large for Taiwan, modest for the role. US$77M for the CEO of the world's most strategically important manufacturer is roughly half of what comparable US mega-cap CEOs (Nvidia, Broadcom, Apple) earn — and Wei runs a more capital-intensive, more geopolitically exposed operation. The NT$ headline reads shocking only because Taiwan's CEO pay distribution is compressed.
• The cohort scales. The 25 non-CEO/CFO officers earned NT$6,298M combined (~NT$250M / US$8M each on average) — appropriate for the bench depth.

3. Are They Aligned?

This is where the case requires the most care. Outside of the Taiwan government's National Development Fund (NDF), insiders own essentially nothing.

Ownership and control. The 6.38% NDF stake is the largest single position; it acts as a political/strategic anchor, not a profit-maximizing insider. Outside of NDF, every D&O combined owns 0.23%. The CEO personally owns 0.03% — about US$606M at current prices, which sounds large but is roughly 8× his single-year compensation. There is no founding-family blockholder, no promoter-style entity, and no dual-class share structure. Voting rights are one-share-one-vote.

Insider buying and selling. TSMC executives don't file Form 4s (Taiwan disclosure regime; not required for foreign private issuers). Issuer purchases of equity in 2025: none ("Not applicable" per Item 16E). Recent third-party trackers report 5 small insider buys in the 90 days to mid-May 2026 totaling roughly US$709K — all buys, no sells — but these are ESPP-routed transactions, not outright open-market accumulation. The actionable read: no insider selling pressure, but no conviction buying either.

Dilution / option grants. TSMC has no stock-option plan for D&Os ("None of our directors and executive officers owned any stock option" per the 20-F). The RSA plans (2022, 2023, 2024 Rules) authorized up to 13.5M shares combined — roughly 0.05% of the 25.93B share count. The 2022 plan reclaimed more than it vested. Share-based compensation is a rounding error on the cap table; share count has barely moved.

Related-party behavior. Four affiliated entities are disclosed:

Each related-party relationship is with a minority-owned affiliate at well-understood economic terms; aggregate flows are under 2% of either revenue or COGS. These are technology and capacity-sharing relationships dating back to the 1990s, not value-extraction conduits. Nothing material.

Capital allocation behavior. TSMC has not bought back stock — Item 16E is "Not applicable" for 2025. Instead, the company has been paying a steadily rising dividend (the source of ~25% of TSMC's annual profit return to shareholders) while reinvesting the remainder at high incremental returns (return on equity ~30%+ for FY2025). For a foundry plowing tens of US$ billions into Arizona, Kumamoto, and Dresden every year, prioritizing investment over buybacks is the correct capital-allocation choice. Dividend is paid quarterly, in cash, with no special dividends or share splits used to manage optics.

Skin-in-the-game: 6/10. Pay structure is aligned (high variable, low base, LTI tied to relative TSR), capital allocation is shareholder-friendly, related-party flows are clean, and there is zero promoter pledging or option dilution. What holds the score below 8 is the absolute level of personal ownership: a US$77M-a-year CEO who has worked here 28 years owns about US$606M of stock. That is real money, but it is one-quarter of one year's free cash flow for TSMC — the alignment comes more from the bonus formula than from a meaningful equity stake. For a company of this strategic weight, that is acceptable, not exceptional.

4. Board Quality

Reading the board:

• Real independence, not formal. Seven of ten directors are independent. Audit & Risk and Compensation Committees are 100% independent. The Chairs of all three committees — Bonfield (Audit), Splinter (Comp), Lin (Nom/Gov/Sust) — are heavyweight independents.
• Semiconductor operating expertise is unusual. Splinter (Applied Materials CEO, ex-Intel EVP), Gavrielov (Xilinx CEO), Bonfield (NXP Chair), Reif (MIT EE professor and former President) — four directors who could each independently challenge management on technology, capex, and yield issues. This is the highest concentration of semi-industry expertise on any global-foundry board.
• Domain coverage outside semis is strong. Burns (Xerox CEO, ExxonMobil director), Elsenhans (Sunoco CEO, Saudi Aramco director), Lin (former Taiwan Finance Minister and Premier) cover capital markets, energy/geopolitics, and Taiwan public-policy interface respectively — exactly the surfaces TSMC needs covered as it builds in Arizona, Japan, and Germany.
• The NDF director is governance-relevant. Chun-Hsien Yeh (since Sept 2025) is currently the Minister of the National Development Council and convenes the NDF Management Committee. He sits at the board not as a director in personal capacity but as the Taiwan government's representative — a structural feature, not a fault, but one to watch as US-Taiwan policy diverges.
• Tenure mix is balanced. Three independents have ≤2 years (Burns, Elsenhans, Lin) — recent refreshments. Bonfield (24y) is the long-tenured anchor; the audit committee chair role gives that depth value, though tenure beyond 15 years is generally considered an independence yellow flag.

5. The Verdict

Governance grade: A−.

Strongest positives.

• Pay is variable and formula-driven; the Compensation Committee is fully independent; LTI is tied to relative TSR.
• Board has unusually deep semiconductor operating expertise — four directors who have personally run chip companies or research institutions.
• No options, minimal RSA dilution, no buybacks (a feature, not a bug, given the capex cycle), no founder-family blockholder, no dual-class shares.
• Related-party flows are immaterial; the affiliate web (VIS, SSMC, GUC, Xintec) is operationally rational and economically small.
• TSMC is the plaintiff, not the defendant, in current trade-secret litigation (Wei-Jen Lo / Intel; Tokyo Electron) — a sign the internal IP-control machinery is functioning.

Real concerns.

• Insider ownership outside the Taiwan government is 0.23%. The CEO holds 0.03%. Alignment runs through the pay formula and the share price, not through a meaningful personal equity stake.
• The Taiwan government (NDF) is the single largest holder at 6.38% — strategic, but it introduces a state-policy lens that does not always optimize for outside shareholders (e.g., onshore-vs-offshore capex pacing, advanced-node export controls).
• Combined Chair + CEO role in C.C. Wei since June 2024 reverses the prior separation (Chang/Liu, then Liu/Wei). A truly best-practice governance setup would name a Lead Independent Director with explicit powers; the 20-F does not flag one by name.

What would change the grade.

• Upgrade to A: split the Chair/CEO roles again, or name and empower a Lead Independent Director; raise CEO direct ownership to ≥US$1B (about 0.05%).
• Downgrade to B+: material new related-party transactions; a successful trade-secret defection that suggests systemic IP-control weakness; or a politically-driven capital-allocation decision (e.g., a US-blocking dividend or domestic-only capex commitment) that places state interest above shareholder return.

The single thing most likely to move the grade is the Chair/CEO consolidation under Wei — it is the one structural choice this board has made that runs against best practice, and the absence of a named Lead Independent Director is the cleanest patch available.