People

People & Governance

Verdict (B). AMD is a widely-held, professionally-managed company with a technically deep board, robust pay-for-performance structure, and clean related-party record — but two real frictions: (1) Dr. Lisa Su has been a one-way insider seller (78 disposals worth $184M across 2025–2026, zero open-market purchases), and (2) Chair and CEO roles are combined. The Lead Independent Director (Nora Denzel) and a 7-of-8 independent board mitigate but do not eliminate the entrenchment risk a 12-year combined-role CEO carries at all-time highs.

Who runs AMD

The bench around Su is unusually deep for a chip company of this scale. Papermaster's CTO seat (since 2011) has spanned the entire Zen architecture run; Norrod owns the data-center business line that delivered the MI300/MI350 ramp; Hu brought a clean Marvell-to-AMD CFO transition in 2023; Grasby got a one-time promotion/retention award in February 2025, signaling the board is paying to keep the commercial engine intact. AMD also added Phil Guido as Chief Commercial Officer in 2024 (the proxy notes a single late Form 4 filing on his initial open-market AMD purchase — administrative, not substantive).

The single-point-of-failure risk is real. AMD has been a Lisa Su story since 2014 — the stock has compounded roughly 100×. Su is 56 and shows no signs of leaving, but with a CEO this central to the equity story, succession depth is what protects shareholders. Norrod and Papermaster are credible internal heirs; both are 60+, which means the next COO/CEO pipeline is also worth tracking.

Compensation — heavy equity, performance-tied, but very large

The structure is conventional best-in-class: cash salary under 3% of total, the rest in PRSU (3-year TSR vs S and P 500 percentile), time-vesting RSU, and options. The 2025 Executive Incentive Plan paid out at 121% of target on financial milestones; the 2022-vintage PRSU paid at 96.84% of target because AMD's three-year TSR (53.55%) trailed the index (55.13%) by a hair — a real failure of the metric to pay 100%, which is reassuring. The board also retooled the 2025 PRSU design to a percentile rank within the index (a cleaner relative measure) and held the change-of-control terms at 2× cash with double trigger.

Where to push back: the absolute size. $55M to the CEO of a company whose three-year TSR essentially matched the index is hard to reconcile with "rigorous" pay-for-performance — that grant was earned on the future outcome, but the current package is benchmarked to peers (NVDA, Broadcom, Qualcomm) whose CEOs are themselves at extreme pay levels. Stockholders have signaled support (Say-on-Pay passed comfortably in 2025 per the proxy's engagement summary), so this is a tolerance issue, not a governance failure.

The alignment question — insider behavior matters more than ownership %

This is where the page gets uncomfortable.

181 Section 16 transactions across the past 18 months. Zero open-market purchases by anyone. Su accounts for ~75% of disposed value, with 9 gift transactions (charity / family trust) on top of 78 outright sales. Almost all sales appear to be 10b5-1 plan executions (predictable, pre-scheduled, defensible), but the pattern — a CEO whose entire compensation comes in stock, who sells everything that vests and never buys with the cash proceeds — is what it is. There is no insider conviction signal here at all-time-high stock prices.

Context that softens the picture:

• AMD's equity-heavy comp design makes selling rational — virtually 100% of after-tax compensation arrives as stock, so executives must sell to convert to cash to live.
• 10b5-1 plans are standard practice and remove timing-discretion liability.
• Su still owns ~4.3M shares (~$1.9B at recent prices); she has not sold to the point of breaching the 6× base-salary ownership requirement.
• AMD has an anti-hedging policy and prohibits pledging without committee approval (none has been granted).

Context that hardens the picture:

• The volume accelerated in 2026 (100 sales / $216M) vs. 2025 (13 / $29M) — a clear 7× step-up.
• Other NEOs (Papermaster, Grasby, Norrod) also sold materially with no buys.
• This is the one governance data point most likely to show up in a critical sell-side note.

The Board — independent, technically credible, but old-AMD heavy

The board has two genuinely strong technical voices — Su herself and Abhi Talwalkar (ex-LSI CEO, current Lam Research Chair) — both of whom can hold management's feet to the fire on roadmap and process choices. Marren's Temasek/TPG capital-allocation lens is the right voice for M and A discipline as AMD digests Xilinx, ZT Systems, and Pensando. Householder's audit chairmanship is unimpeachable (ex-PwC partner, ex-Sempra CFO/COO). McClure (joined Jan 2026, ex-Accenture CFO of 37-year tenure) added a second senior-finance voice; this was a thoughtful refresh, not a token addition.

Two structural critiques:

1. Tenure overhang. Denzel and Householder have been on the board since 2014, the same year Su took over. The board that hired her is the board that still oversees her. The 2026 retirement of Jon Olson (a 2008 director) starts to refresh this; one more long-tenured rotation would be welcome.
2. No standing AI ethics / national-security committee. AMD is now the #2 frontier-AI compute supplier with material China-export exposure (BIS controls, MI300/MI308 customer mix). The Innovation & Technology Committee covers the roadmap; the Audit committee covers regulatory compliance; nothing on the board sits between the two on AI-policy strategy. A capable board of this caliber should formalize that seam.

Ownership — widely-held, no control block

There is no founder block, no promoter, no controlling shareholder. The top-5 holders are entirely passive index funds. Net effect: AMD's accountability mechanism is the market plus index-fund stewardship plus an activist threat that has not (yet) materialized. That is a perfectly normal large-cap structure; it is not a structure that produces founder-level long-term thinking unless the CEO supplies it personally — which, to give Su credit, she has.

Insiders as a group hold under 0.5% of float. Su's own ~4.3M shares are worth roughly $1.9B at recent prices — a vast personal stake in absolute terms, a small one in voting terms. The 6× base-salary ownership guideline ($7.75M minimum) is met >200× over.

Governance hygiene — clean on the formalities

Formally, the housekeeping is excellent. The one late Form 4 (Guido) was disclosed and explained; the proxy notes no related-party transactions met the $120K SEC threshold in FY25; clawback and severance terms are at or above best-practice. There is nothing here that would draw an ISS adverse recommendation.

What would change my grade

Bottom line. AMD's people story is strong on capability and clean on formal governance, weaker on two things that matter most to outside shareholders during a euphoric tape: a CEO selling rather than buying, and the same person serving as Chair and CEO. Neither is a vote-against signal. Both are reasons to watch insider behavior more carefully than a standard semi governance template would suggest.