When you own the whole market but can't out-vote the founder, has AI stewardship hit its ceiling?

UAO Research · May 29, 2026 · Stewardship, Governance & Risk · Feeds: The Universal Owner Risk Radar

A universal owner holds the AI build-out whether it wants to or not. It is in the index, in the credit benchmark, in the infrastructure pipeline. The one thing the owner is supposed to be able to do about a risk it cannot diversify away from is govern it — vote, engage, escalate. Meta's annual meeting on May 27 put a hard number on how far that lever reaches: the strongest AI-governance proposal of the day drew about 27% support and lost, every other proposal lost by more, and the world's largest single fund could not change the arithmetic. The question this raises is not whether stewardship is worth doing. It is whether, at the companies that matter most for systemic AI risk, the vote has a structural ceiling — and what is left for an owner once it hits that ceiling.

What the evidence says

Start with the academic backbone. In Index Funds and the Future of Corporate Governance, Lucian Bebchuk and Scott Hirst argue that the largest diversified owners face agency-cost incentives to under-invest in stewardship and to defer to corporate managers — because the cost of serious engagement is borne by the manager while the benefit is spread across the whole market and across competitors Index Funds and the Future of Corporate Governance: Theory, Evidence, and Policy, Bebchuk & Hirst, Columbia Law Review / NBER w26543, 2019. Their evidence is about index managers specifically, but the logic travels to any owner so diversified that the gains from improving one company are tiny relative to its book: the rational level of effort is lower than the systemic stakes would justify. That is the structural reason engagement so often resolves into a vote rather than a campaign.

Now layer on the voting math. At founder-controlled platforms, dual-class structures put the outcome out of reach before a single outside ballot is counted. Mark Zuckerberg controls roughly 61% of Meta's votes through ten-vote Class B shares; all twelve directors were re-elected with 82%+ support and all ten shareholder proposals failed 36Kr, May 28, 2026. An owner can vote its full position and still lose every contested item. So the ceiling is not only the Bebchuk–Hirst incentive problem; at the AI mega-caps it is also a hard franchise problem.

The season-wide data shows the pattern, not just the instance. Boardroom Alpha's 2026 scorecard counts eight AI-focused shareholder proposals across four US issuers; IBM's AI-bias proposal failed at 2.4% in April, and every named AI proposal in 2025 — Apple, Amazon, Meta, Alphabet, Lyft — failed in the single digits to low teens Boardroom Alpha 2026 Proxy Season Scorecard, 2026. Meta's ~27% is the high-water mark precisely because support is rising — but it is rising from a floor so low that, even extrapolated, a majority on AI governance is years out, and unreachable at the controlled companies.

Where it is contested

The pessimistic reading is too clean. First, the vote is a lagging and partial measure of stewardship; the more consequential work — private engagement, board access, the credible threat of a public rationale — does not show up in a proposal tally. NBIM's choice to publish its reasoning and to withhold its vote on a specific director is a different instrument from a yes/no on a resolution, and it costs the board something even when the resolution fails The Next Web, May 2026.

Second, the Bebchuk–Hirst framework is contested by managers who point to engagement programmes, voting-policy disclosure and the rising salience of AI resolutions as evidence that stewardship is deepening, not hollow. The honest position is that the academic literature establishes an incentive to under-invest, not proof that every large owner does — and a few owners with explicit mandates (NBIM, some Canadian and Dutch plans) plainly invest more than the median.

Third, "rising support, still losing" is genuinely ambiguous. It can mean stewardship is gaining traction and the majority will come; or it can mean the ceiling is real and the trend asymptotes below 50% at the companies that matter. The data cannot yet distinguish these, and a universal owner has to act before it can.

From the allocator's seat

The practical conclusion is not to abandon the ballot but to stop treating it as the whole of stewardship — especially at founder-controlled names where it cannot win. Three moves follow. Reallocate stewardship effort by where the vote can bite: spend scarce engagement capacity at companies with one-share-one-vote structures where a coalition can actually carry a proposal, and shift to public-rationale and director-accountability tools at the controlled platforms. Price the governance gap into the position: if a holding's systemic risk cannot be governed through the ballot, that is a reason to weight the disclosure, the manager engagement and the credit-side exposure more heavily — not a reason to assume the risk is being managed for you. And be explicit with beneficiaries and managers about which lever you are pulling and why, because the credibility of escalation depends on its being legible.

The deeper point connects to this week's brief: the same AI names an owner cannot govern through equity votes are becoming its largest corporate-credit exposures as they borrow to build, with AI-linked issuers heading toward roughly 14% of the US investment-grade index in 2026. The exposure is compounding while the governance lever is capped. That asymmetry — long the risk, short the votes — is the universal owner's defining stewardship problem of this cycle, and the ballot alone will not close it.

What to watch next

Alphabet and Walmart AGMs (June 2026): whether AI-governance proposals there beat or fall short of Meta's ~27% — the clearest near-term read on the trajectory.
Large owners' published voting rationales (rolling): whether NBIM-style public reasoning becomes the standard escalation tool at controlled platforms.
Norway's ethics-framework review (due October 15, 2026): how the world's most-watched responsible-investment regime resets the balance between exclusion, engagement and disclosure.
2026 IG index composition prints: whether AI-linked issuance pushes the sector past the ~14% J.P. Morgan estimate, deepening the credit-side concentration.

Sources

UAO Research. AI-assisted monitoring and drafting; reviewed and edited by the UAO editorial desk before publication. Not investment advice.

Continue the briefing. Read the daily brief · watch the daily video briefing · listen to The Universal Owner · view the chart of the day.

Produced and edited by the UAO editorial desk. Not investment advice.

What the evidence says

Where it is contested

From the allocator's seat

What to watch next

The morning briefing for the people who allocate long-horizon capital.