Every year, the late-May cluster of Alphabet, Amazon and Meta annual meetings gives stewardship teams their clearest read on where investor pressure is heading. The 2026 slate delivered a decisive one: fewer social resolutions, the first-ever double-digit run of AI proposals, and recapitalization votes that quietly exposed the limits of "one share, one vote."
Key takeaways
- More than 175 shareholder proposals have gone to a vote at Alphabet, Amazon and Meta over the past five years — but 2026 saw the total fall, led by a continued decline in social resolutions.
- AI-related proposals reached double digits for the first time, with the strongest support where AI intersected accountability, transparency and long-term risk.
- Recapitalization proposals won a majority of independent shareholders at both Alphabet and Meta — yet failed overall, because founder super-voting shares outvoted them.
- With the SEC no longer reviewing no-action requests, exclusions shaped ballots more than votes did — none of Amazon's seven challenged proposals reached the floor.
The mix is shifting under stewardship teams' feet
For most of the decade, the Big Tech proxy ballot was dominated by social and environmental resolutions — emissions targets, water use, content moderation, civil-rights audits, workforce policy. That center of gravity is moving. In line with a multi-year, market-wide trend, social proposals fell again in 2026, while a new category matured fast: governance of artificial intelligence.
A single 2026 ballot at these companies could still span an extraordinary range — H-1B visa policy, human-rights concerns, hate-speech and child-safety reporting, climate commitments, vote-result disclosure and share-class structure all appearing on one agenda. But the resolutions gaining momentum increasingly cluster around two poles: how these companies are governed, and how they are building and overseeing AI.
Governance still tops the ballot — and dual-class masks the real vote
The most-supported resolutions at Alphabet and Meta were, once again, governance proposals asking each company to unwind its dual-class share structure and move to one share, one vote. On the official tally they "failed" — 31.2% support at Alphabet and 26.5% at Meta. But that headline number understates investor sentiment, because the very structure being challenged is what defeats the challenge.
At both companies, founders control a majority of the vote — roughly 52.9% at Alphabet and 60.8% at Meta — while holding under 20% of the economic value, thanks to super-voting shares carrying ten votes each. Strip those out and a majority of independent shareholders backed one share, one vote at both firms. The same independent majority at Meta supported moving say-on-pay to an annual cadence, and a proposal to break out vote results by share class drew near-majority support.
| 2026 recapitalization vote | Reported support | Founder voting control |
|---|---|---|
| Alphabet | 31.2% | ~52.9% |
| Meta Platforms | 26.5% | ~60.8% |
Reported support is depressed by the dual-class structure itself; among independent shares, both proposals carried a majority. Figures per company proxy disclosures.
AI moves from footnote to headline
For the first time, AI-related resolutions at these three companies reached double digits. They are no longer a single-issue ask. Investors are now probing the full stack of AI risk: board-level oversight of responsible development, the sourcing and use of training data, the reliability and rights implications of deployed systems, and — increasingly — the eye-watering capital and energy required to build it all.
At Alphabet, one proposal asked the audit committee to take formal oversight of responsible AI development and the human-rights risks it creates for users and other stakeholders. A parallel thread ran through all three companies on energy: filers including As You Sow, Mercy Investment Services, Presbyterian and Trillium Asset Management asked each firm to explain how surging data-center electricity demand squares with its stated climate commitments. None of these proposals passed, and support for the energy-and-climate asks actually slipped versus the prior year — but the pattern is unmistakable. Support concentrated where AI met durable financial concerns: accountability, transparency and long-term risk management, rather than reputational framing alone.
That is the signal worth recording. With Big Tech projected to spend on the order of $650 billion on AI infrastructure in 2026, "how is this capex governed?" is becoming a mainstream fiduciary question — not an activist one.
The quiet story: exclusions, not votes, shaped the ballot
A procedural shift did as much to shape 2026 as any vote. With the SEC declining to review no-action requests this season, some feared companies would exclude proposals aggressively. In practice the picture was mixed: Meta sought to exclude nothing, Alphabet just one resolution — but Amazon filed seven exclusion notices, and none of those seven reached a vote. A meaningful share of the year-on-year decline in proposal volume came not from proponents standing down, but from resolutions never making the floor.
"Failed" is not the same as "irrelevant"
The 2026 season is a reminder that vote outcomes are a lagging indicator. The leading signal lives in the detail — and capturing it is an operational problem before it is an analytical one.
- Track independent-adjusted support, not just headline results. A proposal that carries a majority of independent shares is a forward indicator of where engagement is heading, whatever the founder-controlled tally says.
- Treat AI governance as a fiduciary theme with a paper trail. As AI oversight, training-data and data-center energy intersect climate and SFDR-style disclosure, your voting rationale needs to be recorded and defensible — not reconstructed from memory at audit time.
- Expect a widening, faster-moving topic set. When a single ballot spans visa policy, child safety, capital structure and AI, spreadsheets stop scaling. Structured, framework-mapped engagement records are what keep a team defensible.
The through-line of 2026 is that stewardship is broadening and accelerating at the same time. The teams that come out ahead won't be the ones with the strongest opinions — they'll be the ones who can show, engagement by engagement and vote by vote, exactly what they did and why.
Sources
- Bloomberg, "Shareholders Press Big Tech Over AI Energy Use and Climate Goals" (May 2026).
- Insurance Journal / wire coverage of 2026 AI-energy climate proposals (May 2026).
- Alphabet Inc. and Meta Platforms Inc. 2026 DEF 14A proxy statements and 8-K vote results (SEC EDGAR).
- Yahoo Finance, "Big Tech set to spend $650 billion in 2026 as AI investments soar."
Analysis and commentary by engage insights. Figures are drawn from company disclosures and third-party reporting and are provided for information only, not as voting or investment advice.
Amine Gharby
Founder & CEO, engage insights. Former Glass Lewis Director. Writes about stewardship, regulation, and building better tools for asset managers.
Follow on LinkedIn →