by Editorial Team, AdvisorAnalyst.com
A strange thing has happened in American politics: Bernie Sanders and Donald Trump agree on something. In an open letter published this June1, AllianceBernstein's Inigo Fraser Jenkins observes that "Sanders suggests the transfer of a 50% equity stake into an American sovereign wealth fund; Trump's proposal for the US government taking stakes in AI companies seems more limited." The agreement, he notes dryly, mirrors thinking that has emerged from inside the industry itself, including from Sam Altman, while the Financial Times has floated a different remedy entirely — taxation, "specifically suggesting an increase in the capital gains tax."
Jenkins frames this convergence as no accident. He points to "the parallel observations of enormous wealth creation, a looming fear about the impact of AI on jobs and a growing realization that there are also far greater risks to society and mankind from unbounded development of AI" as the backdrop making the ownership debate unavoidable.
The Profit Share Problem
The analytical core of the letter rests on a structural observation about who actually benefits from AI-driven productivity. Fraser Jenkins writes that any productivity gain "can either arise from enhancing the productivity of a unit of labor or from automating and replacing that unit," and that forecasters tend to dodge the question of which dominates. His read of the current arrangement is unambiguous: with corporations controlling what AI gets built and deployed, the system "seems set to increase their profit share further through automation." He ties this directly to AB's house view, noting that America's unprecedented rise in corporate profit share of GDP is "one element... of the narrative of US exceptionalism that leads us to recommend that investors maintain a strategic overweight in US equities compared with the rest of the world." There is, he warns, some ceiling to this dynamic — a point past which profit share becomes "too high, potentially prompting an abrupt, revolutionary-like rejection of corporate power and AI," though he concedes "we do not know where that level lies."
Sovereign Stakes Don't Fix the Real Problem
Here the letter pivots, and it's the sharpest turn in the piece: even if sovereign stakes happened, they would leave the deeper issue untouched. Fraser Jenkins argues plainly that such stakes "leave unresolved other large questions raised by the power transfer implicit in AI," and specifically would not address "the shift in the relative power of governments vs. corporations." He traces this imbalance to the declining effective tax rate of US firms and to governance shifts like "the voting structure of SpaceX," concluding that corporate power "seems unprecedented." Government power, meanwhile, "rests on shaky ground" — G7 government debt averages 120% of GDP, and he cites Niall Ferguson's observation that 2025 marked the first year the US cost of debt servicing exceeded its defense budget, a threshold at which "a great power faces constraints on that power." There's an added irony Fraser Jenkins flags: AB's own research shows corporate profitability tends to benefit from fiscal deficits, meaning "corporations have gained power at the expense of both governments and labor" — which is why he suggests taxing AI firms, rather than taking equity in them, "might restore the power of corporations vs. governments to a more normal level historically."
The Regulatory Capture Question
If taxation is one lever, antitrust is the other — and Fraser Jenkins doesn't shy from naming why it hasn't been pulled. He raises "the degree of regulatory capture that has already taken place" as a likely explanation, alongside the argument that the largest AI firms are now bound up with national security in "the new cold war between the US and China." He poses the uncomfortable hypothetical directly: if AI companies held significant public stakes, "would that potentially preclude effective regulation?" — since a regulator might "pull its punches if there was a direct hit to the public purse."
The Commons Question Nobody Will Touch
The letter's most provocative move is invoking Acemoglu's framework that technology alone never determines who benefits — politics does. Given AI's reliance on "the pillaging of data for training" and its environmental and epistemological costs, Fraser Jenkins concludes there's "a morally defendable case for a commons-like ownership, though we realize that such a suggestion would never fly."
Takeaways for Advisors and Investors
Fraser Jenkins is explicit that he won't resolve the normative question — that's for politicians. But the investment conclusion stands firm: a strategic overweight to US equities remains the best route to long-run real returns, even as he warns that complacency about volatility, suppressed by persistent inflows, "will not always appear so benign." His closing line is a flag advisors should sit with: letting AI compound wealth inequality further "is not a path to any kind of social equilibrium and is profoundly dangerous."
Footnote:
1 Fraser Jenkins, Inigo. "Who Should Own AI? An Open Letter." AllianceBernstein, June 2026.
