The strangest political confluence of 2026 is unfolding: U.S. President Trump and Democratic Senator Bernie Sanders-two extremes of the U.S. political spectrum-have reached a surprising consensus on AI policy. On June 6, Trump publicly stated that the U.S. government might directly invest in AI giants such as OpenAI, Anthropic, and xAI; just a few days ago, Sanders published a column in the New York Times calling for a one-time 50% equity tax on cutting-edge AI companies and the establishment of a federal sovereign wealth fund.
"Making America Great Again " vs " Democratic Socialism ": One Goal, Different Languages
The two politicians 'logical frameworks are very different, but point to the same policy endpoint:
| dimension | Trump (right-wing populist) | Sanders (Democratic Socialism) |
| Core arguments | AI companies use "our data" to train, and the public should share the benefits | AI company "steals" the creative works of millions of people, and the public should be compensated |
| policy tools | The government directly shares shares and becomes a "partner" | One-time 50% equity tax to establish a sovereign wealth fund |
| Target company | OpenAI, Anthropic, xAI ("American Champions") | OpenAI, Anthropic, xAI ("Predator") |
| use of funds | "Beautiful partnership", vague details | Public voting rights, dividend distribution, AI governance |
| political Foundation | MAGA Fundamentals (viewing AI companies as technology oligarchs) | Progressives (viewing AI companies as capital predators) |
Trump's original words were:"You make them partners in this revolution... it will be a wonderful thing. Sanders said: "The wealth of these companies is based on the creative work of millions of people-essentially stolen by some of the richest people in the world. "
Sanders bill: 'revolutionary' design of a 50% equity tax
SandersThe American AI Sovereign Wealth Fund Act was officially proposed in early June. The core provisions include:
One-time 50% tax: for OpenAI, Anthropic, xAI (notably excludes Google or Meta)
Payment method: Pay in stock rather than cash
Fund purpose: Federal sovereign wealth fund, which gives the public voting rights on the board of directors of AI companies and future dividend distribution rights
Legal basis: AI companies use publicly created content training models without permission, and the public should participate in revenue distribution
Altman's "negotiating stance": Not supporting 50%, but supporting "general concepts"
OpenAI CEO Sam Ultraman's response to Sanders was extremely strategic-he "cannot support the 50% threshold" but "is willing to work with Sanders to advocate the general concept of public equity participation in AI companies." This is not a rejection, but a negotiating gesture: moving from a 50%"impossible" starting point to a lower "acceptable" end point.
According toNotus, the Wall Street Journal and the Financial Times confirmed that Ultraman had directly promoted the concept of a government stake to President Trump as early as early as 2025, and discussions have continued recently.
The "political complexity" of the IPO market
This political confluence has a strong impact on the upcomingAI companies pose a direct threat:
Anthropic: Has filed confidential S-1, targets to go public in October, valued at US$965 billion
OpenAI: Working with Goldman Sachs and Morgan Stanley to prepare IPO documents, target to go public in September
SpaceX: Launch on June 12, SPCX ticker
"Headline news about government equity discussions in the IPO roadshow is not ideal for institutional investor sentiment. "Analysis pointed out that if the Sanders Act or Trump executive order is advanced during the IPO window, valuation expectations may be significantly lowered.
The "policy convergence" between Trump and Sanders is the most intriguing political phenomenon in 2026. It reveals the core paradox facing the AI industry: when technology creates trillions of wealth, the distribution justice of wealth becomes an unavoidable political issue. Sanders uses the "steal" framework and Trump uses the "partnership" framework, but both point to the same conclusion-AI companies cannot fully privatize their profits.
YesFor OpenAI and Anthropic, this is the most dangerous "double attack": Sanders '50% equity tax is a "death penalty" and Trump's "government shareholding" is a "life sentence." Ultraman's "negotiating stance" suggests that OpenAI is trying to trade "active cooperation" in exchange for "passive acceptance" exemptions-but the political logic is that once a precedent for government equity ownership is established, the 50% threshold is only a matter of time. The IPO feast for the AI industry may be turning into a preview of "political spoils."