A Republican President and a Socialist Senator Just Agreed on AI. Here’s Why That Should Get Your Attention

This image was generated with help from an AI program

All week I followed the money up the AI stack, the raises, the compute deals, the runaway bills. This is the top of the stack, the layer above all of it: the state itself, now reaching for a piece. And the most revealing part is who’s reaching alongside it.

For nine days I’ve been writing about the AI boom by following a single thread: power, and who holds it. Who builds the agents, who governs them, who funds them, who controls the channel, who owns the compute, who pays the bill. Each story climbed one rung higher up the stack.

This is the top rung. On Friday, President Trump said he’s been in conversations with AI companies about arrangements where, in his framing, the American people could benefit from AI’s success, where the public would, in effect, become a partner in these companies. Reporting from CNBC indicates the administration has specifically discussed taking an equity stake in OpenAI.

After a week of private capital and corporate maneuvering, the most powerful actor of all just pulled up a chair. And here’s the thing that makes this more than a partisan story: look at who’s pulling up a chair on the other side of the table.

The horseshoe nobody expected

Set the politics aside for one second and just look at the positions.

On the right, a Republican president is exploring government equity in private AI firms, consistent with his administration’s earlier move to take a roughly 10% stake in the chipmaker Intel. The mechanism reportedly under discussion could seed something OpenAI itself has floated: a “Public Wealth Fund,” whose proceeds, in the company’s telling, could flow directly to citizens so that ordinary people share in AI’s upside regardless of how much capital they started with.

This story was written with assistance from an AI program.

On the left, in the very same week, Senator Bernie Sanders proposed a one-time 50% tax that the major AI labs, OpenAI, Anthropic, xAI, would pay not in cash but in stock, handed to the public. His argument: give the public a genuine role in steering where this technology goes, and ensure the trillions AI might generate improve life for everyone, not just shareholders.

Stop and absorb that. A president associated with deregulation and a senator who calls himself a democratic socialist have arrived, from opposite ends of the political universe, at strikingly similar destinations: the public should own a piece of artificial intelligence. The mechanisms differ, negotiated equity stake versus mandatory stock tax, but the underlying instinct is the same.

When the two furthest ends of the spectrum independently reach for the same idea, that idea has stopped being ideological and started being structural. Something about AI is forcing this convergence.

That something, I’d argue, is recognition. Both ends of the spectrum have looked at the scale of what’s coming, the compute deals at a billion dollars a month, the trillions in projected value, the sheer centrality of this technology to the economy’s future, and concluded that it may simply be too important to leave entirely in private hands. They disagree on almost everything else. They agree on this.

The case for, taken seriously

I want to give the strongest version of the pro side, because it’s more compelling than reflexive free-market objections allow.

The deepest worry about AI has always been distributional: that it concentrates wealth and power in the hands of whoever owns the models, while displacing the labor of everyone else. If AI really does generate trillions while automating away the jobs that used to distribute that prosperity, then a mechanism for broad public participation in the upside isn’t socialism or cronyism, it’s a pressure-release valve for a society that might otherwise fracture. There are working precedents: Norway’s sovereign wealth fund, Alaska’s oil dividend. Citizens sharing in a national windfall is not a radical idea; it’s a tested one. Seen this way, a public stake is how you keep the AI transition from becoming a wealth transfer to a few thousand people.

That framing is exactly why it resonates across the aisle. The fear of being left behind by AI is bipartisan, so the appeal of “the public gets a share” is bipartisan too.

The case against, taken just as seriously

Now the strongest version of the other side, and notice, again, that it also crosses party lines.

David Sacks, an investor who until recently served as the administration’s AI and crypto czar, said he understands why Sanders’ proposal resonates with many on the right as well as the left, but warned that it would accelerate a dangerous fusion of corporate and government power that he sees the country already sliding toward. That warning deserves real weight, regardless of where you sit politically. When the government owns the company, who regulates whom? An AI lab that is part-owned by the state is a lab the state has every incentive to protect, favor, and shield from competition. National champions get entrenched. Smaller builders, the rest of us, find ourselves competing not against a company but against a company with a sovereign backer. And a technology that shapes speech, information, and decision-making becomes, to some degree, an arm of whoever holds office.

There’s a sharper version still, voiced by skeptics this week: that this is quietly laying the groundwork for a government bailout of OpenAI. Recall the threads from earlier this week, the staggering capital being raised, the billion-a-month compute commitments, and the uncomfortable fact that most companies still can’t measure whether their enormous AI spend actually pays off. If the economics of the frontier labs turn out to be shakier than the valuations imply, a government stake starts to look less like the public sharing in the upside and more like the public being positioned to absorb the downside. “Too big to fail” is a phrase we’ve heard before, and it didn’t end with citizens getting rich.

Why this matters if you build on AI

Here’s the operator’s reading, and it’s significant.

If the state takes equity in the foundational AI labs, those labs stop being ordinary vendors and become something closer to strategic national infrastructure, quasi-utilities with a government on the cap table. For anyone building on top of them, that cuts both ways. On one hand, a government-backed supplier is a stable supplier; the existential “what if my AI provider implodes” risk drops. On the other, you are now building on a platform whose priorities answer partly to politics, not just to customers, subject to national-security carve-outs, favoritism toward incumbents, export controls, and the shifting winds of whoever’s in power. The neutral, commercial substrate you thought you were building on acquires a flag.

It also reframes the bubble question I keep circling. A government stake can read as the ultimate validation, the state betting on AI’s permanence. Or it can read as the first visible crack: the moment the risk got too large for private balance sheets alone to carry. This week gave us both interpretations, and honestly, it’s too early to know which is right. I’d hold both in mind and avoid betting my business on either certainty.

Where I land and where I don’t

I’m not going to tell you whether a government stake in OpenAI is good or bad, and not because I’m dodging. It’s because the honest answer depends on values this technology can’t settle for us: how much you trust the state versus the market, how much you fear concentrated private power versus concentrated public power, how you weigh broad participation against the risk of capture. Reasonable, informed people land in very different places, and the fact that this week put Trump, Sanders, and a former administration insider in such a tangled relationship to one another should make anyone wary of a tidy partisan take. (And the usual caveat: I’m an operator, not a financial or policy advisor, this is analysis, not advice.)

What I’m confident about is the significance. This is the week the AI story graduated from a business story into a sovereignty story. The question is no longer just what these systems can do, who funds them, or what they cost. It’s who owns them, and whether “the public” turns out to mean genuine shared prosperity or merely a new entanglement of corporate and state power wearing the language of the people.

The week, all the way up

So here’s the full climb. NVIDIA showed what agents can do. Microsoft showed how we’ll govern them. Meta showed how they break. Alphabet showed who funds them. Dreambeans showed how we’re made comfortable with them. Apple showed who controls the door. Google and SpaceX showed the compute beneath them. The token reckoning showed the bill. And now Washington has shown the ceiling above all of it: the state, deciding whether AI is a private industry or a public one, with the far left and the far right, remarkably, both insisting the public deserves a seat.

Capability was where this week started. Ownership is where it ended. And the most important conversations about AI from here on won’t happen in keynotes or on cap tables. They’ll happen in capitals.

I’d watch those rooms closely. Whatever gets decided in them will reshape the ground the rest of us are building on.

Genuine question, and I want the disagreement: should the public hold a direct stake in the AI labs, and if so, through a wealth fund, a stock tax, or not at all? And do you read a government stake as a stabilizer or as the first sign of a backstop nobody’s admitting to? I’d like to hear every side of this one.