Good morning. In the Unhedged podcast this week, we asked listeners what we should call the (only possibly real) flight of capital out of bitcoin and into AI stocks and SpaceX. Katie’s suggestion of The BS Trade, but the current leader is the listener on Spotify who came up with “The Great Brotation”. This is your chance to rival Rob’s triumph with the Taco Trade. Heaven knows he’s been dining out on that one, so tell us yours: [email protected].

<img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/t/2/47857/[email protected]/5817747077646525/0/0'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/47857/[email protected]/5817747077646525?pid=1'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/47857/[email protected]/5817747077646525?pid=2'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/47857/[email protected]/5817747077646525?pid=3'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/47857/[email protected]/5817747077646525?pid=4'>

AI groups are not like other companies. By any measure — size of funding rounds, pre-IPO valuations, revenue growth — OpenAI, Anthropic and the like are exceeding historical norms. It’s also unusual for a chief executive to point out that their product might just take all the jobs and then end the world. And finally, it’s strange for some of the fastest-growing companies in the world to suggest giving stakes in their businesses to a sovereign wealth fund. And yet here we are.

The idea of an AI wealth fund has gained some momentum in the past few weeks, after US senator Bernie Sanders proposed a 50 per cent tax on the profits of the AI labs, payable in shares. Donald Trump has seemingly endorsed the idea and been in talks with Sam Altman of OpenAI about a voluntary gift of the ChatGPT maker’s shares to the US public.

If you’re reading this newsletter there’s a decent chance you think governments taking stakes in private companies is a bad idea. Because it usually is. But maybe not this time? Some arguments in favour:

It’s industrial strategy and we all do that now. It’s 2026, and governments taking stakes in strategic sectors is very much back in play. If you’re concerned about liberal democracies winning the AI race, then maybe there’s a case for including AI companies in industrial strategy.

It’s PR/damage control. The AI bros have lost a lot of goodwill telling us their tech will make us all redundant. No surprise then that the public is starting to push back. Left unchecked, the backlash could slow the tech’s adoption and limit its longer-term economic benefits. It is in the interest of Altman et al to appease the masses, who own torches and pitchforks. What better way than equity handouts, especially if they head off extortionary taxation.

It’s the seeds of a compensation programme for the AI losers. Positive supply shocks always create winners and losers. In the heyday of globalisation, economists told us there was no point resisting the forces of trade integration and deindustrialisation. The best we could do was compensate the losers in the form of unemployment insurance, job retraining and relocation. But the compensation never came. If it’s true that AI is going to put us all out of work, then an AI wealth fund is one way to fund the great cost that is going to impose on the state and society.

It aligns public and private incentives. OpenAI and Anthropic have origins in rather unconventional non-profit or public benefit corporate governance structures. This says something about how they think — or once thought — their technology could be harnessed for more than just private profit-seeking. Giving the broader public a stake in their companies would be a literal expression of that public purpose (Could this bind public finances to the AI valuations and disincentivise AI regulation? Absolutely).

It’s socialising the profits. Yes, and so what? The US economy has already become reliant on AI investment as the engine of growth. It may be too early to call the AI companies too big to fail, but their systemic importance is rising. It’s only a matter of time before a failure in the wider AI ecosystem forces the question of government backstops and bailouts. So why not reap some of the returns while the going is good?

There’s a lot we don’t know about the proposals, and there’s a lot that could go wrong. For example, would the shares be bought or donated? Would the wealth fund pay out cheques, like the Alaska Permanent Fund Dividend, or be used to reduce the federal deficit? There are good and bad ways to do this. But if AI is anywhere near as transformative as its creators say, it will require new ways of thinking about corporate governance, economic ownership and redistribution.

(MacFadden)

<img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/t/2/40827/[email protected]/8175842081734281/0/0'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/40827/[email protected]/8175842081734281?pid=1'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/40827/[email protected]/8175842081734281?pid=2'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/40827/[email protected]/8175842081734281?pid=3'><img width='1' height='1' style='display:none;border-style:none;' alt=' src='https://images.passendo.com/extt/2/40827/[email protected]/8175842081734281?pid=4'>

CPI inflation is rising fast and the market doesn’t much care. Yesterday the two-year Treasury, a sensitive gauge of fears about inflation and tightening monetary policy, took a look at the headline annual figure of 4.2 per cent for May and rose a near-meaningless 2 basis points. The reading, the worst in three years, was mostly a result of fuel prices, and the market has long since concluded that energy inflation will prove transitory (Disagree? Quit complaining and sell stocks, which will get clobbered if you’re right).

Stripping out energy and food prices, core inflation is not getting worse. The month-over-month rate slowed, and the three-month average looks stable (the six-month average, the one we really like, is still distorted by the missing data from the end of 2025).

Core goods prices fell between April and May, but that seems unsustainable, for reasons neatly laid out by Sam Tombs of Pantheon Macro: “The pass-through of tariffs to consumers likely now is complete, but the manufacturing surveys indicate that companies soon will start to pass on the hit from the energy shock.”

The problem is services inflation, which is trending solidly above 3 per cent and showing no sustained improvement whatsoever:

You can split the econo-pundit world into two camps. One camp thinks the services inflation will subside because real wage growth is weakening. In other words, the inflation problem will solve itself. The other thinks pandemic-era price rises were never truly beaten, looks at the recent unexpected surge in job creation and concludes the economy is reaccelerating, and inflation will follow. We will write more about this debate soon. Let us know which side you are on.

(Kim and Armstrong)

The new joule order.