Another Merry Week

OpenAI and Anthropic are both readying mighty IPOs of their own, hoping to scoop up billions so they may go on pouring money into an almost miraculous technology of uncertain value.

UFC Octagon Ring in front of the White House. Source: Getty Images

Monday, June 15th, 2026

Bill Bonner, from Youghal, Ireland

---

Another merry week has clattered by. And what did we learn from it? Anything new?

From Monday through Tuesday and on into Wednesday, we watched the assault on Iran gather its head of steam — and the oil market was flung about like a rag doll in the jaws of a puppy.

The President vowed to hit the Iranians “very hard.” Why we were hitting them at all remained, as ever, a mystery wrapped in fog. But no matter: by Thursday the price of oil was sliding again, and we were hitting them no longer. Fox:

Iran latest: Trump cancels tonight’s airstrikes, says deal has been approved

Really? Was the war truly over, we wondered? Had Trump TACO’d yet again?

By Friday the press declared the deal genuine. Pakistan said so. The trouble was that it bore no resemblance to a meeting of the minds — either because the parties never met, or because no minds were on hand to begin with. Antiwar.com:

Tehran and Washington have publicly offered different details for the agreement. According to Iranian media, the deal does not include any restrictions on Iran’s nuclear program, Tehran retains control over the Strait of Hormuz, a ceasefire that includes Lebanon, and the unfreezing of Iranian assets.

After the report, Trump claimed it was false. “The terms that Iran leaked out to the Fake News have NOTHING to do with the terms that were agreed to, in writing,” the President wrote on Truth Social Friday. “What they said, including their weak and pathetic statement on having a deal, bears no relation to the truth. Very dishonorable people to deal with.”

It is now Monday morning, and the terms of the deal — if a deal there be — remain a secret, at least from us. But connecting the dots — the “peace agreement” will probably turn out to be the very twin of the war that bred it: nobody knew what Iran did to make us want to bomb it, and nobody knows what it does to make us stop.

But the grandest spectacle of the week arrived on Friday. The Financial Times:

SpaceX soars 30% in historic debut

To the moon! The company was valued at better than $2 trillion, with Mr. Musk’s personal hoard vaulting past $1 trillion. Speculators craved a slice of it so ravenously that they cheerfully paid the loftiest prices ever.

And the other AI biggies are queuing up. OpenAI and Anthropic are both readying mighty IPOs of their own, hoping to scoop up billions so they may go on pouring money into an almost miraculous technology of uncertain value.

Should it all unfold as planned, these offerings will heap some $4 trillion onto the value of America’s publicly traded companies.

“It’s hard to believe,” said Musk of the IPO. “If people had told me this was going to happen, I was like, man, you must be smoking some really good crack, because I think this company is going to fail.”

Musk may yet prove the truest prophet in the room. There is no plausible road to profit for SpaceX — at any rate, no profit that could justify this morning’s price. Our own guess is that the price will adjust — downward — long before any profitable business model is unearthed.

Another grand event fell on Sunday, when POTUS turned eighty and celebrated with a pageant worthy of Caligula. That Roman emperor adored mass entertainments — the more lurid and low-brow the better; he also waged pointless wars, raised extravagant monuments, persecuted his enemies, and drained the treasury dry. In one famous turn, he proposed to make his horse a consul — whether out of madness or merely mischief, the historians have never settled.

Meanwhile, the New York Times published a piece laying bare how desperately all the President’s men — and women — labored to keep the Epstein files from the light.

“In their public statements, Trump’s advisers were full of bravado, dismissing the crisis. In reality, it was consuming the highest ranks of the administration as no issue had for the president’s team since the Russia investigation in his first term. His aides were determined to keep their rising sense of panic out of public view,” the New York Times reporters wrote.

There is “chaos in the White House,” cries the press — not so much because the President’s men are striving to mislead the public and obstruct the course of justice, but because some traitor in the inner circle is spilling the beans. MSN:

Trump’s fury over Epstein book fuels massive leak hunt

Perhaps the whole Trump menagerie will be waterboarded, one by one, until the leaker confesses.

Our own guess is that the Deep Throat belongs to JD Vance. With one eye already cocked toward the 2028 primaries, he fears that Trump is driving off the independent voters he will need to seize the White House. And so, he feeds the press a version of events in which he alone plays the intelligent voice in a conspiracy of dunces. Whispering to the Times about the goings-on in the situation room is his way of putting a little daylight between himself and POTUS.

By and large, it was a good week. When it ended...and we were still alive. The liquor stores were open. And the show, as jolly and alarming as ever, goes on.

Regards,

Bill Bonner

---

Research Note, by Dan Denning

While the SpaceX IPO went off the launchpad without a hitch, last week raised serious doubts about whether Anthropic (Claude) and OpenAI (ChatGPT) will be able to sell shares to the public and realize the trillion dollar valuations everyone expects. What happened?

On Tuesday of last week, Anthropic launched Fable, the public version of its Mythos software I wrote about in early April. Anthropic has been famously and publicly concerned about ‘ethical’ and ‘safe’ AI. Fable is/was the most powerful version of its model yet, designed to prevent cyber crime, but with ‘guardrails’ to prevent ‘bad actors’ from using it for nefarious purposes (sussing out software vulnerabilities in systems that could be exploited).

It was all going swimmingly until one of Anthropic’s vendors, Amazon Web Services (AWS) called the federal government and said that it was possible to ‘jailbreak’ Fable (force it to do bad things) because AWS programmers had done just that. The US Commerce Department promptly reacted on Friday afternoon by slapping export controls on the Fable model, banning it for all non-US users. This forced Anthropic to not only disable access to the model not only for all its own foreign employees but for all of its current users as well.

There may be more that meets the eye going on here. Anthropic and the Department of War have been engaged in a month’s long back-and-forth over whether its tools come with ‘strings attached’ (restrictions on how the US government can use them). But the kerfuffle couldn’t come at a worst time for Anthropic’s plans to raise cash by going public.

Fable, we have a problem. Will AI pop its own bubble by warning about public safety? Will a ‘two-tier’ AI world where the public gets a ‘safe’ version and the government gets the powerful version generate enough revenues to justify valuations of AI companies? Or did the bubble find its pin last week?