I don’t think the bubble — if that’s what it is — bursts before OpenAI and Anthropic go public. Or until the Knicks win the NBA championship.

BofA’s Michael Hartnett mentioned the latter in this week’s “Flow Show.” “All that, plus the Knicks as champs? I mean, you couldn’t script a better top,” he joked, describing the zeitgeist.

It’s tempting to suggest SpaceX’s record debut — which minted the world’s first trillionaire — is a canary. A bad omen characteristic of the decadent disregard associated with the final days of market manias and speculative frenzies.

But I think there’s probably a ways to go just yet. As Hartnett put it, “we’re getting there, but for now asset allocations are frozen bullish, positioned for late-cycle greed and not at all tempted by 5% yields at the long-end of the Treasury curve.”

I concur with that assessment. Although hardly a scientific observation, I do think there’s something to the idea that the AI fever dream won’t break until the public gets a chance to own the two names synonymous with the delirium. And SpaceX’s hurried transformation from rocket company to AI data center REIT notwithstanding, those two names are OpenAI and Anthropic.

Although “speculative mania is a poor timing indicator,” it’s nevertheless “one of the dynamics that has characterized the peaks of high-valuation, high-concentration bull markets in the past,” Goldman’s Ben Snider remarked.

“Other dynamics that have marked the ends of similar bull markets include disappointing growth, extremely elevated equity issuance and tightening Fed policy,” Snider continued, adding that “none of these conditions describe the environment today, but each appears closer than it did just a few months ago.”

For whatever it’s worth, the table above gives you some context for several metrics Goldman watches for signs of exuberance — “irrational” or otherwise.

To Snider’s point about issuance, equity supply’s a concern, not due to the number of IPOs, but rather to the size of them. When considered with Google’s massive follow-on and the possibility that other hyper-scalers (e.g., Meta) may follow Alphabet’s lead, investors would be derelict not to at least consider the supply-demand equation. Stock prices, after all, are like any other prices: A function of supply and demand.

Although Goldman expects 2026 to be a record year for the dollar value of US equity issuance, that figure has to be adjusted for market cap, which is to say netted against the change implied by stock returns.

As the figure above, also from Snider, shows, adjusted issuance remains negative consistent with the “shrinking market” trend that typified the post-dot-com bust era, with 2021 — year of meme stocks, SPACs and Bored Apes — standing out as the only exception.

“I have to pound the table on this potential supply-demand paradigm shift in stocks away from what had been the virtuous 15-year ‘de-equitization’ period where the float was shrunk and vol suppressed as a massive tailwind for stocks,” Nomura’s Charlie McElligott wrote, in his latest.

In the same note, he described a burgeoning “new world order” defined by big-tech cash burn (to fund AI capex) which “endanger[s] the flow of buybacks.” In other words: Decreased demand. At the same time, McElligott went on, “we have a ‘supply overload’ issue, not just via the sheer size of the tectonic AI-linked IPO calendar, but the hyper-scalers’ own supply.”

That’s a really critical point. Consider how this unfolded. Escalatory AI capex prompted more than $200 billion of hyper-scaler debt issuance to offset cash burn, and concerns stemming from all that new debt compelled the hyper-scalers (Oracle and Alphabet already, with more to come) to consider issuing equity instead, which means AI’s both curbing demand (by siphoning cash that might otherwise go to buybacks) and lifting supply (by compelling suddenly debt-laden tech firms to sell stock rather than more high-grade bonds).

The irony’s almost too much to bear: “We have to borrow to fund some of this spending so there’s money left over for buybacks. Sh-t, now we’ve borrowed too much. We need to think about a secondary.”

The figure above gives you some context for Mag8 equity issuance or, more to the point, underscores the extent to which there really is no context.

“[While] acknowledging that to make this even quasi apples-to-apples” you’d need to adjust for float and/or market-cap, it’s more about long periods of “no issuance at all” when it comes to “the rarity and pivot into supply,” Charlie wrote.

So, when you think about these massive, flashy AI-related supply events (e.g., SpaceX), don’t forget the less-publicized, but hugely important, supply-demand shift going on behind the scenes.

Of course — and the second chart from Goldman nods to this — it’s not as if corporate demand isn’t robust. Indeed, gross buybacks will easily outstrip issuance in 2026, leaving a very large, if diminished versus the prior three years, net corporate bid under the market.

Now to Snider’s “other” point. About Fed policy, I mean. Kevin Warsh simply can’t adopt a dovish cadence while headline inflation sports a four-handle. That’s not a tenable position, particularly not when some (many) market participants (not to mention one former Fed chair who still sits on the board) are already concerned about institutional independence.

Warsh will have to pay lip service next week to upside inflation risks in the US. The new SEP will almost surely reflect the same. There’s no way around that. And don’t forget, the US risks becoming odd man out. The ECB just hiked. The BoJ probably will this month, and the BoE later this summer.

Two-year Treasury yields are ~4.10% against EFFR 3.62%, core inflation’s a full point above target, factory-gate inflation’s running 6.5%, nominal growth’s something like 7%, the UNR’s sitting right at NAIRU and the S&P’s notched five-dozen record highs in the space of 17 months.

You can’t very well lean dovish against that backdrop if you’re Warsh. Not if you don’t want to torch whatever credibility you have (and whatever goodwill the market’s willing to extend you) at your very first presser.

All of that said — and this brings us full circle — it just doesn’t feel like we’re “there yet” vis-à-vis a real tipping point for frothy equities. I don’t see the bottom falling out until OpenAI and Anthropic are public, profit expectations are cut, revisions breadth rolls over, growth slows and/or the S&P trades north of its 2021 multiple. (Go Knicks.)