The Anthropic Mythos: How an Emergency Meeting Between the Fed, Treasury, and Wall Street Could Crack the AI Trade Wide Open

Special Edition April 11, 2026

The Federal Reserve and the Treasury Department do not call emergency meetings with Wall Street CEOs on Friday afternoons because everything is fine. They do it because something is broken, or about to break, and the people in the room know it before the rest of us do.

On Friday, April 10, 2026, at 3:47 PM Eastern Time — seventeen minutes after the equity market closed — a call went out from the New York Fed to a short list of names that should make every serious investor pay attention. Jamie Dimon at JPMorgan. Brian Moynihan at Bank of America. David Solomon at Goldman Sachs. Jane Fraser at Citigroup. Larry Fink at BlackRock. Marc Rowan at Apollo. Jon Gray at Blackstone. The public framing, when it leaked Saturday morning through a Bloomberg scoop, was anodyne: a coordination meeting on “AI infrastructure financing stability.” The language of a press release written to say nothing.

I have been around this business long enough to know what that phrase actually means. “Financing stability” is the polite term regulators use when a market is no longer able to fund itself through normal channels. It is the language of 1998, when the New York Fed’s William McDonough pulled fourteen bank CEOs into a room to bail out Long-Term Capital Management before its unwind took down the entire global derivatives complex. It is the language of October 13, 2008, when Hank Paulson sat nine bank CEOs around a table in the Treasury building and told them they would not be leaving until they accepted TARP capital injections — voluntary only in the sense that the door was locked. It is the language of the March 2020 weekend when the Fed slashed rates to zero and restarted quantitative easing on a Sunday night because the Treasury market had stopped functioning. It is the language of the March 2023 weekend when Treasury, the Fed, the FDIC, and the White House coordinated by phone to backstop uninsured depositors at Silicon Valley Bank before the Asian open on Monday could turn a regional bank failure into a contagion.

Most investors this weekend are watching the price action on Friday’s close and asking whether Monday opens green or red. I am watching something else. I am watching the fact that the most powerful financial officials in the United States felt they could not wait until Monday morning to have a conversation about how the AI trade is being financed. That tells me everything.

Let me put this into perspective. Over the past thirty months, this market has been driven by a single narrative — a narrative so powerful, so totalizing, and so lucrative that it has pulled trillions of dollars of capital, credit, and political will into its orbit. I call it the Anthropic Mythos. Not because Anthropic is the largest or the loudest of the frontier AI labs, but because it is the clearest distillation of the belief system underneath the trade. The mythos is simple. It goes like this: artificial intelligence is the next electricity, the next internet, the next industrial base. The frontier labs building it — Anthropic, OpenAI, xAI, Google DeepMind — are not companies in the traditional sense. They are the substrate of the next economy. Every dollar of GPU capex will be repaid through productivity. Every private round will be absorbed by willing capital. Every narrative of scale will find its capital partner. And if the math looks insane, it is only because we lack the imagination to see the trillion-dollar cash flows on the other side of the curve.

The numbers underwriting this mythos are staggering. Anthropic, the maker of Claude, was last marked at a $183 billion private valuation at the end of 2025 — up from $61 billion twelve months earlier. OpenAI sits near $500 billion in its most recent secondary round. xAI has quietly crossed $200 billion. Hyperscaler AI-related capex — Microsoft, Google, Meta, Amazon — is now running at an annualized pace north of $400 billion, a figure that exceeds the peak annual capex of the entire telecom buildout in 2000. The Magnificent Seven, whose earnings and capex are almost entirely tethered to the AI narrative, now represent roughly 35% of the total S&P 500 market capitalization — the highest concentration in any index in American history, including 1929, 1972, and 2000.

That is the mythos. That is the trade. And on Friday afternoon, the Fed and the Treasury decided it could no longer wait until Monday.

This is a special Sunday edition because I do not think my readers should walk into the April 14 open without understanding what likely happened in that room, what it means for the AI trade, and — most importantly — which scenario this emergency meeting is tracking: the 1998 tail-risk playbook, or the 2008 break-glass playbook. When the Fed and the Treasury start calling emergency meetings about how artificial intelligence is being financed, pay attention. The cracks in the story almost always show up in the financing first. They always have.

Link Substack to Discord