Who Is Left to Say No When Everyone at the Table Is on the Same Side of It

Eighty-six times revenue, a 23% software move, and a front end that declined to comment

Weathervane. The wind is still coming from the long end. The policy rate has been frozen at 3.50% to 3.75% since July whilst the term premium does the tightening, and the price of long money is set by deficits, issuance and a global duration bid rather than by the committee. Higher-for-longer has stopped being a Fed decision and become a bond-market fact. The cruise ship holds this heading, and Thursday moved it neither way.

Ahoy there, Trader! ‍‍⚓️

It’s Phil…

Here is the sequence that has been bothering me since Wednesday night, and I want to lay it out plainly before I try to interpret it, because the facts are firmer than my reading of them.

Wednesday, on the earnings call, Nvidia said its constraint is supply. Demand runs beyond even the 70% growth guided for fiscal 2028. It cannot build fast enough.

Wednesday night, per The Information, it agreed to pay $12.9bn for Hugging Face. Reuters puts that company’s annualised revenue near $150m.

Thursday, Salesforce closed 23% higher on a deepened partnership with Anthropic. Salesforce Ventures led the 2023 round that valued Hugging Face at $4.5bn. Anthropic has separately committed roughly $45bn over six years to compute running on Nvidia’s next-generation systems.

Each of those is ordinary on its own. Together they describe a small room.

Which leaves the question I keep circling.

So: if an industry is financing its own demand, what is the interest rate actually for?

Walking the dots. The striking thing about Thursday is what did not happen. The S&P rose 0.72% whilst the equal-weight version of the same five hundred companies fell 0.3%. Ten of eleven sectors declined. The 10-year eased two basis points, the 30-year two, the 2-year one. Volatility fell 4.73% to 14.50 on the eve of a new Fed Chair’s first Jackson Hole keynote. Not one instrument outside the AI complex treated any of this as information.

Textbook versus reality. The textbook says capital is allocated by parties with opposed interests. A buyer wants to pay less, a seller wants more, a lender wants covering, and the price that survives the argument carries information. The interest rate is the referee in that argument: it is what the marginal outside investor demands before parting with money.

That mechanism needs one condition, rarely stated. The parties have to be independent.

What I cannot work out is what happens to the referee when the buyer is also the supplier, the investor and, increasingly, the owner. Nvidia’s finance chief addressed this directly on the call, saying investments reflect independent commercial judgement with no tie-in, and I have no basis to doubt her. But even taking that entirely at face value, an eighty-six-times-revenue price agreed inside a small group of firms is not a price the outside world argued its way to. It is a price that group agreed among itself.

Which may be exactly why the 2-year moved one basis point. Not because policy is irrelevant, but because the marginal buyer of AI compute is not currently borrowing at the policy rate. He is being financed by his supplier.

I do not know whether that is a durable structure or a late-cycle tell. I know that if it holds, Warsh’s twenty minutes this morning are addressed to a part of the economy that has stopped listening.

Phil’s Musing

The bit I keep turning over is that none of this needs anyone to be doing anything wrong. Every transaction here is defensible on its own terms and probably correct on its own terms. It is the aggregate that looks strange, and aggregates do not have an author. My honest lean is that when a sector can fund its own customers, the cost of capital stops disciplining it, and the thing that eventually disciplines it instead is demand from outside the room. I would rather find out which quarter that arrives in than argue about the multiple.

Happy trading,

Phil
Less Brain, More Gain
…and may your trades be smoother than a cashmere codpiece

P.S. – Phils Footnote I nearly wrote this letter as an outrage piece about eighty-six times revenue, and I am glad I did not, because the multiple is not the interesting part. Multiples are opinions. The interesting part is that I could not find the counterparty whose job it was to disagree with it, and I looked. If you can find one, tell me, because I would like to be wrong about this before Monday rather than after.

 

 


You may also like

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}