And some companies more or less already are, implicitly if not explicitly.

• Microsoft just cut off Claude Code licenses, and Tom Warren at the Verge claim that this is at least in part because of costs.

  

• Target has expressed some anxiety about pricing models for AI agents.

• Starbucks just shut down an AI inventory experiment that they had been experimenting with because they realized that it couldn’t be trusted:

None of this should be surprising; I have been pointing out the errors since the week ChatGPT was introduced in 2022, and study after study after study (I will go into this more in a later newsletter) has shown that most companies don’t seen a significant return on their AI investment.

The overall situation is this:

• Three companies that have not yet shown themselves to be profitable are expected to soon IPO for a total of something like four trillion dollars.

• Index funds, the staple of many people’s retirement funds, are going to be more or less forced to rapidly absorb these exercises in fantastical thinking.

• Those exercises in fantastical thinking are premised on the notion that customer demand will be essentially endless.

• But we are already seeing cracks in that fantasy.

• If enough customers have second thoughts, none of the three IPO’ing companies will ever hit their long-term projections.

• In which cases those stocks will eventually fall.

• A lot of banks may take a hint as well.

• Get ready for “too big to fail”, and watch your portfolio drop.

Disclaimer: I can’t give you investment advice (let alone tax advice), and the market might remain irrational for longer than any of us can remain solvent.

But I see what I see, and it just doesn’t make sense.