Sixth Street Warns of Triple Mismatch Signals Converging in AI Boom
nashnova research
Sixth Street CEO Alan Waxman warns that AI infrastructure is flashing three simultaneous red flags — credit expansion, rising speculation, and supply-chain distortion — a combination that has historically preceded market dislocations.
"The largest capex boom ever" — what's the core problem?
Waxman calls the current AI buildout "the largest capital-expenditure boom in history."
The fundamental issue: nobody truly understands AI's unit economics, because those economics keep shifting across the entire stack.
This means → the money is flowing in, but the industry itself cannot calculate the return per dollar — and that uncertainty is the root of every risk that follows.
Banks with "indigestion," chips piling up in warehouses — what's breaking?
On credit: some banks are trying to offload data-center loans, showing signs of "indigestion." Waxman says several worry they "didn't really read the documents the way they should have."
On supply chains: roughly $200 billion worth of chips are sitting in warehouses because the data centers meant to house them aren't ready.
In plain terms = the chips arrived, but the buildings didn't; the loans went out, but the banks only later realized the risks — a double mismatch hitting both sides at once.
How severe is the power shortage?
Waxman frames the power problem as simple arithmetic: there isn't enough electricity available to buy.
What is available is getting more expensive and takes longer to connect.
This means → even if funding and chips are both in place, no power means no running data centers — electricity is the hard bottleneck for the entire chain.
Why are financing costs about to jump?
Insurance-company lenders have been the main capital source for AI chip financing, but they are approaching single-name concentration limits on clients like Microsoft and Nvidia (rules capping how much they can lend to any one borrower).
Borrowers will be forced to turn to pension funds and sovereign wealth funds, which demand returns of 8%–9% — well above the 5%–6.5% insurers typically charge.
Waxman notes: "When that covers 70% of the debt structure, it's a material difference."
In plain terms = the cheap money is running out, and what comes next costs significantly more — the entire AI buildout faces a step-change in funding costs.
Demand is real — but what does a correction look like?
Waxman stresses that "demand is real" and the AI transformation will ultimately succeed.
But he argues the severity of any correction depends on three variables: connectivity (some call it "cyclicality"), leverage, and asset-liability mismatch — "when those three things are true at the same time, that's when dislocation typically occurs."
This reflects a deeper signal: Waxman is not bearish on AI itself — he is warning that the resonance between jumping financing costs and idle chip inventories may become the key test of whether dislocation materializes.
市场有风险，内容仅供研究参考，不构成投资建议。