What happened
Morgan Stanley told clients that the four US hyperscalers chasing AI capacity will post a combined cash shortfall close to $600 billion in 2027, per a note surfaced by CryptoBriefing on Saturday. The call frames the gap as capex minus free cash flow across Microsoft, Meta, Alphabet and Amazon, the group now anchoring every serious AI infrastructure forecast. The bank's view is that operating cash flow, even at current margins, won't come close to covering the GPU, land, power and construction spend already in the pipeline.
The balance gets plugged with debt. That is the part the market has been slow to price. Capex guidance from the same group has climbed every quarter for six consecutive prints, and the Morgan Stanley number assumes that trajectory doesn't bend down.
If anything, the bank is treating current consensus as a floor.
Why it matters
Crypto has traded as a liquidity asset for most of this cycle, and the liquidity story is downstream of what happens in investment-grade credit. A $600 billion funding need from the most creditworthy issuers on the planet pulls supply into IG, nudges spreads wider at the margin, and competes for the same pool of capital that reaches for risk when yields compress. The AI capex boom has been the macro tailwind behind the Bitcoin-Nasdaq correlation since mid-2024.
It's also the reason power stocks, uranium, and anything plugged into data-center demand have run alongside crypto. If the funding side of that trade starts to look expensive, the asset side gets re-rated. The second-order read matters more than the headline.
Hyperscalers aren't going to default. But if they consume credit market oxygen at this scale, lower-rated issuers pay up, refi windows tighten, and the risk-on bid that has supported crypto through every macro wobble this year loses a leg.