Although I’ll take it vis à vis my early-October suggestion that US yields had likely seen their near-term peak, I’m by no means convinced the two-week Treasury rally constitutes an “out of the woods” moment for bonds.
Most obviously, the macro environment remains very challenging for bond bulls. All nascent “growth scare” concerns aside, nominal GDP in the US is up a staggering 65% since the Q2 2020 COVID lows, inflation remains well north of the Fed’s target and we live in a world defined by populist politics and fiscal dominance.
At the same time, intense competition for scarce capital to fund the AI buildout has to be viewed in the context of the shifting buyer base for government bonds.
Already, the US Treasury market was adjusting to a new reality defined by actual price discovery as the price-agnostic bid from the Fed and the foreign official sector dissipated, leaving more supply for price-sensitive investors.
To the extent their mandates allow, some of those investors are also buyers of high-grade corporate debt, and quite a bit of the AI-related IG issuance is long tenor.
The figure above’s a reminder: More than half of the compute debt deluge is 10 years and out.
As Treasury yields moved ever higher over the course of 2026, market observers repeatedly suggested that hyper-scaler supply (and expectations of more to come) was a major factor in explaining the US long-end’s inability to catch, let alone hold, a bid.
Now, though, with spreads on hyper-scaler long-tenor debt having ballooned from ~55bps in September of 2025 to ~130bps currently, the competitive environment could shift in Treasurys’ favor if a higher cost of capital makes the big-tech spenders think twice before leveraging the balance sheet any further.
Through that lens, it’s fitting (and in some sense ironic) that the pullback in tech shares and accompanying widening in all things “AI spreads” should be the catalyst that flipped the stock-bond correlation back negative.
The figure above gives you a sense of just how quickly things can change.
“‘Bonds as your hedge’ has found itself working again because the market has suddenly assigned increasingly credible delta to an outright US economic growth scare scenario, as the ‘cracks in the AI trajectory’ story is metastasizing wrong-way,” Nomura’s Charlie McElligott said Friday, adding that “the AI outlook has shifted from the prior ‘heads I win, tails you lose’ to now ‘glass totally and completely empty’ view.”
That latter point’s important. Markets tend to overshoot, particularly in situations like this one where a crowded trade (short bonds) gets stopped out and the resulting squeeze amplifies the price action, which is in turn interpreted as “validation” for what was an otherwise tenuous narrative (in this case the US growth scare story).
Of course, fading this sort of cascading move can be treacherous — because it’s very difficult to know whether the squeeze has truly run its course. Put as a question: Is positioning clean now, or just cleaner?
I don’t have the answer(s). But I would note that the Treasury rally came “ahead of schedule,” so to speak, in the context of the November QRA, which may find Scott Bessent tipping actual cuts to coupon auction sizes as soon as H1 2027.
If Treasury were to pull any such reductions forward and make them a “now” thing next month, well then look to November/December 2023 for your bond rally extension potentials.
