There’s a new big short. Or “best short.” Somebody call Michael Burry and tell him to un-retire. Again.

We’ve likely seen “peak easy” financial conditions in the US and therefore “trough credit spreads,” which in turn means wider CDS, particularly for the AI hyper-scalers, whose spending plans are the talk of the proverbial town.

That’s all according to BofA’s Michael Hartnett, whose popular “Flow Show” series is to the work week what charred, smoldering vehicles are to highway traffic jams: Something tangled and confused to ogle after an annoyingly long wait.

There’s a chart. It could’ve (and very well might’ve) walked right out of a blog post penned by a couple of stay-at-home dads in whose hands a Bloomberg terminal’s a deadly weapon.

(I just had a lot of fun at a lot of people’s expense. For those of you who caught all the jokes, congratulations: You’re a nerd who spent the better part of two decades immersed in the narrow world of sell-side commentary and the financial bloggers who leeched off it.)

No, but seriously: It’s a decent chart. Oracle’s becoming the poster child for the AI hyper-scaler spending / cash flow juxtaposition at the heart of many a bear narrative.

Now for the caveats. First of all, Hartnett’s using two different y-axes up there. It’s a beautiful thing, really: The best chart crimes are misdemeanors in which the scaling’s too subtle for most people to notice (20bps in this case). The two series will map pretty well onto one another regardless (when you’re talking about blue-chips, default risk tends to be cyclical not idiosyncratic), but not quite that well.

Oracle’s not going to just stop spending on AI (the opposite in fact) which means its CDS is likely to keep trading wide. Given that, the only conclusion one can come to about that chart is that it’s meant to suggest the most widely-referenced gauge of indicative credit risk for dozens of high-grade corporates is going to abruptly ~double in a “catch-up” trade to the widening in Oracle’s five-year CDS.

Spoiler alert: That’s not going to happen, certainly not overnight as the chart seems to wink at. What you’re seeing in Oracle’s CDS is hedging by people with exposure to the company’s debt in the face of ramped up borrowing to fund AI infrastructure. Sure, someone’s naked on that CDS (i.e., just gambling), but there’s a structural component to it.

All of that said, I take Hartnett’s broader point, which he summarized as follows:

So long as the dollar [doesn’t appreciate too sharply] there’s no reason to discount [tighter financial conditions], but the bulk of FCI easing is behind us [which is] why markets are more nervy about credit cracks and financing capex booms, and why the best short is AI hyper-scaler corporate bonds.

That works as a talking point, and there’s nothing implausible about it as a trade, even as the vast majority of investors obviously won’t be shorting IG corporate bonds (or buying default protection on hyper-scalers).

But to suggest — accidentally, obliquely or otherwise — that the broader high-grade universe is on the brink of a paroxysm based on the historical “correlation” between Oracle’s five-year CDS and CDX.IG is to traffic in wild extrapolation.