What Yesterday's Most Active Corporate Bonds Say About Who's Actually Buying
By Paul Rubillo ·
One of the features we’re most excited about sharing on TreasuryBonds.com is the Most Active Bonds section (Muni’s and Corporates). It’s hard to find that data and one of the things we learned quickly about fixed income data is how sacred it is. Vendors have many stipulations on what they will share and then what we can even do with the data in terms of displaying it. Six of the twelve bonds on yesterday's most-active list carried a name tied directly to the AI buildout: two from Nvidia, two from Alphabet, one from Amazon, and one from Analog Devices, a chip maker that supplies the same data centers those other four are financing. The rest of the list belonged to Aon, the insurance broker, and Verizon, names with nothing to do with AI at all.
Let’s look at Nvidia's 2036 note, which carries a 4.95% coupon. The 10-year Treasury closed the same session at a 4.979% yield. A company coupon sitting essentially on top of the risk-free rate on a comparable maturity is a striking data point, though it needs a caveat the trading screen itself doesn't give you: coupon isn't yield to worst, and this table doesn't show YTW for any of these bonds. A 4.95% coupon bond trading above par yields less than 4.95% to a buyer today; trading below par, it yields more. Without the price or the YTW column filled in, what we know for certain is narrower than it looks: Nvidia can borrow at a stated rate that undercuts the government's own 10-year borrowing cost, which says a great deal about demand for that name specifically, even before you know what an actual buyer earns on it today.
Amazon's 2036 note carries a 5.30% coupon, about 32 basis points over that same 10-year Treasury. Alphabet's 2046 note carries a 6.25% coupon against a 20-year Treasury yielding 5.385%, an 86 basis point gap, and its 2056 note runs 6.375% against a 30-year Treasury at 5.351%, just over 100 basis points wide. Aon's bonds, by contrast, run wider still at comparable maturities, its 2036 note at 5.95% against that same 4.979% 10-year, a gap of about 97 basis points, and its 2056 note at 6.45% against the 30-year, about 110 basis points over.
Better-regarded borrowers pay less to borrow, and Aon isn't Nvidia. That said, Aon's 2036 note was the single most active bond of the day at $34.8 million in volume, ahead of every AI-linked name on the list including Nvidia's own two issues combined. That's not a story about AI dominating fixed income trading outright. It's closer to AI-linked paper and old-economy paper sharing the same actively traded shelf space, at noticeably different spreads, on a day when Treasury yields themselves moved in a way that touches both.
Every maturity on the Treasury curve fell in the same session, from the 1-month bill down to the 30-year bond, none of the moves large on their own, all of them pointed the same direction. One curve detail worth flagging on its own terms: the 20-year yielded 5.385%, higher than the 30-year at 5.351%. That's not an inversion in the sense that matters for recession signaling. It's a long-documented quirk specific to the 20-year tenor, which has traded cheap relative to its neighbors on the curve fairly consistently since Treasury reintroduced the maturity in 2020, largely a function of thinner liquidity and lighter index demand for that specific point on the curve rather than anything being said about growth expectations.
Alphabet's 2056 note and Aon's 2056 note are both 30-year paper from today's vantage point, and Verizon's floating-rate note runs to 2059. Duration measures how much a bond's price moves for a given change in yield, and it isn't the same number as years to maturity, though the two move together in the same direction. A 30-year corporate note at these coupon levels likely carries an effective duration somewhere in the neighborhood of 15 to 17 years, an estimate rather than a figure taken from this data. Using that as a rough illustration, a one percentage point rise in the yields affecting that bond could produce something like a 15% to 17% drop in its price, before accounting for convexity or any change specific to the issuer's own credit. That's a real number to sit with if you might need to sell before 2056 rather than hold to maturity. It matters much less if the plan is to collect the coupon for three decades and let the bond mature at par, in which case the price swings along the way are noise rather than a loss you'd ever actually realize.