In February, investors were practically fighting for a seat at the table. Alphabet's 100-year sterling bond, priced at a 6.125% coupon and maturing somewhere around 2126, drew an order book nearly 10 times the size of the £1 billion on offer. Five months later, that same bond is trading below 90 pence on the pound. Anyone who bought at par has already lost more than 10% of their principal. Not over a decade. In months.
The bond was part of a much larger multi-currency debt raise, with estimates putting the full package between $20 billion and $32 billion. Alphabet needed the cash for a reason. The company has flagged capital expenditure plans for 2026 in the range of $175 billion to $205 billion, almost all of it pointed at AI infrastructure. The century bond was a small slice of that, but it carried a certain symbolic weight: it was the first 100-year bond from a major tech company since Motorola did the same back in 1997. For perspective, Motorola in 1997 was selling StarTAC flip phones. The company has since been broken apart and sold off in pieces. Alphabet is betting it ends up nothing like that.
The buyers, as expected for paper this long, were pension funds and insurers. These institutions need assets that match their liabilities, things like pension obligations stretching 30, 40, 50 years into the future. A century bond fits that brief on paper. The spread at issuance was 120 basis points over 10-year gilts, which seemed tight but reasonable given Google's standing as one of the few companies on earth you'd actually trust to still be around and paying coupons in 2126.
Why the price fell, and what it actually means
Rising interest rates did what they always do to long-duration bonds: they crushed the price. This is duration risk in its most vivid form. The longer the bond's maturity, the more sensitive its market price is to any shift in rates. A 100-year bond is essentially the most extreme version of this trade imaginable. Even a modest rate move can wipe out years of coupon income when the duration is that long.
The coupon of 6.125% sounds generous for an investment-grade corporate issuer. It isn't irrelevant. But yield is not the same as total return, and the institutions that piled in at 120 basis points over gilts are now sitting on a position that behaves less like a bond and more like a volatile equity. The mark-to-market loss is real, even if the coupon keeps arriving on schedule.
Alphabet's aggressive borrowing to fund AI infrastructure is not unique in this cycle. The bet across the industry is that AI generates returns large enough to justify spending at this scale. Whether $175 billion-plus in capex this year alone pays off is a question that won't have a clean answer for years. The century bond, by design, won't mature until your great-grandchildren are thinking about retirement.
This article is for informational purposes only and does not constitute financial advice. Bond prices can fall as well as rise.
