By James Mackintosh
Lending money to Alphabet for a century might seem to raise some serious questions. Will we still use Google to search the internet in the 22nd century? Will the internet even exist? Will America make it to its 350th birthday?
In fact these are risks investors in the rare century bond issued on Tuesday can safely ignore, thanks to the mathematics of bonds (more on that later). The risks investors should be focused on are more mundane, but also closer to hand: The prospects for artificial intelligence, the increasing debt load of Big Tech and the risks that come with joining a crowd throwing money at a fashionable industry.
Start with the crowd. Century bonds get issued when money is easy. The first wave came in the mid to late 1990s, when companies had a lower yield compared with safe Treasurys than any time since. Coca-Cola issued the last of these 100-year bonds in May 1998, not long after the last technology 100-year issue, from Motorola. The cost of corporate debt compared with Treasurys jumped after hedge fund Long-Term Capital Management imploded.
The second wave came when money was actually free during the period of zero interest rates. Austria managed to issue zero-coupon 100-year bonds in 2020, while in the late 2010s investors started by lending to universities and Mexico, but eventually were willing to lend for 100 years to flaky sovereigns including Argentina, attracted by what seemed like high yields compared with earning nothing on cash.
It didn't end well, with Argentina defaulting after just three years and Austria's bonds now worth just 5% of what they were worth at issue, as zero rates proved temporary.
We'll have to wait to see if Alphabet's sterling century bond yielding 6.05% is the harbinger of another wave of issues, but it is certainly tapping the markets when money is cheap for companies.
The spread of corporate yields over Treasurys last month hit the lowest since just after Coke's 1998 bond, amid strong demand for the safety of high-quality issuers. This is a great time for companies to borrow; it isn't obviously a great time to lend to them.
Ordinary credit risks are easy to dismiss for Alphabet. It is sitting on $126 billion of cash and marketable securities, borrows less than half that and is rated AA+. That doesn't mean its debt is safe for investors, though, because Alphabet is engaged in a race to spend as much as possible as fast as possible on AI.
Alphabet's AI chatbot, Gemini, has proved popular, but is up against OpenAI's ChatGPT, Anthropic's Claude, Chinese developers and others to grab market share -- and this is before it is able to charge anything like the cost of running it to customers.
The business model of AI remains, to be polite, in flux. If businesses and individuals eventually prefer low-cost open-source AI models, or demand collapses when prices are raised to cover costs, all the current leaders might suffer.
Alternatively, if one proves far better than the others it might grab most of the market, as Google did with search engines. For Alphabet, there's also the danger that AI replaces traditional web search, undermining its core cash engine.
If it goes wrong, the wasted cash will mostly be a problem for shareholders. But bondholders are still exposed, as less of a cash cushion means weaker credit quality, and higher yields.
The final risk is the borrowing to come. Big Tech firms are pouring cash into AI and data centers, and raising hundreds of billions of dollars in debt to finance them. As well as changing the nature of the companies from capital-light software toward capital-heavy AI infrastructure providers, it will test the willingness of lenders.
Alphabet's oversubscribed issue shows that for now investors are quite happy to finance its AI spending spree, but as the industry borrows more, it may hit a limit and lift yields across the tech sector.
Back to the mathematics. Investors know that what matters isn't whether they will get their money back in 100 years' time, but the interest payments over the next few decades. The final payment of principal, if it happens, will be just 0.28% of the total payments Alphabet is promising on the bonds. They should behave very like a conventional 40-year bond, available from tech companies including Oracle, Cisco Systems, Intel and Apple in recent years.
One way bond investors measure risk is duration, the time it takes for a bond to pay back the original investment. For Alphabet's 100-year bond, that's just under 17 years. If it was a 40-year bond with the same terms, it would still be more than 15 years, not a huge difference.
We should know long before then if Alphabet's heavy spending has paid off and made it an AI winner. Investors should worry about that, not the state of the world in 100 years.
News Corp, owner of The Wall Street Journal, has a commercial agreement to supply content on Google platforms.
Write to James Mackintosh at james.mackintosh@wsj.com
(END) Dow Jones Newswires
February 11, 2026 12:00 ET (17:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.