Artificial intelligence may be partly to blame for rising bond yields, but it’s also creating opportunities for income investors. Issuance by hyperscalers is surging as companies seek to fund the buildout of artificial intelligence. This year, Alphabet, Amazon, Metaplatform and Oracle have issued nearly $223 billion in corporate bonds as of Aug. 20, according to LSEG. This is more than double the amount seen in all of 2025, the company found. This flood of supply could be responsible for the rise in sovereign bond yields, said Krishna Guha, head of economics and central banking strategy at Evercore ISI. “High-quality hyperscalar bonds are close competitors to government bonds and are also hedged in the sovereign debt market. Relative value investors seek to isolate risk in selected issuers and offset underlying macro and broader market risks,” Guha wrote in a note on Wednesday. “This issuance is becoming increasingly global across currencies, helping to explain the global nature of yield movements.” He added that upward pressure on yields also reflects investors’ expectations that supply will increase significantly going forward. The 30-year Treasury yield topped 5.33% on Tuesday, its highest level in 19 years, but fell Wednesday after the Treasury announced it would ramp up bond buybacks. However, yields rose again on Thursday. Wall Street also blamed the ballooning budget deficit and rising inflation concerns. Total U.S. government debt surpassed the $40 trillion mark as of Tuesday, more than doubling in a decade. Capturing Attractive YieldsThese days, investors can capture yields of 4.75% to 8% on hyperscalar bonds, depending on the specific issuer and maturity date, said Dominic Pappalardo, chief multi-asset strategist at Morningstar Wealth. Bonds are mostly investment grade and have long maturities. Pappalardo said the assets are inherently riskier because more debt means the company has to be better able to repay and service it. But these companies are very large and healthy, he noted. “Debt levels have risen quite a bit, but from my perspective that’s not necessarily a cause for concern, because they’re very profitable, they still have a lot of cash on hand, and their business is still growing at a very meaningful pace,” he explained. Leslie Falconio, head of taxable fixed income strategy in UBS America’s chief investment office, believes there is value in high-quality bonds from large hyperscalers. The supply was a surprise to the market, he said, but there was demand from insurance companies, pension funds and foundations. “This is a great time to lock in those yields,” Falconio said. “It’s just going to widen some of the (credit spreads) because there’s more supply coming into the market.” Credit spread refers to the difference in yield between a risky bond and a government bond of the same maturity. Widening spreads indicate that investors are demanding higher rewards for taking on risk. Still, Vishal Khanduja, head of broad-market fixed income at Morgan Stanley Investment Management, said that while the yield is “very attractive,” investors should do their homework because there is wide variation in the market. “Not every issue that comes up will have the same creditworthiness, structural quality and project effectiveness that you’re looking for,” he said. “Thus, for the same group of AI infrastructure debt being issued, the outcomes after two to three years can be very different and very diverse.” He estimates that investors could earn about 6.5% on long-term debt issued by investment-grade AI leaders. Spotting Opportunities UBS Falconio is a long-term investor with a focus on larger, higher quality issuers with good cash flows and strong balance sheets. “You could earn a lot of money with compound interest,” she says. “When we buy these bonds, especially long-term bonds, most of the time we’re not doing it to flip. We’re doing it to lock in the yields that we’re seeing.” Khanduja and his team do research and analysis before buying each security. They consider things like who is going to take the deposit, which is basically the customer who is buying the project, such as the data center operator. Additionally, consider whether the contract includes a condition that the off-taker will make a payment if the lease is not accepted. “The location is very important. What type of power agreement? Because there is a huge lack of that in the U.S. right now,” he said. “What kind of government approvals do they have? What is the timeline? Who are the contractors? What is their experience and success rate in completing projects?” Oracle is one of the stocks in the Eaton Vance Total Return Bond ETF, which Khanduja manages. Check your exposure Individual investors who already own core bond funds already have some exposure to hyperscalers. About 8% of the investment-grade index is included in such assets, Falconio noted. Morningstar Wealth’s Pappalardo said those who want to stick with exchange-traded funds but want more exposure might consider corporate bond funds. Otherwise, investors can work with a financial advisor to buy individual bonds, he said. Make sure your portfolio is diversified on both the bond side and the stock side. The equity side may already have a lot of high-tech investments. Beware of the risks Although most of the hyperscalers’ debt is investment grade, some AI investments remain speculative, Pappalardo said. “We don’t yet know how much return we’re going to get from that investment,” he says. “In many ways, these companies are counting on revenue and profits that haven’t yet materialized. I’m not saying they will or won’t materialize. We just haven’t seen it yet.” Additionally, companies are likely to continue issuing bonds, potentially flooding the market with supply again.
