The race to fund AI infrastructure has turned into a tug-of-war for capital between hyperscalers and governments. Bank of America strategists say that is pushing up financing costs for developing nations, layered on top of pressure from geopolitics, elevated oil and tighter central-bank policy. Average yields on dollar bonds from emerging markets have risen by nearly a full percentage point in the second half, reaching their highest since April last year.
At BofA, Merveille Paja, a sovereign credit strategist covering Eastern Europe, the Middle East and Africa, warned that refinancing could get tricky for some governments. "The risk that hyperscalers actually are crowding out emerging markets really is a risk for our asset class," she said, noting that tighter global financial conditions could snag borrowers waiting for a better window.
A Bloomberg index shows how tight things got earlier this year: in June, EM sovereigns were paying as little as a 1.5 percentage point premium over Treasuries, roughly half the five-year average. Yields on EM dollar debt now sit at an 18-month high.
By Bloomberg's count, Amazon.com Inc. and Alphabet Inc. together have sold about $70 billion of bonds in the second half. In contrast, emerging-market sovereign eurobond issuance since July 1 has fallen to $45 billion, down from $91 billion over the same period a year earlier. For investment-grade EM issuers, volumes have dropped 56% to $29 billion, and October has not recorded a single sale so far.
Deadlines are looming, too. By year-end, developing countries have $27.7 billion in external debt obligations denominated in dollars and euros coming due. The largest near-term needs are Saudi Arabia at $5.2 billion and Turkey at $3 billion. For consistency, the tech totals are in dollar-equivalent terms, while the EM sovereign numbers refer to dollar-denominated external bonds.
When the largest borrowers crowd a market, smaller ones pay more. Market Briefs covers that spillover free every weekday.
Worldwide, AI-related borrowing has already moved past $445 billion this year, Bloomberg calculates. Bloomberg Intelligence projects that hyperscaler borrowing could reach $500 billion next year. Bloomberg also reported that Elon Musk's SpaceX is discussing raising $40 billion to buy Nvidia chips, potentially including $30 billion in investment-grade bonds.
This cash hunt is happening as budget shortfalls remain large in the US and Europe and the developing world's financing needs grow, forcing investors to be choosier as the cheap-money era fades. Investors are comparing richer yields from profitable AI-linked companies with EM sovereign bonds that must contend with the ongoing US-Iran war and an upswing in US Treasury yields.
The squeeze is felt most where investors can easily swap one bond for another. "Some pressure has materialized among the tightest-trading investment-grade sovereign issuers," said Micah James, Vanguard's global head of emerging-markets research. "Investors can more easily substitute between highly-rated EM sovereign debt and investment-grade technology issuers."
It is not all zero-sum. "For us, it is not an either/or situation of picking an EM or a hyperscaler bond," said Eoghan McDonagh, Allianz SE's lead portfolio manager for EM debt. "We monitor the spread between the two, to help judge when valuations are attractive. There has been good demand from US accounts who have been using EM sovereign investment-grade names as a diversifier from some of their hyperscaler exposure."
And in a world where money costs more, some expect a sharper pecking order. "When money becomes expensive again, choosing becomes essential," said Edouard Carmignac, chairman and CIO at Carmignac, an investor in Mistral, SpaceX and Revolut. "Scarcer money favors the most promising companies, particularly those harnessing augmented intelligence, while crowding out governments suffering from fiscal incontinence."
For you, the upshot is simple enough: there is more highly rated, AI-fueled corporate paper competing for attention and fewer new EM sovereign deals since July. With yields higher and year-end EM maturities approaching, the price of funding governments could stay under pressure in the near term.
AI financing is reaching emerging market borrowing costs. Join Market Briefs free and follow the connection.
