Over the last 12 months, the US hyperscale technology companies have borrowed huge sums to finance their gargantuan artificial intelligence capex plans. So far, they have relied mainly on senior bonds, with some also turning to the equity market.
The scale of the investment boom inevitably raises the question: could hybrid capital bonds play a part?
Estimated 2026 capex for the five largest hyperscalers — Amazon, Microsoft, Alphabet, Meta and Oracle — has climbed to $775bn, from roughly $448bn in 2025.
Bank of America expects these five companies to spend around 90% of their operating cash flow on capex this year, up from 65% last, loading pressure on free cash flow and debt metrics.
So far this year, the hyperscalers have raised $62bn-equivalent of Reverse Yankee bonds in euros, sterling, yen, Swiss francs and Canadian dollars, according to CreditSights analysis.
The latest came on Monday: Amazon's largest ever Maple bond, scooping C$14bn in five tranches.
On top of this, these firms have raised $107bn in their home market, CreditSights said.
These senior markets, however, are not bottomless. Although demand is still rampant for any new hyperscaler deal, there is likely to be an upper limit to how much investors are willing to take down. And the $169bn of bonds so far issued are a long way from $775bn of capex.
Having already squeezed plenty of juice out of senior bonds, several hyperscalers have turned their attention to raising equity instead.
Oracle, for instance, told investors it would raise up to $50bn of debt and equity financing this year. So far, the Austin-based firm has issued $25bn of senior bonds in a single shot, as well as a $5bn convertible bond. It has yet to return for the final $20bn of equity fundraising.
Last week, Alphabet said it would briefly pause its debt capital markets onslaught and raise $80bn of equity instead.
But so far, none of these firms has mentioned hybrid bonds.
Logical appeal
Every corporate DCM banker in Europe would have the pitch ready. For issuers with large investment programmes, hybrids can offer a useful alternative to issuing equity.
Obtaining 50% equity credit is now straightforward, with well established structures, and the extra cost over issuing senior bonds is less than ever. SSE, the UK electricity company, paid 90bp-100bp over senior for €1.3bn of hybrids with 5.25 and eight year call dates last week.
For that pittance, SSE got €650m of equity, from an accounting point of view, protecting its credit ratings without diluting shareholders.
These arguments have been convincing more and more companies, especially in the US. Moody's simplified its rating methodology in 2024, eliminating a previous rule that for US companies, hybrids had to be the same level of subordination as preferred shares to get 50% equity credit.
With that gone, these instruments now offer US issuers the same benefit European companies have enjoyed for more than a decade: 50% equity credit and interest that is tax-deductible for the issuer.
US issuers have reacted by increasing hybrid issuance from $7bn in 2023 to $32bn and then $44bn last year, according to the Intercontinental Exchange.
Firms with heavy capex plans, like utilities and telecoms companies, have dominated hybrid issuance this year. These two sectors have produced 55% of the supply in euros and 93% in dollars, data from GlobalCapital's Primary Market Monitor and Dealogic show.
Other more cyclical borrowers such as Stellantis and General Mills have also used this instrument to defend investment grade ratings with great success.
Recent hybrid trades have also attracted larger books than many senior deals, as investors chase yield by buying high quality subordinated paper. Euro hybrids have been on average 4.9 times covered at final terms, against 3.5 for senior bonds, PMM data show.
Investment firm Twenty Four Asset Management has also speculated about whether hyperscaler hybrids could be coming.
Too big to care
The arguments for hybrids make perfect sense — on paper.
But for most of the hyperscalers, this market as it stands is beneath their notice. They are simply too big and rich. As the table below shows, most of them are in the double-A rating band, and despite their enormous spending plans, their rating outlooks are stable. Bond investors are so far happy to keep pouring money into their begging bowls.
If the balance sheet metrics start looking out of whack, hey, they can raise some equity. Alphabet's planned $80bn equity raise would have been by some distance the world's biggest ever equity capital markets deal. After launch it increased it, to $84.75bn. Even at that colossal scale, it was barely a scratch on Alphabet's then $4.6tr market capitalisation — about a 1.8% dilution.
Hyperscaler funding so far this year
Hyperscaler (Moody's/S&P ratings, stable unless stated) | Market cap (as of Tuesday) | Expected 2026 capex (% of market cap) | Senior bond borrowing year to date | Planned equity funding for 2026 |
Alphabet (Aa2/AA+) | $4.4tr | $200bn (4.5%) | $51.9bn-equivalent | $84.75bn |
Amazon.com (A1/AA) | $2.6tr | $180bn-$190bn (7.3%) | $67.5bn-equivalent | None yet |
Meta Platforms (Aa3/AA-) | $1.5tr | $125bn-$145bn (9.7%) | $25bn | FT reports plans |
Microsoft (Aaa/AAA) | $3.1tr | $190bn (6.1%) | None yet | None yet |
Oracle (Baa2 negative/BBB negative) | $609bn | $50bn (8.2%) | $25bn | $25bn |
Source: company reports, Dealogic, rating agencies
With shareholders as supportive as Alphabet's, why would a hyperscaler bother fiddling about with a hybrid instrument to raise synthetic equity? To the CFO, it would probably just feel like expensive debt.
Straggler of the flock
The most likely exception is Oracle. It is in a different ballpark, with a far smaller market cap and mid-triple-B ratings on negative outlook. Oracle has already wisely decided to raise half this year's capex — $25bn — as equity.
If it feels the need to splash out again next year, would it have the nerve to call on shareholders again? Yet issuing a ton more senior debt might worry the rating agencies, and at this level of the ratings spectrum, that has a cost in credit spreads.
In this context, a hybrid issue could look like a viable option.
This speculation connects with a wider debate — how big can hybrids get, and is it a separate market from corporate bonds?
Oracle is small enough that its capex needs are closer in size to more normal industrial companies like telecoms group Verizon (market cap $192bn) or power company NextEra Energy ($177bn). Those two have led the way recently in US issuance of hybrids.
Another important deal was that by Stellantis, the Italian-US car maker, on March 10. This was 10 days into the US-Israeli war on Iran, Stellantis had never issued a hybrid before, and it had just been downgraded by Moody's and S&P to Baa3/BBB- after it took a €22bn hit on changing its electric vehicle strategy.
Yet it raised €5bn of Ba2/BB rated debt, one of the biggest hybrid issues ever, and was showered with €19bn of demand.
Different level
Stellantis's transaction showed the hybrid market was capable of more than it had ever demonstrated before. This deal is getting towards the scale at which hybrids could make a difference to an issuer like Oracle.
But it is still a long way from anything that would be useful to the likes of Alphabet, Amazon or Meta.
It is dangerous to say 'never' in capital markets. Verizon staggered observers and redefined what was possible in corporate bonds with its $49bn acquisition financing in September 2013. It was almost three times as big as the previous largest corporate bond, a $17bn issue by Apple, but investors loved it.
If the right issuer needed it, the same could happen with hybrids. A gaggle of investment banks would love to try.
But if such a deal came, all pretence would have to be abandoned that corporate hybrid capital is somehow a separate market from senior and high yield bonds.
Specialist hybrid funds would be swamped by that quantity of paper. The vast majority would have to come from the same regular corporate bond investors that buy the issuer's senior debt. Hybrids would be a way to pay them a bit more to take some extra risk.
Investors are likely to be enthusiastic — especially in these days of megadeals of all kinds. Supply creates its own demand, and one day, investors will be called upon. The question is when an issuer picks up the phone.
Hybrids are a way to deal with constraints. For now, the big hyperscalers do not seem to feel any. When shareholders turn the cold shoulder, hybrid bankers will polish up their pitchbooks.