If the issuers in the hyperscaler sector persist and continue to be large issuers, the concern is similar to the equity market: Does it become too concentrated in the corporate credit space?
And the reality is that it should not be nearly as big of a concern as we’ve seen in equity indices.
AI Financing Beyond Hyperscaler Balance Sheets
The hyperscaler market has certainly been an important part of the issuance pattern for corporate credit in 2026, and it will continue to be so going forward.
We do think it will be evolving, though. It will be an evolving story where a lot of that issuance will be moving from specific hyperscaler corporate balance sheets and, as you can see on the chart that we’re showing here, expanding much more in 2027 and 2028 into data centers and chips and compute financing, which is required for overall AI financing needs.
So that is going to be a meaningful change. Ultimately, that should result, as you can see in the chart in 2027 and 2028, in actually decreasing financing from what we call traditional corporate hyperscalers, but pretty meaningful increasing in the financing of data centers and chips.
A lot of that issuance in data centers, and probably in the chips and compute, is going to be somewhat in the high-yield market, not focused as much on investment grade like we’ve seen recently.
Why Broader Corporate Credit Spreads Have Held Up
As the hyperscalers have done a lot of issuance, you’ve definitely seen different spread patterns of hyperscaler spreads relative to overall corporate spreads.
But you can see, for spreads overall on the chart, have been really pretty stable through this period, throughout the course of this year. You can see some of the spread widening that we saw in July from the hyperscaler sector.
The reality is other spreads have not been subject to this crowding out. There’s not been a concern around the fundamentals as it pertains to either banks or consumer-type companies.
And as long as the fundamentals to that part of the economy remain stable, which is what we expect, combined with a level of interest rate that maintains demand levels for investment-grade credit in particular, we think that can persist – that reasonably tight spreads away from hyperscalers should persist.
Why Concentration in Credit Differs from Equities
If the issuers in the hyperscaler sector persist and continue to be large issuers, the concern is similar to the equity market: Does it become too concentrated in the corporate credit space?
And the reality is that should not be nearly as big of a concern as we’ve seen in equity indices.
Historically, we have seen this phenomenon already in credit. We saw that in the banking sector. There are five or six money center banks – large banks in the U.S. They already account for 10 or so percent of the credit market in a very small number of credits. That’s the type of size that we would expect the hyperscalers to eventually end up being as a part of this market.
So we’ve already experienced it through the banking sector.
And then secondly, because of just the phenomenon of how prices change in bonds relative to equities, you shouldn’t have this market appreciation where the credits, just due to their market capitalization, take over an index like we saw in the equity market.
Decision Points is a new video series that looks at the data driving fixed income markets. Bringing together views across Neuberger’s Fixed Income Team, these perspectives offer a closer look at the opportunity set across public and private markets.
In episode 2, David Brown, Global Co-Head of Investment Grade and Multi-Sector, looks at the next phase of AI financing. As borrowing extends beyond the hyperscalers, what could change for corporate credit investors?
Three key takeaways
- We expect AI financing to shift beyond hyperscaler balance sheets in 2027 and 2028—toward data centers and chips—with some borrowing coming through high yield.
- Heavy hyperscaler issuance hasn’t pushed spreads wider across the rest of corporate credit. We see that resilience continuing if fundamentals and demand hold up.
- Concentration should be less concerning in credit than in equities. Large banks are an existing precedent demonstrating that bond-price appreciation shouldn’t cause hyperscalers to take over an index
For investors, the next phase of AI financing could bring a broader mix of borrowers, including more high-yield issuers. So far, heavy borrowing by hyperscalers hasn’t pushed spreads wider across the rest of corporate credit.
