A BrAIve New World for Credit

lesbarclays1 pts0 comments

A BrAIve New World for Credit - by Les Barclays

This Is Not Investment Advice

SubscribeSign in

A BrAIve New World for Credit<br>Credit moves from passenger to driver<br>Les Barclays<br>Aug 12, 2026

20

Share

Welcome back!<br>Credit markets are playing a more central role in financing AI since we’ve entered the “show me the money” phase of the AI theme and it’s reshaping credit markets. The hyperscalers are tapping public and private debt markets to plug the gap for funding AI infrastructure as cash flows dip into negative territory. Borrowers have optionality vis-à-vis financing as lenders race to fund the AI buildout - so much so it’s the driving force behind the rise in global corporate bond issuance, which has reached $3.68tn in H1 2026.1<br>Amanda Lynam at Goldman Sachs noted that 2026 AI-linked debt issuance stands at $489bn, which is ~51% above their own $322bn FY25 forecast, and 2026 isn’t finished. Even Morgan Stanley’s $585bn call for 2026 looks conservative.<br>Roughly 40% (~$234bn) of this year’s AI debt supply is coming from the hyperscalers.

GPU ABS - an investable asset class! - can maintain its value if (1) GPUs are ring-fenced in an SPV, (2) they have high utilisation rates, (3) there’s contractual cash flows, (4) insurance wraps [known as RVGs] are included and (5) there are dynamic LTV adjustments. Returns for GPU-backed loans are in the low teens, so it’s no surprise some PC funds want to get involved. If GPU residual risk is too large (loan value drops below the outstanding loan balance), loan prices will collapse.<br>Dylan Patel points out that GPU borrowers carry credit risk but there’s still some pie-in-the-sky thinking in that article. Equities can get drunk on narratives, but credit will force people to sober up. In credit the upside is capped at principal + coupon and there’s disproportionate downside. You can’t make interest payments with potential and promise.<br>As hyperscalers flood IG benchmarks, credit curves could steepen if supply keeps outpacing demand for longer maturities. Within tech, expect divergence in spreads between the winners and everyone else.<br>AI’s debt wave shares similarities with the post-GFC bank recap era when financials flooded IG indices. Heavy issuance widened spreads but also created opportunities for anyone willing to roll up their sleeves. I avoid lazy ‘08 analogies, but the parallels with AI concentration in credit are hard to ignore.

More runway for hyperscalers? Not quite | Apollo<br>Since the hyperscalers are rated AA- on average, the AI debt wave has been almost entirely an IG story. The tech sector alone accounts for more than 10% of the index for IG & leveraged loans, and we’re 8 months in. Tech could hit 20% of U.S. IG within two years.<br>Spreads have surprisingly stayed tighter than expected this year. That was until spreads on some hyperscaler debt widened by around 15 bps as the market absorbed $75bn of unexpected supply in June and July. Hyperscalers currently have below 1x net leverage and although some issuers could borrow substantially more while remaining below 2x net leverage, doing so could result in multiple-notch downgrades, increasing financing costs, and reducing demand from ratings-sensitive investors.<br>The recent weakness in hyperscaler bonds actually matters for fund managers, since IG corporates are a snooze fest. Even small price moves can swing relative performance. Hyperscalers are so deep in the AI arms race that they can’t afford to slow down, that partially explains why they’re tapping the bond market for tens of billions irrespective of market conditions.<br>AI’s appetite for capital is massive - they're in a game of capex chicken as no one wants to be the first to blink. Hyperscalers are expected to spend over $5tn by 2030 on tech and data centres, according to Goldman. Private markets will step in to help fill the gap.<br>The depth and breadth of the private financing ecosystem have featured heavily in the AI-related investment cycle. “We expect the ‘real assets’ categories of private infrastructure and real estate will play an even larger role in the years ahead,” Amanda Lynam, chief credit strategist at Goldman Sachs Research, writes in a report.

Private infrastructure funds raised a record $221bn in 2025, and the average fund size jumped to $1.8bn. The fundraising has been concentrated, with experienced managers acquiring the majority of the capital. Infrastructure funds returned 12.8% last year, outperforming all other private market categories except for private equity and venture capital, Lynam writes.<br>The scale of AI financing means borrowers are tapping everything from IG bonds to private credit, across every structure and currency they can grab.<br>With their massive AI capex, hyperscalers already account for a big chunk of new corporate issuance. They could soon hit market saturation as public debt investors run into concentration limits and index weights for these names keep climbing.<br>“We expect liquid credit market saturation and...

credit hyperscalers private debt market financing

Related Articles