Goldman Sachs Summit Spotlights AI Computing Infrastructure Mega-Build as Trillion-Dollar Funding Gap Sparks Financing Wave
Alina Collins
At Goldman Sachs' annual leveraged-finance conference, AI data-center and power-infrastructure financing dominated the agenda — AI-linked junk-bond issuance has topped $20 billion in two months, while the absence of traditional M&A deal flow underscored a massive capital rotation toward AI buildout.
What signal did this conference send?
Goldman's 11th annual Leveraged Finance & Credit Conference drew over 400 investment executives and 85 borrowers to Dana Point, California.
Attendees spanned legacy industries and AI newcomers: American Airlines and Caesars Entertainment alongside Applied Digital and Cipher Digital.
This means → the leveraged-finance market's attention has tilted sharply from traditional corporate debt to AI infrastructure — not a one-company story, but an entire funding chain redirecting.
How big is the AI financing wave?
Over the past two months, AI-linked companies raised more than $20 billion in the U.S. junk-bond market.
Apollo Global Management and Blackstone are rallying investors for a $36 billion financing deal to back Anthropic's AI infrastructure buildout; Anthropic separately announced it has confidentially filed a draft IPO registration.
Miriam Wheeler, Goldman's global head of leveraged finance, said capital-expenditure demand across data centers, power, and chips "touches virtually every market we are involved in."
In plain terms = Wall Street's largest institutions are treating AI infrastructure as the defining financing opportunity of this era — the scale has moved well past pilot stage into main-battlefield territory.
Where is the risk once the money floods in?
Chris Bonner, Goldman's head of Americas leveraged finance, warned that most AI-facility bonds currently trade at nearly identical prices — no differentiation yet.
As supply saturates, a sorting mechanism will kick in: borrowers that fail to deliver data centers on schedule will see their bond pricing start to diverge.
This means → the market's current "treat-all-AI-the-same" pricing will not last; companies that fall behind on execution will pay the price in higher capital costs.
Why hasn't traditional M&A come back?
Despite the AI frenzy, Wall Street's appetite for traditional M&A debt remains strong. Bonner opened the conference bluntly: "We all want to see more issuance."
Marquee deals — the Electronic Arts acquisition, Paramount Skydance's planned purchase of Warner Bros. Discovery — have added bond supply, but a steady deal pipeline has not yet formed.
Caesars Entertainment agreed to a $5.7 billion takeover by Fertitta Entertainment, the most talked-about non-AI deal at the conference, co-led by Goldman and Morgan Stanley.
This reflects a traditional M&A market that is not short on demand but still in a "sporadic-landing" phase, with capital and attention heavily siphoned by AI buildout.
What does the broader debt-market backdrop look like?
Wheeler struck a cautiously optimistic tone: the debt market is constructive in both size and pricing.
She acknowledged rate volatility over the past week but argued the current backdrop still favors more deal activity.
In plain terms = money is not scarce and rates have not shut the financing door — what will determine the next leg is whether AI projects deliver on time and when traditional M&A volume picks up.
Content is for reference only, not financial advice.