The 11th Goldman Sachs Leveraged Finance and Credit Conference has concluded, delivering a clear central message: in the backdrop of a continued absence of traditional merger and acquisition activity, AI infrastructure is rapidly consuming every corner of the global credit market at an unprecedented pace.

Over 400 investment executives and 85 borrowers—ranging from American Airlines to Caesars Entertainment, and from Applied Digital to Cipher Digital—gathered, with the keyword 'Artificial Intelligence' dominating the proceedings.

In just the past two months, the U.S. high-yield bond market has absorbed over $20 billion in new supply. Apollo Global Management and Blackstone are arranging a record $36 billion private credit deal for Anthropic, while tech giants like Oracle, Google, Microsoft, and Meta have tapped the global bond markets for over €60 billion.

Beyond the conference walls, SpaceX is advancing its IPO roadshow with a valuation starting at $180 billion. This signal, intertwined with the presence of Caesars Entertainment's CEO negotiating a $5.7 billion buyout, sketches a new credit market landscape for an 'AI-first' era.

"The capital needs for data centers, power, and chips are enormous and this has permeated every market we are in," stated Miriam Wheeler, Goldman Sachs' Global Head of Leveraged Finance, in an interview during the conference. "For our capital solutions team, AI is likely the number one theme where we are spending the most time right now."

**The AI Frenzy: Center Stage in Credit Markets**

The dominant theme at Goldman Sachs' 11th Leveraged Finance and Credit Conference was artificial intelligence.

Despite lingering anxieties among executives about the slow recovery of traditional M&A, high interest rates, and potential energy price spikes from Middle East conflicts, the AI narrative was sufficient to keep the atmosphere buoyant.

The numbers presented at the conference underscore this fervor. The U.S. high-yield market alone absorbed over $20 billion in new issuance in the past two months, with a significant portion flowing to AI-related infrastructure projects.

Zooming out, the AI infrastructure sector has become the fastest-growing segment within the non-investment grade credit market. Data from Octus shows that by early May 2026, total committed and funded debt for AI infrastructure issuers exceeded $107 billion.

Globally, blue-chip companies are actively tapping overseas funding. Alphabet and Amazon have collectively raised over €60 billion this year through euro-denominated bonds, extending their financing reach from Tokyo to London.

With debt from acquisitions yet to materialize on a large scale, AI has become the 'hot money' filling the void. Morgan Stanley estimates that from 2025 to 2028, AI infrastructure build-out will require about $3 trillion, with nearly half—$1.5 trillion—needing external financing.

**A Landmark Deal: The $36 Billion 'Chip-Backed' Loan**

A signature event of the conference was the revelation that Apollo Global and Blackstone are arranging an approximately $36 billion debt financing package for AI developer Anthropic, potentially marking one of the largest private credit deals in history.

The structure of this financing is unusual. The bulk of the $36 billion package is reportedly intended to purchase custom Tensor Processing Unit (TPU) chips from Google, which Anthropic will then lease to power its Claude models.

Broadcom, Google's co-development partner on the TPUs, is providing payment guarantees for a large portion of the financing, effectively adding a credit 'anchor'.

This transaction is significant not only for its size but also for its structure, potentially marking a watershed moment for 'chip-backed' debt as a regular financing tool. It signals private credit stepping in to fill the void left by traditional banks in financing AI infrastructure.

The commercial logic is based on Anthropic's rapid expansion. The company recently announced a $6.5 billion equity raise, valuing it at $96.5 billion post-money. Its reported annualized revenue has surged past $30 billion.

**Bubble Concerns: 'Winner-Takes-All' and Pricing Divergence**

While the AI theme energized most participants, several top credit executives highlighted potential risks. Currently, most corporate bonds issued for AI facilities are priced at nearly identical spreads, but bankers widely predict a coming 'divergence mechanism'.

"You will get to an inflection point where companies that are lagging in execution will see that directly reflected in their cost of capital," said Chris Bonner, Goldman Sachs' Head of Americas Leveraged Finance.

Another emerging trend is the market shifting from monetizing the 'AI concept' to monetizing 'AI execution capability'. Projects lacking in data capacity, engineering deployment, or commercial viability may face sharply widening credit premiums.

**In AI's Shadow: The Absence of Traditional M&A**

Despite AI's dominance, the hope for a recovery in traditional M&A was a persistent undercurrent. While there have been notable debt deals like the Electronic Arts acquisition and the proposed Paramount-Skydance acquisition of Warner Bros. Discovery, a steady flow of M&A transactions remains elusive.

In this context, Caesars Entertainment's agreement to a $5.7 billion acquisition by Fertitta Entertainment became a hot topic. The company's CFO attended the conference for private meetings with institutional investors.

The proposed Paramount-Skydance acquisition of Warner Bros. Discovery also garnered significant attention, with Paramount launching related debt tender and exchange offers in mid-May.

Deal-makers are anxiously awaiting a genuine restart of traditional leveraged buyouts in the second half of 2026 to bolster the broader credit asset pool.

**Long-Term Risks: When AI Investment Returns Meet the 'Cost of Capital' Rule**

A more fundamental question raised repeatedly at the conference was whether the future cash flows supporting the current debt issuance can cover their rising cost of capital.

"The debt markets are certainly constructive from a size and pricing perspective... but we think this backdrop is unequivocally better for more deal activity," noted Miriam Wheeler.

However, Chris Bonner offered a starker warning: the AI credit party will eventually move away from a 'rising tide lifts all boats' dynamic. Companies lagging in execution will pay a higher cost of capital.

This concern is not unfounded. Oracle's $1.6 billion data center financing reportedly faced investor demands for a higher premium due to concerns over sufficient collateral, with the bonds ultimately pricing at a 7.5% coupon.

Meanwhile, hyperscalers are extending their debt reach into global markets via foreign currency bonds. Analysts warn that if AI stumbles, it "could trigger more volatility."

With borrowing costs remaining elevated and AI build-out cycles stretching for years, the challenge of covering interest payments with core business revenue looms over every financing deal. The possibility of AI infrastructure shifting from a 'credit darling' to a 'vulnerability' is beginning to surface in the collective consciousness of market participants.