Key Takeaways:
Rising financing costs are testing the appetite for AI infrastructure debt as Goldman Sachs Group Inc. markets a $5.4 billion package for a Blackstone Inc.-backed data center project tied to Microsoft Corp.
Goldman Sachs is gauging investor demand for roughly $5.4 billion in debt to fund a QTS data center development in Georgia that Microsoft has agreed to lease, according to people familiar with the matter. The financing includes about $4.9 billion in secured bonds and approximately $500 million in term loans.
The deal comes as investor sentiment toward AI infrastructure financing has deteriorated over the past several weeks, pushing up borrowing costs for new issuers. Concerns have mounted over whether the massive capital spending on data centers and graphics processing units will generate sufficient returns, particularly as hyperscalers including Microsoft, Amazon.com Inc. and Alphabet Inc. collectively plan to spend more than $200 billion on AI infrastructure this year.
"Investors are becoming more selective about AI-related debt after a wave of issuance that pushed yields higher," said a senior credit market strategist at a major Wall Street bank. "The market is starting to differentiate between projects with clear anchor tenants and speculative builds."
The QTS project, owned by Blackstone's data center platform, benefits from Microsoft's long-term lease commitment, which provides a predictable revenue stream to service the debt. QTS is one of the largest data center operators in the U.S., with more than 40 facilities across the country.
The financing structure — secured bonds backed by the physical assets and lease cash flows — is designed to offer investors a degree of protection that unsecured corporate debt does not. Even so, the widening spreads on recent data center bond issuances suggest the market is demanding a higher premium for AI infrastructure risk.
For Goldman Sachs, the deal represents a significant underwriting fee opportunity at a time when investment banking revenue has been under pressure. The bank has been one of the most active arrangers of data center financing, capitalizing on the surge in AI-related construction.
The broader implications extend beyond this single transaction. If financing costs continue to rise, it could slow the pace of data center development, particularly for speculative projects without committed tenants. That would be a headwind for companies like Digital Realty Trust Inc. and Equinix Inc., which rely on access to capital markets to fund expansion. Conversely, projects backed by hyperscaler leases — like the QTS-Microsoft deal — are likely to retain access to funding, albeit at higher rates.
Microsoft has committed to spending over $50 billion on AI infrastructure this fiscal year, with data center leases forming a growing portion of that total. The company's Azure cloud business is racing to add capacity to meet demand for its Copilot AI assistant and enterprise cloud workloads.
This article is for informational purposes only and does not constitute investment advice.
