HSBC CEO urges scrutiny of AI data-center tenants and contract prices

AI can transform a bank without making every data-center project worth financing. HSBC CEO Georges Elhedery drew that distinction in an October 20 interview with CNBC’s Dan Murphy, urging investors to examine tenancy agreements and contract prices. His warning was about choosing individual deals, not rejecting the technology HSBC itself is working to adopt.

A financier that also runs data centers

HSBC has a stake on both sides of the infrastructure market: it finances data centers and operates them. These facilities house the servers that support AI and cloud computing, placing the bank’s comments at the intersection of its lending business and its own technology needs.

The London-headquartered lender had close to $3 trillion in assets and ranked as Europe’s biggest bank in S&P Global’s 2025 ranking, CNBC noted. Its 2025 annual report identified data centers and AI among the sectors represented in its customer base, alongside an ambition to lead in fast-growing transition ecosystems.

Elhedery’s financing test focused on the agreements behind a project, rather than a blanket judgment about the sector. He said some contracts were harder to support than others. The distinction leaves room for investment while rejecting the idea that all projects deserve the same backing.

You have to look at the quality of the tenancy agreements. You have to look at the pricing of these contracts. Not all of them are of that level of virtuosity that you can support strongly. Some of them may be more challenging.

Georges Elhedery, HSBC CEO, speaking to CNBC on October 20, 2026

Enthusiasm inside the bank

His assessment of AI’s role within HSBC was markedly enthusiastic. Elhedery described it as a major means of redesigning the bank’s businesses and activities, saying the technology would help re-engineer the institution. He said HSBC was already on a journey to adopt it.

He also argued that AI would give employees more power through higher productivity. That was his expectation for the technology’s business value, rather than a measured productivity result presented in the interview. His optimism about adoption sat alongside, rather than displaced, his caution about the terms of infrastructure investments.

A larger contest for financing

The scale of the proposed buildout explains the financial stakes. McKinsey estimates global data-center spending could reach $7 trillion by 2030. Large technology companies are increasingly using private credit and other debt to develop facilities, CNBC reported, bringing outside financing into the race to add computing capacity.

Earlier in October, Allianz chief economic adviser Mohamed El-Erian described competition between government borrowers and AI-related companies as a source of pressure in bond markets. Speaking at the LSEG and Eurasia Group GZERO Summit in London, he said governments were both borrowing and refinancing existing debt at higher rates.

El-Erian said hyperscalers—the largest cloud operators—had raised $200 billion through bonds in 2025 and would probably raise $1 trillion in 2026. He described those borrowers as insensitive to financing yields and spreads because they were betting on a transformative technology. The 2026 figure was his projection, not a completed borrowing total.

JPMorgan Chase separately estimated earlier in 2026 that AI-related debt issuance would reach $4.1 trillion by the decade’s end. That forecast covers borrowing by data-center and other AI-adjacent companies. It measures debt issuance, unlike McKinsey’s spending forecast; the two figures are not separate costs to add together.