Stay connected with BizTech Community—follow us on Instagram and Facebook for the latest news and reviews delivered straight to you.
A consortium of major banks is in advanced discussions to provide up to $38 billion (£30 billion) in loans to fund a network of new data centres linked to OpenAI, the creator of ChatGPT. The financing, which could rank among the largest corporate debt deals in recent history, underscores the escalating costs of building the computational backbone for artificial intelligence, as tech firms race to meet surging demand for advanced models.
The proposed loans, structured as project financing backed by future revenues from AI services, involve lenders including JPMorgan Chase, Goldman Sachs and Bank of America. They aim to support OpenAI’s expansion plans, which include gigawatt-scale facilities in the US and Europe to power training and deployment of next-generation systems like GPT-5.1. Sources familiar with the talks said the deal could close by mid-2026, contingent on regulatory approvals and energy supply agreements.
OpenAI, valued at over $150 billion following its latest funding round, has been grappling with capacity constraints as user numbers exceed 800 million weekly. The company has already secured multi-billion-dollar commitments from Microsoft for Azure cloud resources, but the new debt would enable diversification and faster scaling. “This is about securing the infrastructure for the AI economy,” one banker involved told the FT, adding that the loans carry interest rates of 6-8 per cent, reflecting the high-risk, high-reward nature of the sector.
Financing the AI Boom
The deal highlights a broader trend of debt financing flooding into AI infrastructure, with global data centre investments projected to reach $500 billion by 2030, per McKinsey estimates. Traditional equity raises have dominated early-stage AI funding, but as projects mature, banks are stepping in to offer leveraged options, often secured against hardware assets or long-term contracts with hyperscalers like Amazon Web Services.
Read also this: https://biztechcommunity.com/news/miami-mayor-francis-suarezs-bitcoin-salary-yields-300-gains/
For OpenAI, the funds would address acute bottlenecks: Training a single frontier model can consume energy equivalent to a small city, and delays in chip deliveries from Nvidia have hampered progress. The proposed sites, located in low-cost energy regions like Texas and Virginia, incorporate sustainable features such as nuclear micro-reactors and advanced cooling systems to mitigate environmental concerns.
However, the scale of the borrowing raises eyebrows. Critics warn of over-leverage in a nascent market, where monetisation remains uncertain—OpenAI reported $3.7 billion in revenue for 2025 so far, but losses topped $5 billion due to compute expenses. “It’s a bet on AI’s trillion-dollar future, but one wrong turn could trigger defaults,” said Daniel Ives, an analyst at Wedbush Securities.
Regulatory and Market Reactions
US regulators, including the Federal Reserve, are scrutinising such mega-deals for systemic risks, particularly given the concentration of lending among a handful of Wall Street firms. In Europe, where data privacy laws are stringent, similar projects face hurdles under the AI Act, potentially delaying cross-border expansions.
The talks come amid a flurry of AI financing announcements, including Meta’s $70 billion capex guidance and Google’s $155 billion cloud backlog. As the sector’s energy demands—forecast to rival Japan’s total consumption by 2027—clash with global decarbonisation goals, sustainable lending terms will be key.
For OpenAI and its backers, this could unlock unprecedented scale; for the financial system, it’s a high-stakes test of AI’s economic viability.
