The buildout behind artificial intelligence is increasingly being financed with debt. A new Financial Times analysis found that several OpenAI partners have borrowed roughly 96 billion dollars to fund data center and chip expansion, putting the financial side of AI infrastructure under sharper scrutiny.
Borrowing follows the AI infrastructure race
The report names major players like Oracle and Softbank, alongside specialized providers such as Coreweave. Together, they represent different parts of the same expansion push: more data center capacity, more chips, and more capital committed before the resulting revenue is fully visible.
That matters because AI infrastructure is not just a software story. The systems behind large-scale AI require physical facilities, computing hardware, and long-term financial commitments. When those commitments are funded through borrowing, the question becomes not only who can build fast enough, but who can carry the balance-sheet pressure that comes with building at that pace.
The Financial Times analysis, as described in the source article, places the debt linked to several OpenAI partners at roughly 96 billion dollars. That figure gives the expansion a clear financial frame: the AI boom is being supported by large obligations, not only by operating cash or existing infrastructure.
Coreweave shows the pressure on specialized providers
Coreweave stands out in the report because its liabilities and lease obligations far exceed its expected annual revenue of five billion dollars. That comparison does not, by itself, determine the company’s future. But it highlights the scale mismatch that can emerge when a provider expands aggressively to meet AI infrastructure demand.
Lease obligations are especially important in this context because they reflect commitments tied to capacity. For a business building around demand for AI computing, that capacity can be central to growth. At the same time, obligations must still be serviced, and expected revenue needs to support the financial structure around them.
The source article does not describe Coreweave’s contracts, margins, or customer mix. What it does show is the basic tension: specialized infrastructure providers may be positioned close to AI demand, yet they can also accumulate obligations that are large compared with expected annual revenue.
The borrowing extends beyond OpenAI partners
The debt wave is not limited to a small circle of companies. Bank of America says the five largest tech firms, including Amazon and Microsoft, have issued 121 billion dollars in new debt this year. According to the source article, that amount is about four times their usual annual average.
This broader figure is important because it suggests that AI infrastructure financing is part of a wider shift in tech capital strategy. The largest companies are also turning to debt markets at an unusually high level, even though the source article does not separate how much of that borrowing is directly tied to AI.
For readers following the AI economy, the key point is the pattern. Data center expansion and chip capacity are costly, and the companies linked to that expansion are increasingly using borrowing as one way to fund the next stage of growth.
- OpenAI partners: Several have taken on roughly 96 billion dollars in debt for data center and chip expansion.
- Coreweave: Its liabilities and lease obligations far exceed expected annual revenue of five billion dollars.
- Largest tech firms: Bank of America says the five largest, including Amazon and Microsoft, issued 121 billion dollars in new debt this year.
- Investor concern: Deutsche Bank reports rising credit default swap costs for companies like Oracle.
Lenders are watching the risk more closely
Deutsche Bank reports rising credit default swap costs for companies like Oracle. In plain terms, the source article presents that as a signal of growing concern among lenders and investors.
That concern does not mean the AI infrastructure buildout will stop. It does mean that financial markets are beginning to price the risk of large borrowing more visibly. When debt grows quickly, investors tend to pay closer attention to whether future revenue can justify present commitments.
The issue is not simply that companies are borrowing. Large technology companies and infrastructure providers often use debt as part of their financing mix. The more specific concern raised by the source article is scale: the amounts are large, the buildout is capital intensive, and some obligations appear substantial when compared with expected revenue.
What this means for the AI buildout
The AI infrastructure race is often discussed through products, models, and performance. This report shifts attention to the financial machinery underneath. Data centers and chips require upfront investment, and the companies trying to supply that demand are taking on major obligations to do so.
For OpenAI partners, the roughly 96 billion dollars in debt identified by the Financial Times analysis points to how much capital is being mobilized around AI infrastructure. For the broader technology sector, the 121 billion dollars in new debt issued this year by the five largest tech firms, as cited by Bank of America, shows that borrowing has become a significant part of the current expansion cycle.
The practical takeaway is straightforward: AI growth is not only a question of technical capability. It is also a test of financing, lender confidence, and the ability of infrastructure providers to turn expensive capacity into durable revenue.