OpenAI may need to find new sources of capital to fund its aggressive data center spending. And the clock is ticking.

The maker of ChatGPT will need $207B in new financing by 2030 to pay its data center rent bills, according to a report from HSBC. First reported by the Financial Times, the HSBC analysis raises questions about whether OpenAI will be able to meet more than a trillion dollars in spending it has already committed to in cloud computing contracts and data center deals.

“One unknown parameter is the flexibility that OpenAI may have to adjust its commitment vs effective demand or financial capacity,” HSBC analysts wrote.  “Capital injections, debt issuance, or higher revenue than in our model would help closing the funding gap.”

OpenAI has $1.8T in cumulative cloud and data center development deals already signed for a total of 36 gigawatts of computing capacity that is to be delivered in the years ahead, according to HSBC.

This includes a $250B deal with Microsoft and a $38B deal with Amazon, both for cloud services, that OpenAI has signed since the beginning of October to secure close to 4 GW of capacity.

OpenAI has also reportedly agreed to spend $300B on cloud services from Oracle by 2030. Also included in HSBC's calculation are agreements to pay for the computing hardware for OpenAI’s Stargate data center development partnership with Oracle and SoftBank. This includes its nonbinding commitment to purchase 10 GW of Nvidia products and an October deal to acquire AMD chips worth tens of billions of dollars.

As all these deals are executed, OpenAI’s cumulative data center rent will be $792B between now and 2030, according to HSBC. That total rises to $1.4T by 2033. Both figures are roughly in line with OpenAI’s own estimates.

While the study anticipates significant revenue growth for OpenAI, that new revenue is more than matched by these rising data center costs. Even when other funding sources are taken into account, such as a promised $100B investment from Nvidia, the firm’s liquidity will still be around $200B short of required data center spending, according to HSBC’s analysis.

Still, the report acknowledges that higher-than-expected revenue growth could eliminate this potential shortfall, and OpenAI has several fundraising levers it may be able to pull to make up the difference. And if it can’t, it may have some ability to back out of its data center commitments.

“Less capacity would always be better than a liquidity crisis,” the analysts wrote.