Investment bank projects $207 billion shortfall as infrastructure costs dwarf revenue projections
KEY POINTS:
- HSBC Global Investment Research projects OpenAI will need $207 billion in new financing by 2030 to sustain its expansion plans, according to Fortune and The Register.
- The company’s cloud and data center costs are projected to reach $792 billion through 2030, with a $620 billion data center rental bill alone, per HSBC’s analysis reported by Fortune.
- Despite projected revenues exceeding $213 billion by 2030, OpenAI’s cumulative free cash flow will remain negative, according to the bank’s assessment.
The investment bank HSBC has delivered a stark assessment of OpenAI’s financial trajectory, projecting that the company behind ChatGPT will remain unprofitable through 2030 despite its explosive user growth—and will require at least $207 billion in additional financing to sustain its ambitious infrastructure expansion.
According to Fortune, HSBC’s semiconductor analyst team, led by Nicolas Cote-Colisson, updated its OpenAI forecasts following recent multiyear cloud computing commitments, including a $250 billion agreement with Microsoft announced in October and a $38 billion deal with Amazon disclosed in November. Fortune reported that these deals came without any new capital injection and represent the latest in a series of capacity expansions targeting 36 gigawatts of AI compute power by decade’s end.
The Financial Times’ Alphaville blog, which first reported on HSBC’s analysis, described OpenAI as a “money pit with a website on top,” according to Fortune. Diginomica, citing the analysis, reported that HSBC characterized the company as a “burning platform” requiring $207 billion just to continue losing money through 2030.
HSBC projects OpenAI’s cumulative free cash flow by 2030 will remain negative, Fortune reported. The bank’s analysts model the company’s cloud and AI infrastructure costs at $792 billion between late 2025 and 2030, with total compute commitments reaching $1.4 trillion by 2033. Fortune noted that HSBC referenced OpenAI CEO Sam Altman’s publicly stated plan for $1.4 trillion in compute expenditure over eight years.
The funding gap emerged after OpenAI committed approximately $588 billion across three major cloud providers, according to The Register. These include the $300 billion Oracle deal announced as part of the Stargate project, the $250 billion Microsoft Azure commitment, and the $38 billion AWS arrangement.
OpenAI announced the Microsoft partnership on October 28, 2025, with the company contracting to purchase $250 billion in Azure services while Microsoft retained a stake valued at approximately $135 billion, representing roughly 27 percent ownership, according to OpenAI’s official announcement. Microsoft confirmed it would no longer hold right of first refusal to be OpenAI’s exclusive compute provider.
The AWS deal, announced November 3, 2025, marked OpenAI’s first major contract with the cloud market leader, according to CNBC. Under the seven-year agreement, OpenAI will access hundreds of thousands of Nvidia graphics processing units, with all capacity targeted for deployment before the end of 2026.
Despite the daunting infrastructure costs, HSBC’s projections assume significant user growth. The Register reported that HSBC predicts ChatGPT’s consumer products will attract 3 billion regular users by 2030, up from the 800 million weekly active users that OpenAI CEO Sam Altman announced at Dev Day in October, according to TechCrunch.
The bank’s revenue projections exceed $213 billion by 2030, Fortune reported, based on assumptions of higher subscription conversion rates—10 percent versus the current 5 percent—and an assumption that large language model providers will capture a portion of digital advertising revenue.
Fortune reported that HSBC outlined several options to close the funding gap: dramatically increasing the proportion of paid subscribers from 10 to 20 percent could add $194 billion in revenue; capturing a larger share of digital advertising; or extracting extraordinary efficiencies from compute operations. Even under optimistic scenarios, however, the company would still require fresh capital beyond 2030, according to the analysis.
The bank noted potential challenges with debt financing. Fortune reported HSBC described raising additional debt as “possibly the most challenging avenue in the current market conditions,” citing Oracle and Meta’s recent significant debt issuances to finance AI-related capital expenditure, which have raised market concerns about general AI financing.
The stakes extend beyond OpenAI itself. The Register, citing the HSBC research paper, reported that the most exposed partners to OpenAI’s success or failure include Oracle, Microsoft, Amazon, Nvidia, AMD, and SoftBank, which holds an 11 percent stake in OpenAI.
Oracle has already experienced volatility following its Stargate commitments, The Register reported. After announcing its deal with OpenAI in September, Oracle’s share price initially surged 30 percent, briefly making co-founder Larry Ellison the world’s richest person. The company has since relinquished those gains.
The analysis arrives as questions mount about AI’s broader economic returns. Fortune noted that HSBC referenced Nobel laureate Robert Solow’s observation that “you can see the computer age everywhere but in productivity statistics,” with the bank noting that weak productivity gains remain “an unfortunate characteristic of today’s developed economies.”
Harvard economist Jason Furman calculated that without data centers, GDP growth would have been just 0.1 percent for the first half of 2025, Fortune reported—raising questions about how long growth can be sustained on anticipated future AI returns that remain unproven.
