HSBC Global Investment Research has released a report updating its projections on OpenAI's computing power demands and cloud leasing costs, estimating that the AI unicorn may require up to $207 billion in new financing by 2030. The research team, led by Nicolas Cote-Colisson, noted that OpenAI's recent expansion of long-term procurement agreements with cloud providers has significantly increased its funding needs.
According to HSBC's analysis, OpenAI has signed a $250 billion cloud computing procurement deal with Microsoft (MSFT.US) and a separate $38 billion, seven-year cloud services contract with Amazon (AMZN.US). Meanwhile, competitor Anthropic (backed by Amazon and Google) is accelerating its expansion, securing a supply of 1 million TPU AI chips from Google (GOOG.US, GOOGL.US) worth "tens of billions of dollars" and planning a $50 billion infrastructure investment. Additionally, Anthropic has entered a $30 billion computing power commitment agreement with Microsoft and Nvidia (NVDA.US), with expectations of receiving up to $15 billion in funding.
Cote-Colisson's team reassessed OpenAI's computing power expansion path and leasing costs, concluding that the company will still need an additional $207 billion by 2030. Whether this funding gap can be narrowed depends on several variables, including OpenAI's ability to adjust procurement commitments based on actual demand and whether it can secure equity investments, debt financing, or significantly higher-than-expected revenue to alleviate financial pressure.
The report highlights that a new AI supercycle is emerging, with OpenAI projected to bear $1.4 trillion in computing costs over the next eight years—raising investor concerns about return on investment. Compared to its estimated 2025 revenue of approximately $12.5 billion, the vast disparity between costs and revenue also impacts risk assessments across the AI supply chain.
However, analysts point out that Microsoft's CEO has stated that the simultaneous commercialization efforts of AI model companies, infrastructure providers, and chip manufacturers will help enterprise clients realize AI value faster. Therefore, even without considering AGI (Artificial General Intelligence), AI's potential for productivity gains remains substantial.
Among the companies covered by HSBC, those most vulnerable to OpenAI's success or failure include Oracle (ORCL.US), Microsoft, Amazon, Nvidia, AMD (AMD.US), and SoftBank, which has heavily invested in AI.