OpenAI, the company behind ChatGPT, is taking unprecedented steps to secure its position in the global artificial intelligence hardware race, ordering hundreds of billions of dollars’ worth of chips despite its revenues being a fraction of that amount.
According to industry reports, OpenAI has committed to purchasing 26 gigawatts of high-end data processors from major semiconductor manufacturers, including Nvidia, AMD, and Broadcom. The order — estimated at more than 10 million units — would require enough power to rival the output of 20 nuclear reactors.
Analysts say this scale of procurement underscores both the potential and the peril of OpenAI’s ambitions. “They will need hundreds of billions of dollars to meet these obligations,” said Gil Luria, managing director at D.A. Davidson. “It’s an extraordinary financial challenge for a company that isn’t yet profitable.”
Despite generating an estimated $13 billion in annual revenue, OpenAI continues to forecast multi-billion-dollar losses, with profitability not expected until 2029. The company has declined to comment on how it plans to finance its growing hardware needs, though co-founder Greg Brockman hinted in an interview that “creative financing” will be crucial to sustaining growth amid the soaring demand for AI capabilities.
Nvidia, AMD, and Broadcom have all avoided discussing their direct arrangements with OpenAI. However, reports suggest Nvidia is considering investing up to $100 billion in OpenAI over several years — a move that would allow the chipmaker to recoup its investment through a circular financing model, where OpenAI uses Nvidia’s funds to purchase Nvidia chips.
Meanwhile, AMD has proposed a different type of deal, offering OpenAI the option to acquire equity stakes in the chip manufacturer. Analysts interpret this as AMD’s effort to boost its standing in an AI chip market still dominated by Nvidia.
The enormous commitments have triggered debate over whether OpenAI is fueling another tech bubble similar to the late-1990s dot-com crash. Some investors fear that the company’s unprecedented spending, combined with uncertain returns, could destabilize the broader AI investment landscape.
However, others argue that unlike the speculative internet boom, today’s AI demand is tangible and rapidly expanding. Harvard Business School’s Josh Lerner noted that, “There is very real demand for AI in a way that feels materially different from the 1990s.”
Still, the financial risks remain enormous. Even if OpenAI sells equity at its reported $500 billion valuation, the proceeds would not be enough to fully cover its chip obligations, leaving the company likely to borrow against its hardware assets as collateral.
CFRA analyst Angelo Zino believes OpenAI’s growth trajectory and its 800 million ChatGPT users justify its partnership-heavy financing model. Yet, the uncertainty persists. As Lerner observed, “It’s a real dilemma — how to balance massive future potential with the speculative nature of such investment today.”
