OpenAI plans to spend $50 billion on computing power this year alone, a figure that dwarfs the annual capital expenditure of most Fortune 500 companies and underscores the staggering financial appetite of the race to build artificial general intelligence. The disclosure, made under oath by OpenAI President and co-founder Greg Brockman during testimony in the company's high-profile legal battle with Elon Musk, puts a hard number on the infrastructure arms race that is reshaping the technology industry.
Brockman revealed the spending projection during his second day of testimony in the closely watched trial, where Musk is seeking $150 billion in damages and leadership changes, alleging OpenAI abandoned its original charitable mission. The figure represents an exponential escalation from the roughly $30 million the company spent on compute in 2017, a nearly 1,700-fold increase in less than a decade. OpenAI's internal projections point to a cumulative compute spend of approximately $600 billion by 2030, a trajectory that would make it one of the most capital-intensive enterprises in history.
The Scale of the Bet
To grasp the magnitude, consider that $50 billion in a single year exceeds the entire annual research and development budgets of pharmaceutical giants like Pfizer and Roche combined. It is roughly equivalent to the GDP of Luxembourg. The spending reflects the enormous computational demands of training and serving increasingly large AI models to a rapidly expanding user base, with ChatGPT now commanding hundreds of millions of weekly active users.
Much of the financing comes with strings attached. Reports indicate that multibillion-dollar investments from partners like Amazon and Nvidia are structured as conditional discounts, effectively requiring OpenAI to lease massive compute capacity from those same companies. The arrangement blurs the line between investment and vendor lock-in, raising questions about how much of the $50 billion represents genuine spending power versus pre-committed obligations.
Brockman also disclosed during testimony that his personal stake in OpenAI is now worth nearly $30 billion, a figure that places him among the 100 wealthiest people on Earth and illustrates just how much value has accrued inside the company since its founding as a nonprofit in 2015. The company is also actively exploring an initial public offering, which would mark one of the most anticipated tech listings in years.
Washington Takes Notice
The spending disclosure arrives at a moment when AI infrastructure has become a central concern in Washington. In a separate appearance before the Senate Commerce Committee at a hearing titled 'Winning the AI Race: Strengthening U.S. Capabilities in Computing and Innovation,' OpenAI CEO Sam Altman urged lawmakers to support massive investment in domestic AI infrastructure. Altman told senators that 'infrastructure is destiny, and we need a lot more of it,' framing the buildout as essential to maintaining American competitiveness against China and other rivals.
Altman pointed to the Stargate Project, a $500 billion initiative announced earlier this year alongside Oracle and SoftBank with the backing of the Trump administration, as evidence of the scale required. He argued that 'the cost of AI will converge to the cost of energy,' a formulation that links the future of artificial intelligence directly to the nation's power grid and energy policy. The hearing also featured testimony from Microsoft President Brad Smith, AMD CEO Lisa Su, and CoreWeave CEO Michael Intrator, all of whom pressed lawmakers to streamline federal permitting for electrical infrastructure and resist heavy-handed regulation.
A Capital Arms Race With No Ceiling in Sight
OpenAI is far from alone in its infrastructure ambitions. Microsoft has committed over $80 billion in capital expenditure for fiscal year 2025, much of it directed at AI data centers. Google parent Alphabet disclosed $75 billion in planned capital spending. Amazon Web Services continues to expand its GPU fleet at a pace executives describe as unprecedented. Together, the hyperscalers and their AI partners are on track to pour more than $300 billion into infrastructure this year, a figure that would have seemed fantastical even two years ago.
The question is whether the returns will justify the outlay. OpenAI is projected to generate roughly $12 to $13 billion in revenue this year, meaning it is spending approximately four dollars on compute for every dollar it earns. That ratio is sustainable only as long as investors believe the technology will eventually produce profits at a scale commensurate with the investment, a bet that rests on the assumption that AI will become as foundational as the internet or electricity.
What to Watch Next
Three developments will determine whether this infrastructure bet pays off. First, the outcome of the Musk trial could reshape OpenAI's corporate structure and governance, potentially complicating its ability to raise future capital. Second, the pace at which AI revenue grows relative to compute costs will signal whether the economics of large-scale AI are improving or deteriorating. Third, Washington's willingness to fast-track energy and permitting reform will determine how quickly companies like OpenAI can actually build the physical infrastructure their spending plans require. For now, the $50 billion figure stands as the clearest signal yet that the AI industry has entered a phase of capital deployment unlike anything the technology sector has ever seen.
"Infrastructure is destiny, and we need a lot more of it."— Sam Altman, CEO, OpenAI