The $1 Trillion Question

OpenAI’s ambition to lead the next era of artificial intelligence may come with a price tag that dwarfs even Silicon Valley’s most capital-intensive ventures. According to Citi Research, the ChatGPT developer would need to spend more than $1 trillion within the next five years to meet its computing capacity commitments under deals with Nvidia, Broadcom, and AMD. The company has pledged to deploy 26 gigawatts of compute power—an amount nearly equivalent to the electricity consumption of the entire state of New York during peak summer demand.

At an estimated $50 billion per gigawatt in hardware, energy infrastructure, and data center construction, Citi analyst Chris Danely projects OpenAI’s total capital expenditure at $1.3 trillion by 2030. That estimate doesn’t include CEO Sam Altman’s even bolder internal target of 250 gigawatts by 2033, which would cost over $12 trillion.

Funding Gap and Investor Anxiety

The scale of OpenAI’s spending ambitions has raised questions about its financial feasibility. Citi forecasts the company’s 2030 revenue at just $163 billion, a fraction of the capital needed to fund its expansion. That mismatch underscores growing investor anxiety over whether the current AI boom is sustainable—or merely inflating a market bubble reminiscent of the dot-com era.

Analysts say Wall Street’s exuberance over AI stocks has already propelled valuations to historic highs. Nvidia, the sector’s clear leader, has more than doubled in value this year. Yet the viability of AI-driven growth depends on the economics of scaling massive computing infrastructure—an equation that may not add up.

“Altman has the power to crash the global economy for a decade or take us all to the promised land,” said Bernstein analyst Stacy Rasgon, emphasizing the risk of overbuilding in a market where demand forecasts are uncertain.

The Power Problem: Can Infrastructure Keep Up?

Beyond financial constraints, OpenAI faces a growing logistical challenge—energy supply. The company’s commitments to massive “Stargate” data center projects in the U.S., Norway, and the UAE depend on power infrastructure that may not exist yet. The U.S. power grid, already strained by industrial demand and decarbonization efforts, may struggle to deliver enough capacity to sustain AI expansion at this scale.

OpenAI’s $300 billion partnership with Oracle and its $22 billion deal with Nvidia-backed CoreWeave highlight the capital-intensive nature of this AI build-out. Yet even these staggering sums represent only a fraction of what would be needed to meet Citi’s projections.

If OpenAI’s vision materializes, the chip industry stands to gain enormously. Nvidia could generate up to $500 billion in revenue, and Broadcom more than $100 billion, from AI infrastructure contracts. But without sufficient capital and power, these optimistic projections risk becoming theoretical.

What’s Next: Between Hype and Reality

The trillion-dollar spending projection marks a pivotal moment for AI investment strategy. While OpenAI’s ambitions align with global enthusiasm for generative AI, the company’s financial trajectory highlights the limits of exponential growth in a capital-constrained world.

Investors and policymakers will be watching closely in the coming year as OpenAI and its partners navigate infrastructure bottlenecks, regulatory scrutiny, and funding pressures. The AI revolution may well hinge not only on innovation—but on whether the world can afford the power and capital to sustain it.