NextFin News - OpenAI is targeting approximately $600 billion in total compute spending through 2030, according to a source familiar with the matter on February 20, 2026. This massive capital projection comes as the San Francisco-based AI pioneer, led by Sam Altman, prepares for a potential initial public offering (IPO) that could value the firm at up to $1 trillion. The disclosure follows a robust fiscal 2025, where OpenAI reported $13 billion in revenue—exceeding its $10 billion forecast—while maintaining disciplined spending of $8 billion against a $9 billion budget. According to The Business Times, the company is currently in the final stages of a fundraising round exceeding $100 billion, which includes a pivotal $30 billion investment from Nvidia. This round is expected to value OpenAI at roughly $830 billion, marking one of the largest private capital raises in corporate history.
The scale of this expenditure reflects a fundamental shift in the artificial intelligence industry from software innovation to heavy infrastructure development. Altman has previously committed to a broader vision of spending $1.4 trillion to secure 30 gigawatts of computing resources, a capacity equivalent to powering 25 million American homes. This infrastructure push is being executed against a backdrop of shifting domestic policy. Since the inauguration of U.S. President Trump on January 20, 2025, the administration has emphasized "American AI Supremacy," favoring massive domestic energy and data center deregulation. This political environment has emboldened OpenAI to project total revenues exceeding $280 billion by 2030, split almost equally between its consumer and enterprise divisions.
However, the financial path forward is not without significant headwinds. While revenue is surging, the cost of maintaining and operating these models—known as inference—is growing at an even faster rate. According to The Information, OpenAI's inference expenses quadrupled in 2025 alone. This surge in operational costs has led to a contraction in adjusted gross margins, which fell to 33% in 2025 from 40% the previous year. The data suggests that as AI models become more ubiquitous and integrated into enterprise workflows, the "cost of serving" is becoming a primary bottleneck for profitability. This margin compression highlights the "compute trap": the more successful an AI model becomes, the more capital must be reinvested into the hardware required to sustain its usage.
The $600 billion figure represents more than just a budget; it is a strategic moat. By locking in such vast amounts of compute power, OpenAI is attempting to outpace competitors who may lack the capital access or the deep-tier partnerships with hardware giants like Nvidia. The projected $30 billion stake from Nvidia is particularly telling, as it creates a circular ecosystem where the world’s leading chipmaker becomes a primary stakeholder in its largest customer. This vertical integration of capital and silicon is likely to define the next five years of the industry, as the barrier to entry for "frontier models" moves from the billions into the hundreds of billions of dollars.
Looking ahead, the success of OpenAI’s $1 trillion IPO ambition will depend on its ability to stabilize margins through hardware efficiency and sovereign energy solutions. The Trump administration’s focus on domestic energy production may provide the necessary power grid expansion to support Altman’s 30-gigawatt goal, but the sheer volume of capital required will necessitate a transition from private venture funding to public equity markets. As OpenAI moves toward 2030, the industry will watch whether the company can convert its $600 billion infrastructure bet into a sustainable, high-margin software business, or if it will evolve into a new breed of AI-centric utility provider, defined by massive scale and capital-intensive operations.
Explore more exclusive insights at nextfin.ai.
