OpenAI is building a new digital infrastructure, and the cost of this operation is beginning to dwarf the budgets of medium-sized countries. Recent reports suggest that Sam Altman’s company plans to spend a staggering $600 billion on computing capacity by 2030. This is strategically laying the groundwork for an IPO that could value the company at around a trillion dollars.
The year 2025 proved to be an operational breakthrough for the San Francisco-based giant. Revenues of $13 billion significantly beat forecasts, and cost discipline allowed it to close $8 billion in expenses – a billion below its target. These figures lay the foundation for an ongoing funding round in excess of $100 billion, with Nvidia emerging as a key player with a $30 billion investment. With a market valuation of up to $830 billion, OpenAI is emerging as the absolute leader in the private equity market.
But beneath the cloak of explosive growth lie structural challenges that investors are watching closely. Although the company expects its revenues to reach $280 billion by 2030, symmetrically distributed between the consumer and corporate sectors, profitability is being challenged. Inference (inference) costs, the ongoing maintenance of running models, quadrupled last year. The result is a drop in adjusted gross margin from 40% to 33%.
It is becoming clear to business leaders that the days of ‘cheap artificial intelligence’ are coming to an end. Altman himself has announced the need to invest $1.4 trillion to develop 30 gigawatts of energy resources. This is the scale that turns OpenAI from a software company into an energy and infrastructure giant. The success of this venture depends on whether revenue growth manages to outpace the models’ appetite for power and silicon.
