OpenAI chief executive Sam Altman has told business leaders that companies which put off adopting artificial intelligence will lose ground to competitors that move first, framing AI as a matter of competitiveness rather than an optional experiment.
In remarks reported in early April, Altman set out three priorities for executives. First, secure computing capacity now, because access to AI hardware is likely to be the scarcest resource over the next several years. Second, stop treating AI as a pilot project and start measuring what it actually delivers in productivity and output. Third, senior leaders should use AI tools in their own daily work, since, in his view, executives who do not use the technology themselves tend to underestimate what it can do.
OpenAI’s own build-out
The advice on compute reflects OpenAI’s own spending. The company has said its computing capacity roughly tripled over the past year to about 1.9 gigawatts, a measure of the electrical power its data centers draw, and that it plans to keep adding infrastructure on a near-continuous basis. OpenAI has separately said it has secured commitments for around 10 gigawatts of U.S. capacity under its Stargate program, ahead of an earlier 2029 target, according to DataCenterDynamics.
Those figures underline how much the AI race has become a contest over energy and physical infrastructure as well as software. A gigawatt is roughly the output of a large nuclear reactor, and the leading AI developers are now negotiating for power on a scale previously associated with heavy industry.
A forecast for 2028
Altman has also repeated a prediction he made earlier this year that by 2028 more of the world’s intellectual capacity could reside in data centers than outside them. He has paired that forecast with a warning that the labor market faces a difficult transition as AI systems take on more cognitive work, and that the concentration of such capability in a few companies raises questions about who controls it.
The claims are projections from the head of a company that benefits from the expectation of rapid AI growth, and they should be read in that light. Even so, Altman’s practical message to executives, secure capacity, measure results and use the tools, is consistent with what many enterprise technology advisers are recommending as AI moves from experimentation to everyday operations.
