OpenAI has turned down some opportunities this year because it doesn’t have enough computing power to support them, according to Sarah Friar, OpenAI’s chief financial officer.
“We have some really tough deals going on right now that we’re not pursuing because we don’t have enough computing power,” Friar told ARK Invest CEO Cathie Wood in an interview published this week.
Frier said this issue will become especially acute in 2026 as demand for AI continues to soar globally.
“Right now, I’m spending a lot of time trying to find the last bit of computing that will be available here in 2026,” she said.
OpenAI President Greg Brockman echoed that pressure in an interview with the Big Tech Podcast published Wednesday, saying the company is struggling to meet demand.
Their comments highlight the growing constraints across the AI industry. Even the most advanced companies have limited access to the computing power needed to train and run models.
“If you don’t have [compute]I have no income. That’s one thing I know for sure,” the monk said.
Computing limitations create trade-offs
This shortfall is forcing OpenAI to make strategic tradeoffs.
Brockman said the company is prioritizing a few core use cases, such as personal AI assistants and tools that can solve complex tasks, because “it’s impossible to access everything” given current computing limitations.
That dynamic is already shaping product decisions. OpenAI has pulled back from several initiatives, including discontinuing its video app Sora, to focus resources on its core revenue-generating AI products.
The company currently serves about 900 million consumers and more than 1 million businesses and has raised significant funding, Frier said. The company recently closed a $122 billion funding round aimed at securing future computing power.
“We can’t build compute fast enough to keep up with demand,” Brockman said, explaining what he called “very tough decisions” about what to launch and where to allocate resources.
OpenAI has a “multi-year commitment” to secure future capacity, Frier said.
Other AI companies appear to be facing similar constraints.
Anthropic recently tightened its limits on Claude model usage during peak hours, a sign that even major model makers are struggling to keep up with the surge in demand.
For now, basic constraints remain. Even in the age of AI, you can’t scale without the hardware behind it.