OpenAI may raise additional capital to address a deepening compute crunch and meet surging demand for its artificial intelligence services.

The move signals that even record-breaking investment may not be enough to keep pace with the massive infrastructure requirements of generative AI. As the company scales, the race for computing power has become a primary bottleneck for growth and deployment.

Chief Financial Officer Sarah Friar discussed the funding possibilities during an interview on Bloomberg Television with host Caroline Hyde on May 15, 2026. Friar said the company might seek more money despite recently completing what she described as the largest private fundraising round ever.

"The compute crunch is real, and we need to secure additional compute capacity to meet the surging demand," Friar said [2].

The financial need stems from the aggressive scaling of hardware. OpenAI ended 2025 with 1.9 gigawatts of compute capacity [3]. This represents a 9.5-fold increase in capacity since 2023 [3].

According to Friar, the company has maintained an annual compute capacity increase of roughly three times [3]. This rapid expansion is necessary to support the processing requirements of newer models and a growing user base, but it requires constant infusions of liquidity to fund data center expansion, and chip procurement.

Friar said the company may raise more capital even after the previous record-setting round [1]. The pursuit of additional funding highlights the extreme costs associated with the current AI arms race, where the ability to iterate is tied directly to available hardware.

“The compute crunch is real, and we need to secure additional compute capacity to meet the surging demand.”

The admission that the 'largest private fundraising round ever' is insufficient suggests that the capital intensity of frontier AI development is scaling faster than traditional venture capital cycles can support. By focusing on the 'compute crunch,' OpenAI is signaling that its primary constraint is no longer just algorithmic breakthroughs, but the physical availability of power and silicon.