OpenAI's CFO says the company is passing on opportunities because it does not have enough compute

OpenAI's Chief Financial Officer, Sarah Friar, has revealed that the company is facing significant challenges due to limitations in computing resources, leading them to forgo certain opportunities this year. In a recent interview with ARK Invest's CEO, Cathie Wood, Friar emphasized the tough choices the organization is making, stating, "We're making some very tough trades at the moment and things we're not pursuing because we don't have enough compute." The urgency of the situation is particularly pronounced as demand for AI technologies is expected to surge in 2026. Friar mentioned that she is actively seeking any available compute resources, highlighting the increasing pressure on OpenAI to meet its operational needs. In a parallel conversation on the "Big Technology Podcast," OpenAI President Greg Brockman echoed Friar's sentiments, noting that the company is struggling to keep pace with the overwhelming demand for its services. This situation underscores a broader issue in the AI sector, where even leading firms are constrained by their access to the necessary computing power required for training and deploying advanced models. Friar pointed out the direct correlation between compute availability and revenue generation, stating, "If you do not have it [compute], you do not have revenue." This scarcity has compelled OpenAI to make strategic decisions, focusing on a limited set of core applications such as personal AI assistants and complex problem-solving tools, as they are unable to address all potential projects with their current compute capacity. As a result, OpenAI has decided to scale back on some initiatives, including the discontinuation of its video app, Sora, in order to concentrate on core, revenue-generating products. Serving around 900 million consumers and over 1 million businesses, OpenAI is in the midst of raising substantial funds, having recently completed a $122 billion funding round to enhance its compute capabilities. Brockman articulated the painful nature of these decisions, stating, "We cannot build compute fast enough to keep up with demand," and acknowledged that the firm is making long-term commitments to secure future capacity. Other AI companies, such as Anthropic, are experiencing similar constraints, having recently imposed usage limits on their models during peak times. Ultimately, the message is clear: in the rapidly evolving landscape of AI, scaling operations hinges on the availability of robust hardware.