Being the CEO of a leading AI company often comes with a specific kind of confidence. When asked about financial projections, the response can reveal much. Sam Altman, CEO of OpenAI, recently engaged in a significant dialogue with Brad Gerstner of Altimeter Capital regarding OpenAI's financial trajectory amid its large infrastructure investments.
OpenAI has reportedly committed over $1.4 trillion to establish its AI capabilities. However, its disclosed annual revenue is around $13 billion. During a podcast conversation aired in late 2025, Altman pushed back against the concerns raised by Gerstner, claiming that OpenAI's actual revenue sits well above the figure presented in public discussions. He projected that the company could exceed $20 billion in revenue by the end of 2025, aiming for even higher rates as they approach 2030.
Microsoft’s CEO, Satya Nadella, also contributed to this conversation, highlighting the importance of economic scaling in AI technology. It is noteworthy that OpenAI's valuation reached an impressive $500 billion following a secondary share sale, which underscores confidence from investors.
What does Altman's response imply for investors? Gerstner is not just any investor; he leads a major investment firm that has backed OpenAI. When he raises questions about the financials, it attracts attention. If OpenAI achieves $20 billion in revenue, this still represents only a small portion of its substantial infrastructure investments. The critical takeaway for investors is the expectation that AI revenue must scale exponentially in the coming years.
Although increasing from $13 billion to $20 billion represents notable growth, further expansion from $20 billion to hundreds of billions by 2030 involves considerable challenges. This growth entails embedding AI solutions into enterprise workflows, governmental systems, and industries that have yet to adopt AI.
Moreover, considering operational margins is essential. Running AI at scale to meet user demand involves substantial operational expenses. As hardware costs may decline, the sheer computational requirements mean that revenue growth alone will not guarantee profitability.