OpenAI CEO Sam Altman responds to investor’s revenue concerns by offering to buy back shares

There’s a particular kind of confidence that comes with being the CEO of the most talked-about AI company on Earth. The kind where someone asks you a pointed financial question and you respond by offering to find a buyer for their shares.

That’s essentially what happened when Brad Gerstner, founder of Altimeter Capital and host of the BG2 podcast, pressed OpenAI CEO Sam Altman about the company’s revenue trajectory relative to its staggering infrastructure commitments. Rather than directly address the concern, Altman offered to help find a buyer for Gerstner’s shares.

The numbers that prompted the question

OpenAI has reportedly committed over $1.4 trillion toward AI infrastructure. The company’s disclosed annual revenue sits at approximately $13 billion.

During the podcast conversation, which aired in late October or early November 2025, Altman pushed back on the framing. He asserted that OpenAI’s actual revenue is significantly higher than the $13 billion figure and projected a trajectory that would see the company exceed $20 billion by the end of 2025, eventually reaching hundreds of billions by 2030.

Microsoft CEO Satya Nadella also joined the discussion, adding weight to the conversation about the economics of scaling AI.

OpenAI’s estimated valuation after a secondary share sale reportedly reached $500 billion.

Why the deflection matters

Gerstner isn’t a random Twitter user lobbing questions into the void. He runs Altimeter Capital, a significant investment firm that has previously financed OpenAI’s growth. When someone of that stature asks about the math, the audience pays attention.

Even if OpenAI hits $20 billion in revenue by year’s end, that figure still represents a small fraction of the $1.4 trillion in infrastructure commitments. The implied bet is that AI revenue will scale exponentially, not linearly, over the next five years.

What this means for investors

Going from $13 billion to $20 billion in annual revenue is impressive growth in absolute terms. Going from $20 billion to hundreds of billions by 2030 requires OpenAI to become embedded in enterprise workflows, government systems, and industries that haven’t yet adopted AI at scale.

There’s also the question of margin. Running inference at scale — the process of actually serving AI responses to millions of users — is extraordinarily expensive. Even as hardware costs decline over time, the sheer volume of compute required means that top-line revenue growth doesn’t automatically translate to bottom-line health.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.