Key Takeaways

  • Anthropic achieved a $30 billion annual revenue run rate in under a year, driven by over a thousand enterprise customers paying $1M+.
  • Brad Gersonner described this as the “largest revenue explosion in the history of technology,” suggesting a “near infinite TAM” for intelligence.
  • Chamath Palihapitiya questioned the actual profitability, pointing to the massive capital expenditure on compute and the unknown gross margins.
  • David Sacks viewed the rapid revenue growth as vindication for the hundreds of billions invested in AI infrastructure.

The Disagreement

Brad Gersonner celebrated Anthropic’s revenue ramp as unprecedented. He highlighted the scale, saying, “you have the largest revenue explosion in the history of technology.” Gersonner argued this growth proves a “near infinite TAM” for intelligence, claiming that the market for AI is “radically different than anything that we’ve seen before.” He even pushed back on concerns about margins, suggesting, “you might actually expect to see these companies their gross margins are exploding higher.”

Chamath Palihapitiya countered this optimism, focusing on profitability. He expressed skepticism about the net revenue, stating, “The thing that we need to understand is how gross margin negative is this revenue growth. We don’t know that and at least we don’t as outsiders.” Palihapitiya’s point is that top-line growth can mask immense costs, especially with the enormous compute required for AI models. Without knowing the cost of goods sold, revenue figures alone are misleading.

Who’s Right (and When They’re Wrong)

Both Gersonner and Palihapitiya hit on essential truths, but from different angles. Gersonner is right about the sheer demand for AI capabilities. The rapid adoption by enterprise customers, many paying seven-figure sums, confirms that businesses are willing to pay for intelligence. This does suggest a massive, previously untapped market. The fact that Anthropic scaled so quickly indicates a real need for these tools.

However, Palihapitiya’s skepticism is warranted, especially for founders considering AI ventures. Generating revenue is one thing; generating profitable revenue is another. The AI industry’s reliance on incredibly expensive compute resources means that gross margins are not guaranteed, even with high prices. David Sacks provided some balance, seeing the revenue as “justification” for the hundreds of billions in AI capex. His argument suggests that while the costs are high, the market can support them.

The truth likely lies in between. AI companies can achieve extraordinary revenue, validating the investment in underlying infrastructure. But profitability will hinge on efficiency, scale, and defensible advantages beyond raw compute. For most startups, simply acquiring customers isn’t enough; the unit economics must be sound from day one. An “infinite TAM” doesn’t mean infinite profit at any cost.

What to Do With This

If you’re building an AI product, stop celebrating top-line revenue until you understand your gross margins. Before you launch, model your COGS for every single inference. Factor in GPU costs, energy, and network. Then, find five customers willing to pay for your product at a price that ensures positive gross margins after all compute expenses. Do not scale until those unit economics are proven viable.