The AI lab that positioned itself as the responsible choice is now asking public markets to trust it without giving them actual control.

The Summary

The Signal

Anthropic's updated S-1 filing reveals a company hitting scale but losing momentum. Revenue climbed to $17 billion in Q3, but the growth rate slowed from prior quarters. More telling, the AI lab swung back to an operating loss after reporting profitability earlier in the year. The pattern suggests the easy revenue has been captured. What comes next requires more spending for smaller gains.

The numbers matter, but the real story is governance. NYC Comptroller Brad Lander, who oversees $280 billion in pension funds, is publicly challenging Anthropic's IPO structure. His argument cuts to the bone: Anthropic wants to raise money from ordinary people while preventing those same people from having any say in how the technology develops.

"Anthropic's approach threatens to disempower shareholders and prevent ordinary people from reining in a potentially dangerous form of technology."

This isn't abstract corporate governance theater. Anthropic has positioned itself as the safety-first AI lab, the one that takes constitutional AI and responsible scaling seriously. That brand attracted capital and talent precisely because it promised a different path than the move-fast-break-things approach. Now it wants to go public with a structure that insulates leadership from shareholder pressure.

The timing is pointed. Anthropic is hitting the market as:

  • Compute costs continue climbing faster than revenue
  • Competition from OpenAI, Google, and open-source models intensifies
  • Questions about AI safety and alignment remain unsolved
  • Regulatory scrutiny of AI companies accelerates globally

Lander's criticism exposes the contradiction at the heart of the AI boom. These companies claim they're building world-changing, potentially existential technology. Then they design governance structures that make it nearly impossible for investors or the public to influence decisions. You can't have it both ways. Either the technology is powerful enough to warrant special governance, or it's a normal company that should answer to normal shareholder dynamics.

The IPO structure debate extends beyond Anthropic. Multiple AI labs are exploring public listings with dual-class shares, special purpose vehicles, or other mechanisms that concentrate control. The argument is always the same: AI development requires long-term thinking that quarterly earnings pressure would destroy. But "trust us" has never been a sufficient governance model for technologies with systemic risk.

The Implication

Watch how this plays out. If Anthropic proceeds with a structure that sidelines shareholders, it sets precedent for every AI company that follows. Public markets become pure funding vehicles with no oversight rights. That's a formula for concentrated power in an industry already defined by it.

For institutional investors like pension funds, this is decision time. Take diluted shares in companies building AGI, or walk away and cede the entire sector to private capital and insiders. Neither option is great. But Lander is drawing a line that other large allocators may follow, especially as AI safety concerns move from theoretical to concrete.

Sources

Bloomberg Tech | Fortune Tech