The operative question in every AI safety debate is where the accountability mechanism sits in the deployment chain, not whether AI poses risks. Mark Zuckerberg of Meta and Jensen Huang of Nvidia have given the same answer: market forces already create the right incentive for companies to ship AI products safely, and the responsibility for ensuring that falls on individual companies.
The constraint here is incentive alignment, not development pace. Both executives are making a structural argument. When a company releases a product that harms users, it loses the market. That feedback loop, in their shared framing, is already the load-bearing governance mechanism in the industry's economics, and it does not require an externally imposed slowdown to function.
The position carries a direct implication. If market accountability is already the operative mechanism, there is no case for voluntary deceleration at either Meta or Nvidia. Development at both companies continues at its current rate, and both executives frame that rate as the natural output of a market already calibrated toward safety.
What their argument does not settle is whether the mechanism is performing. The self-regulation thesis is a claim about where enforcement should sit in the AI governance stack. Whether the market feedback loop is actually catching safety failures before they cause harm is precisely the question that remains open, and the one their critics will keep pressing. For now, Zuckerberg and Huang have placed the answer firmly with individual companies.