The Self-Regulatory Moat

As the federal government struggles to regulate frontier AI, three leading labs—Google, OpenAI, and Anthropic—have created the Standards Authority for Frontier AI (SAFA). The triad aims to standardize benchmarks, pre-deployment review protocols, and incident disclosure procedures. While this move may appear to be a responsible approach to avoid federal regulation, it is actually a strategic move to establish a regulatory moat around their own operations.

Creating self-regulatory standards carries a fixed cost structure, but for these labs, the cost is minimal compared to their multi-billion-dollar compute clusters. For smaller labs or startups, these same protocols represent a significant financial barrier to entry. Historically, industries that have adopted voluntary standards—like railroads and pharmaceutical manufacturing—have found that these bodies create a regulatory capture buffer. This allows the dominant incumbents to define the regulations that will eventually become law.

By establishing SAFA, the three labs ensure that their algorithms meet a shared baseline for safety and compliance. This creates two key economic benefits. Firstly, it reduces price inelasticity at the top of the AI stack. Enterprise procurement teams, wary of risk, will look for labs that have SAFA certification rather than running their own safety audits. This gives the established labs a competitive advantage in securing enterprise contracts, even if their models are not necessarily superior in performance.

Secondly, self-regulation shifts liability. By being part of SAFA, labs can avoid liability claims related to AI safety incidents, as they will have demonstrated compliance with the cartel's standards. This creates a disincentive for smaller labs to challenge the incumbents, as the cost of doing so would outweigh the potential benefits.

The real economic impact of SAFA is that it creates a minimum efficient scale for frontier AI competition. By formalizing standardization, the dominant labs lower the barrier to entry for larger players, while raising the cost of competition for smaller entities. This concentration of market power allows the three labs to preserve pricing power in the API token market, not through superior model performance, but by offering corporate insurance and compliance guarantees.