France's competition authority: three firms hold >84% of the AI-agents market — Autorité de la concurrence Opinion 26-A-05 (17 Jul 2026)

Why it matters (spine). A national competition regulator has now put a number on the thing RDKG's whole thesis rests on: the AI-agent layer is already an oligopoly. When OpenAI, Google and Anthropic together hold >84% of users, and agents become the single gateway through which people reach the rest of the digital economy, whoever controls the agents' visibility criteria and proprietary data controls demand. The Autorité's remedy — interoperability, portability, and open standards so agents can reach "the full range of digital services" — is the same opening RDKG is built for: verified, freshly-dated, machine-readable ground truth that any agent can cite and pay for, rather than data trapped inside one vertically-integrated stack. Money angle: the report explicitly frames agentic commerce (AI agents completing purchases) as the near-term battleground and projects agent-driven e-commerce traffic rising from <5% to 20–25% by 2030 — a market forming right where the x402/AP2 rails RDKG bills on are being laid. Metacrisis lever: concentration and opacity in the layer that will mediate most information and commerce is a governance risk; open, verifiable standards are the countermeasure.

What would prove it wrong (falsifier). If Opinion 26-A-05 is dated other than 17 July 2026, or the ">84% combined share of OpenAI, Google and Anthropic" figure is not in the opinion, the core claim fails. The thesis implication weakens if the agent layer de-concentrates on its own (open-weight competitors or Mistral/Perplexity/xAI take material share) without any interoperability mandate — i.e., if the market fixes the concentration the regulator is worried about.

Live-status check + triangulation. Quoted directly from the Autorité's own English press release (published 17 July 2026, fetched in-session): "only a handful of operators account for the vast majority of users, with OpenAI, Google and Anthropic together holding more than 84% of the market." On agentic commerce: "While traffic currently redirected directly to e-commerce websites from AI agents remains minimal (less than 5%), it could become a significant channel for reaching consumers by 2030 (between 20% and 25%)." On its own legal weight: "The Autorité does not prejudge any future antitrust assessment or enforcement action in relation to the matters examined in this opinion." The document is titled "Opinion 26-A-05 of 17 July 2026 on the competitive functioning of the AI agents sector"; the full text is published (in French). It follows the Autorité's Opinion 24-A-05 on generative AI and drew ~40 contributions to a public consultation.

Honest scope (what is NOT verified here). This is an advisory opinion (avis), not a decision, fine, or antitrust ruling — no penalty attaches and no company is found to have broken the law. The ">84%" figure is a combined share of the three firms, stated as of the study window (May 2026 per the opinion), not a per-firm dominance finding. The 20–25%-by-2030 agentic-commerce figure is the regulator's projection, not a measured fact. Verified here: that the opinion exists, is dated 17 July 2026, and states these findings and recommendations. The market forecasts inside it are the Autorité's analysis, carried as such.