French AI Company Mistral Criticizes Anthropic for Monopoly, Both Sides Have Opposing Stances on AI Regulation Pace
French AI company Mistral, along with several other European AI firms, openly opposes Anthropic's proposal to "slow down advanced AI development," with both sides having directly opposing stances on the pace of AI regulation.
These European companies accuse Anthropic of using "safety concerns" rhetoric to push for stricter regulatory thresholds, which would objectively increase compliance costs for new entrants and latecomers, thereby consolidating Anthropic's market advantage as an early American enterprise.
The core of the controversy lies in Anthropic's long-standing public advocacy for caution in the development and deployment of cutting-edge AI models, supporting the introduction of stricter safety assessments and regulatory review mechanisms; whereas Mistral and other European firms interpret this stance as a competitive strategy rather than purely a safety consideration.
Mistral, one of the most watched AI startups in Europe, has consistently emphasized the need for Europe to accelerate its autonomous AI capability building to avoid long-term dependence on American companies at the foundational model level, which is also the direct background for its opposition to the "slowdown theory."
From a market competition perspective, if regulatory thresholds are raised in the name of safety, it directly pressures European latecomers with smaller funding and computing power compared to leading American companies like Anthropic and OpenAI; meanwhile, the established American leaders with model capabilities and funding advantages may benefit relatively from higher compliance thresholds, as these thresholds will raise the entry costs for followers. This is also the core logic behind Mistral and others linking Anthropic's safety claims to market position issues.
Source: Public Information
ABAB AI Insight
Since its establishment, Anthropic has centered its narrative on "AI safety," with several members of its founding team coming from the early safety-related team at OpenAI, advocating for a responsible scaling policy and repeatedly calling for stricter evaluation and regulatory checkpoints for the training and deployment of cutting-edge models at U.S. Congress and various AI safety summits. This stance has helped Anthropic establish an image as the "most cautious frontier laboratory" within policy-making circles, granting it relatively advanced discourse power in regulatory discussions.
Mistral, founded by several French researchers from Meta and Google DeepMind, has always positioned itself as a representative of "European sovereign AI," with funding and policy resources mobilized towards "reducing dependence on U.S. foundational models." French and EU-level industrial policies also tend to support local foundational model companies. Mistral's public opposition to Anthropic's "slowdown theory" essentially aims to maintain the time window for European companies to catch up with leading American firms during the regulatory window period; any regulatory actions that raise compliance thresholds will directly compress this window.
This controversy intertwines "safety narratives" with "market positioning," resembling the early cloud computing industry's leading firms pushing for complex compliance certifications that raised entry barriers for smaller cloud providers, and is similar to some traditional industry's leading companies lobbying for higher entry standards to indirectly exclude latecomers. The European AI industry is currently in a "catch-up phase," while leading American laboratories have entered a "rule-making phase," and the differences in their industrial positions directly determine the completely opposite interpretations of the same regulatory proposal.
This controversy essentially represents a transfer of discourse power in the path of regulatory change—safety standards and regulatory thresholds may seem technically neutral, but who proposes them and at what pace they are advanced directly determines who will dominate in the next phase of AI competition. This mechanism occurs because the AI regulatory framework is still in the formative stage, with rule details not yet fixed; early entrants have the motivation to package their achieved capability levels as "safety baselines," while latecomers must interpret the same baseline as "competitive barriers," with both sides actually competing for the authority to set future regulatory standards.
ABAB News · Cognitive Laws
The earlier safety standards are set, the more early entrants win.
Latecomers are not opposing safety, but the timetable set by others.
The regulatory window period is a window for players to re-rank.