The EU just drew a line at 45 million users, and the companies on either side of it now play by completely different rules.
The Summary
- The EU designated ChatGPT, Reddit, and Roblox as "Very Large Online Platforms" under the Digital Services Act, triggering strict content moderation and transparency requirements
- Anthropic's Claude avoided the designation by staying under the 45 million monthly active user threshold in the EU
- The move signals increased compliance costs and regulatory scrutiny that will reshape growth strategies for AI platforms and social networks operating in Europe
The Signal
The Digital Services Act just created two classes of platforms: those big enough to deserve Brussels' attention, and those small enough to avoid it. ChatGPT, Reddit, and Roblox crossed the 45 million monthly active user line in the EU, earning them the Very Large Online Platform label and everything that comes with it. Anthropic's Claude didn't.
This isn't about product quality or safety risk. It's purely arithmetic. The DSA designation triggers obligations that smaller platforms, even sophisticated AI systems, can sidestep by staying under the threshold. ChatGPT now faces mandatory risk assessments, independent audits, and content moderation transparency reports. Claude doesn't.
"The 45 million user threshold creates a regulatory moat that has nothing to do with capability and everything to do with reach."
For OpenAI, this is the cost of market dominance. ChatGPT's rapid adoption across Europe, its integration into enterprise workflows, and its consumer brand recognition all worked against it here. The designation brings increased compliance costs that will impact growth strategies and investor confidence, forcing OpenAI to divert engineering resources from product development to regulatory infrastructure.
Reddit and Roblox face the same calculus, but for different reasons. Reddit's designation reflects its role as a de facto public square for millions of Europeans. Roblox's reflects the EU's growing concern about platforms where user-generated content dominates and minors congregate. Both now need legal teams in Brussels, content moderation systems that meet EU standards, and transparency tools they didn't budget for this year.
The interesting question isn't why these three platforms got tagged. It's why Anthropic managed to build a frontier AI model that competes directly with ChatGPT while keeping its EU user base just under the regulatory tripwire. Three possibilities:
- Anthropic is genuinely less popular in Europe than in the US
- The company throttled EU growth intentionally to avoid DSA obligations
- Enterprise adoption, which is Anthropic's strength, doesn't generate the same monthly active user counts as consumer chatbots
The Implication
If you're building an AI agent platform or consumer AI product, the 45 million EU user threshold is now a strategic planning number. Cross it and you're in a different regulatory universe. Stay under it and you can move faster than your larger competitors, at least in Europe.
For OpenAI, the path forward is compliance, not retreat. ChatGPT's value to European users and businesses justifies the overhead. But expect smaller AI companies to optimize around this line, potentially fragmenting how AI products scale globally. The EU just made "stay small in Europe" a viable competitive strategy.