Meta will allow rival AI companies to offer chatbots on WhatsApp in both Europe and Brazil, ending a months-long standoff with regulators who called the company’s AI lockout anti-competitive.
The decision came within 48 hours in both regions: Europe on March 5, Brazil on March 6. Together, these markets represent over 700 million WhatsApp users who will now have access to AI assistants from companies other than Meta.
It’s a significant forced opening in what has become the primary distribution channel for consumer AI.
What Happened
In October 2025, Meta updated its WhatsApp Business API terms to ban general-purpose AI chatbots from the platform. The company claimed the API “isn’t designed to be a platform for the distribution of chatbots.”
The policy would have blocked AI providers like OpenAI, Perplexity, and Anthropic from reaching users through WhatsApp, while Meta’s own “Meta AI” assistant remained available to all 2+ billion users globally.
Regulators didn’t buy it.
The European Commission opened an antitrust investigation in December 2025. By February 9, 2026, it issued a Statement of Objections setting out its preliminary view that Meta had breached EU competition rules.
Brazil’s CADE (Administrative Council for Economic Defense) moved even faster. On January 12, 2026, it ordered Meta to suspend the policy through a temporary injunction, ruling that an outright ban on external AI services “would not be proportionate” given WhatsApp’s dominance in Brazilian messaging.
Meta appealed. It lost.
The New Terms
Rather than fight extended legal battles, Meta capitulated. The terms:
Europe: Rival AI companies will pay between €0.0490 and €0.1323 per “non-template message” their chatbots send or receive, with rates varying by country. The arrangement lasts 12 months.
Brazil: AI providers will pay $0.0625 per non-template message, effective March 11, 2026.
For context: if an AI chatbot has a conversation of 20 messages with a user in Brazil, the AI company owes Meta $1.25. At scale, this becomes serious money. If 10 million Brazilian users have ten 20-message conversations per month, that’s $125 million annually flowing to Meta from competitors.
Why This Matters
The antitrust reasoning reveals how regulators are starting to think about AI distribution.
Teresa Ribera, the EU’s executive vice-president for competition, stated the Commission was investigating whether to “act quickly to prevent any possible irreparable harm to competition in the AI space.”
The concern isn’t just that Meta makes a popular AI assistant. It’s that Meta controls the surface where billions of people communicate. If only Meta AI can reach users through WhatsApp, other AI companies face a structural disadvantage no amount of technical innovation can overcome.
CADE’s ruling in Brazil made this explicit: WhatsApp’s “relevance in the Brazilian instant messaging services market” meant Meta couldn’t simply lock out competitors.
The Distribution Problem
WhatsApp has over 2 billion users globally. In Brazil alone, it’s estimated that 99% of smartphone owners use the app. It’s the default way millions of people communicate with businesses, receive notifications, and increasingly interact with automated services.
When Meta launched Meta AI inside WhatsApp, it gave its own assistant instant access to this massive user base. Competitors had no equivalent distribution channel. You can build the best AI in the world, but if you can’t reach users, it doesn’t matter.
This is the same dynamic that’s played out repeatedly in tech. Google’s dominance in search gave it an advantage in advertising. Apple’s control of the iPhone gave it advantages in payments, music, and app distribution. Now messaging platforms are emerging as the key distribution channel for AI.
The regulatory response suggests that owning a dominant communication platform doesn’t automatically entitle you to own AI distribution through that platform. At least, not exclusively.
What AI Companies Get
Access to WhatsApp’s Business API means AI companies can build bots that:
- Respond to user messages in real-time
- Integrate with business workflows
- Offer customer service, search, and assistance
- Compete directly with Meta AI for user attention
The fee structure creates a new business model: AI-as-a-service through messaging. Companies like OpenAI, Anthropic, and Perplexity can now potentially reach WhatsApp’s user base, though they’ll pay Meta for the privilege.
Whether the fees are reasonable is already contested. AI rivals told MLex that Meta’s solution “perpetuates competition problems” because the fees still create barriers and the 12-month European arrangement provides no long-term certainty.
The Bigger Picture
This matters beyond WhatsApp.
Apple faces similar questions about AI distribution through Siri and the iPhone. Google controls AI access through Android and Chrome. Microsoft dominates enterprise AI through Office 365 and Teams.
If WhatsApp can’t lock out competing AI services, can these other platforms? The antitrust logic would seem to apply equally: if you control a dominant distribution surface, you may not be able to use that control to exclude AI competitors.
The regulators in Europe and Brazil have established a principle: messaging platforms with dominant market positions must provide access to third-party AI services under reasonable terms.
That principle will almost certainly be tested against other platforms in other markets.
What to Watch
Pricing battles: AI companies will argue the per-message fees are too high. Meta will argue they’re necessary to maintain platform quality. Regulators will have to decide what “reasonable” means.
European deadline: The 12-month arrangement expires in March 2027. Watch for whether Meta tries to reimpose restrictions or whether the Commission makes the opening permanent.
Other markets: WhatsApp dominates messaging in India, Indonesia, and much of Latin America and Africa. Regulators in these markets may follow Europe and Brazil’s lead.
Apple and Google: The logic that forced WhatsApp open applies to Siri and Google Assistant. Don’t be surprised if similar investigations follow.
The Bottom Line
Meta tried to use its messaging dominance to control AI distribution. Two of the world’s most aggressive competition regulators said no.
The immediate effect is that AI companies can now reach WhatsApp users in two major markets. The longer-term effect may be establishing that owning the communication channel doesn’t mean owning the AI layer on top of it.
For users in Europe and Brazil, this means more AI choices. For Meta, it means sharing the most valuable real estate in consumer AI with competitors. For the AI industry, it’s a test case for whether dominant platforms can be forced to provide access.
The answer, at least in these two markets, is yes.