After January’s blanket ban and March’s $0.0625-per-message fee, Meta has filed a fresh European Commission proposal: free WhatsApp access for OpenAI, Perplexity, and others, up to a usage cap, then a fee.
Meta has filed a fresh European Commission proposal that would give rival AI chatbots, including OpenAI’s ChatGPT, Perplexity, Anthropic’s Claude and others, free access to WhatsApp in Europe up to a usage threshold, then start charging beyond it, Reuters reported on Monday, citing people familiar with the matter.
The proposal was submitted last week after the European Commission signalled it was considering a formal order forcing Meta to open WhatsApp to third-party AI services under the Digital Markets Act.
Meta blocked all third-party AI assistants on WhatsApp on 15 January, an action that affected ChatGPT, Microsoft Copilot, Perplexity, and several smaller third-party assistants that had built distribution inside the messaging app.
The blanket ban triggered antitrust attention in both Brussels and Brazil. Meta partially relented in March, agreeing to open WhatsApp to rival chatbots but at a $0.0625-per-message fee, a structure both regulators and the affected AI companies treated as commercially prohibitive.
The new proposal is the third iteration of the same posture. Free access up to a usage cap is the structure that regulators across multiple jurisdictions have been pushing for since the original ban. Meta is attempting to land a compromise that satisfies the European Commission’s antitrust review without entirely abandoning the per-message-revenue framework it tried to introduce in March. The Commission is reviewing the proposal alongside the broader Digital Markets Act gatekeeper obligations workstream.
What the new structure actually delivers in practice depends on the cap. Meta’s submission may also include a one-month preferential access window for rival companies to integrate with WhatsApp’s developer infrastructure before pricing takes effect.
If the threshold is high enough to allow normal consumer use without triggering the fee, the proposal materially changes the competitive dynamics of the European AI-chatbot market. If it is calibrated to trigger the fee within days, the proposal is functionally a rebadge of the March position.
The stakes for the affected AI companies are significant. WhatsApp has roughly 2bn monthly active users globally and approximately 500m monthly active users in Europe; the platform is, in regions like Brazil, India, and parts of the Mediterranean, the default consumer messaging infrastructure.
Direct chatbot distribution inside WhatsApp has been an established product category for several years and Perplexity and ChatGPT have been individually accessible inside the app even during periods when the blanket ban was supposed to be in effect.
The fact that the ban has not produced a clean technical shutdown is part of what has given the regulators leverage in the current negotiation.
The wider Digital Markets Act context is the part this story sits inside. The European Commission has been running a comparable enforcement track against Apple over App Store anti-steering and adjacent gatekeeper-platform questions, and the WhatsApp case is structurally similar: a dominant platform whose interoperability obligations under the DMA cut against the platform’s own commercial incentive to favour first-party AI products.
Meta’s first-party AI product, Meta AI, sits inside WhatsApp by default; every consumer interaction the platform routes to Meta AI is a consumer interaction not routed to OpenAI or Perplexity. The Commission’s posture, on the available evidence, is that the DMA’s interoperability obligations apply to that competitive trade specifically.
The competitive read for the agent-distribution category is sharper. Dust’s $40m Series B landed earlier this month on a thesis that the AI agent that wins is the one with frictionless distribution inside the user’s existing communication surface. Whether OpenAI, Perplexity and Anthropic can build sustainable economics around WhatsApp-routed consumer AI usage depends on where the Commission lands the cap.
Meta did not respond to Reuters’ request for comment. The Commission’s review timeline is not public but is, on standard Article 8 DMA cadence, expected to produce a formal decision inside the next several months. The previous January-to-March cycle, documented in industry coverage, suggests the operative compromise can move quickly once both sides have a structure to work from. The free-up-to-a-cap framework is that structure.
