EU Orders Meta to Open WhatsApp to Competing AI Chatbots

On June 9, 2026, the European Commission ordered Meta Platforms to restore free access to its WhatsApp messaging platform for rival AI assistants — and gave the company five working days to comply [1][2]. The decision marks the Commission's first use of interim antitrust measures in 17 years, a signal of how seriously Brussels views the competition threat posed by Meta's control over messaging infrastructure at the dawn of the AI assistant era [3].

The order is the latest escalation in a confrontation that began in October 2025, when Meta rewrote its WhatsApp Business Solution Terms to ban general-purpose AI chatbots from the platform [4]. That policy, which took effect on January 15, 2026, locked out OpenAI's ChatGPT, Perplexity, Anthropic's Claude, and dozens of smaller AI providers — cutting off access for over 50 million WhatsApp users who had been using ChatGPT alone through the platform [5][6].

The Legal Basis: Article 102, Not the DMA

The Commission's order does not rely on the Digital Markets Act's interoperability mandate — the regulation that has forced WhatsApp to open its messaging protocol to third-party chat apps like BirdyChat and Haiket in Europe [7]. Instead, the action is grounded in Article 102 of the Treaty on the Functioning of the European Union (TFEU), the EU's foundational prohibition against abuse of a dominant market position [8].

The distinction matters. The DMA's messaging interoperability provisions require WhatsApp to allow person-to-person communication with rival messaging services, including text, images, voice messages, and files [7]. The Article 102 case is different: it focuses on whether Meta used WhatsApp's dominance in the messaging market to foreclose competition in the adjacent — and rapidly growing — market for general-purpose AI assistants [8][9].

The Commission opened formal proceedings under Article 102 on December 4, 2025 [9]. It issued a Statement of Objections to Meta on February 9, 2026, stating that Meta's ban on rival AI chatbots "appears at first sight to be in breach of EU competition rules" [10]. When Meta responded in March 2026 by offering to restore access — but at a per-message fee that the Commission described as "not economically sustainable for competitors" — Brussels escalated to interim measures [11][3].

The interim order requires Meta to reinstate third-party AI assistants on the WhatsApp Business API "under the same terms and conditions that were in place" before the October 2025 policy change [3]. This access must remain in place until the investigation concludes, with no set deadline — the Commission has stated the order could last until June 2029 [2].

If Meta fails to comply, it faces fines of up to 10% of its total worldwide annual turnover [1]. Based on Meta's 2025 revenue of $200.97 billion, that theoretical maximum exceeds $20 billion [12].

What's at Stake: WhatsApp's EU Footprint

WhatsApp's market position in Europe is difficult to overstate. The platform has over 3.3 billion monthly active users globally [13], and penetration rates in major EU markets are among the highest in the world.

Source: Statista / Infobip

Data as of Dec 1, 2025CSV

In Germany, WhatsApp reaches 94% of the internet population. In Spain and Portugal, penetration exceeds 91%. Italy sits at 90%, the Netherlands at 85% [13][14]. These numbers make WhatsApp effectively the default communications infrastructure in much of the EU — a status that gives Meta substantial control over which AI services can reach European consumers through their primary messaging channel.

Meta's European revenue in 2025 totaled $46.57 billion, roughly 23% of the company's global revenue of $200.97 billion [12]. While Meta does not break out WhatsApp-specific revenue — the platform's monetization comes primarily through the WhatsApp Business API rather than advertising — the company's "Family of Apps Other Revenue" category, which includes WhatsApp Business paid messaging, has been growing steadily [12].

Source: Meta 10-K Filing

Data as of Feb 1, 2026CSV

The Technical Question: APIs, Not Encryption

A central element of this dispute is what kind of access is actually at issue. The Commission's order concerns the WhatsApp Business API — the programmatic interface that businesses use to send and receive messages on WhatsApp [15]. Before the October 2025 ban, AI providers like OpenAI had built integrations through this API, allowing users to message a WhatsApp number and interact with ChatGPT directly [5].

This is not the same as the DMA's messaging interoperability requirement, which involves message-passing protocols, contact-list access, and end-to-end encryption key exchange between different messaging platforms [16]. The Business API operates at a different layer: messages sent through it are encrypted in transit using the Signal Protocol, but because the API endpoint is controlled by a third-party business (in this case, the AI provider), WhatsApp does not consider these messages end-to-end encrypted in the same sense as peer-to-peer WhatsApp messages [17][18].

Meta has argued that granting third-party AI access raises security and privacy concerns [4]. However, independent analysis of the Business API architecture suggests that the encryption question is largely a red herring in this specific context. The WhatsApp Business API has always operated with the third-party business having access to message content — that is, by design, how business messaging works [18]. AI chatbots accessed WhatsApp through the same API infrastructure used by airlines for boarding passes, banks for transaction alerts, and retailers for customer service. No new encryption vulnerability is created by allowing an AI provider to use the same interface [17].

Meta's own engineering blog acknowledged in 2024 that DMA-mandated messaging interoperability with third-party messaging apps posed genuine cryptographic challenges — specifically around ensuring that the Signal Protocol's end-to-end guarantees hold when messages cross platform boundaries [16]. But those challenges apply to the DMA's messaging interoperability mandate, not to the Business API access that is the subject of the current antitrust order.

Who Benefits?

The Commission framed the order as protecting "a growing market for general-purpose AI assistants" and giving "smaller players and new entrants" space to compete [1]. The AI companies most immediately affected include OpenAI (ChatGPT), Perplexity, Anthropic (Claude), and smaller but fast-growing firms like Luzia and Poke [5][6].

Source: First Page Sage / StatCounter

Data as of Jun 1, 2026CSV

The current AI chatbot market is heavily concentrated. ChatGPT holds 54.7% of global web visits, followed by Google's Gemini at 27.4% and Anthropic's Claude at 8.2% [19]. In Europe specifically, ChatGPT's dominance is even more pronounced, with 70.83% of the regional market [19].

Meta's own AI assistant, Meta AI, was exempt from the October 2025 ban and continued to operate on WhatsApp while rivals were locked out [9]. The Commission's preliminary view is that this self-preferencing — promoting Meta AI while excluding competitors — constitutes the core abuse of dominant position [8].

Meta's counterargument is pointed. A company spokesperson stated: "The European Commission has decided that OpenAI and some of the largest companies in the world can use the paid-for WhatsApp Business product for free" [3]. Meta has characterized the order as "regulatory overreach" and confirmed it will appeal [3].

There is a legitimate question about whether the order primarily benefits large incumbents rather than small startups. OpenAI's ChatGPT had 50 million WhatsApp users before the ban — a distribution channel that smaller AI companies could not match [5]. Restoring free Business API access may disproportionately benefit companies with the engineering resources and scale to build and maintain WhatsApp integrations.

The GDPR Collision Course

The order raises unresolved tensions between EU competition law and EU privacy law. When a user messages an AI chatbot through WhatsApp, the message content is processed by the AI provider's infrastructure. Under GDPR, this raises questions about data controllership, legal basis for processing, purpose limitation, and cross-border data transfers [20].

The European Data Protection Board (EDPB) and the European Commission issued joint guidelines in October 2025 on the interplay between the DMA and the GDPR [20]. These guidelines address general principles — consent must be freely given, specific, informed, and unambiguous — but do not specifically address the scenario of AI chatbot access through messaging APIs [20][21].

There is no public indication that the EDPB was formally consulted on the competition case against Meta or on the specific privacy implications of routing WhatsApp user messages through third-party AI infrastructure [20]. This gap is significant. The EDPB has separately been working with the Commission's AI Office on guidelines for the AI Act's interaction with data protection law, but those guidelines remain in progress [20].

Privacy advocates have noted an underlying tension: the DMA and competition law push for openness and interoperability, while GDPR pushes for data minimization and purpose limitation [21]. Requiring Meta to let third-party AI companies process WhatsApp messages could expand the number of entities with access to user communications — an outcome that data protection regulators have historically opposed.

Precedent and Track Record

The Commission's track record on interoperability mandates provides mixed evidence on whether such orders produce real market change.

The most relevant recent precedent is the DMA's messaging interoperability requirement for WhatsApp itself. After the DMA took effect on March 7, 2024, Meta was required to enable third-party messaging services to interoperate with WhatsApp within three months of receiving a request [7]. By late 2025, only two small services — BirdyChat and Haiket — had launched interoperable chat with WhatsApp in Europe [7]. The limited uptake suggests significant practical barriers to interoperability, even when legally mandated.

Apple's iMessage was not designated as a "core platform service" under the DMA because it did not meet the user thresholds required for designation, a decision the EU reached in February 2024 [22]. This means there is no direct precedent for a DMA interoperability mandate on a major messaging platform being tested at scale.

In other DMA enforcement areas, results have been more visible. After the DMA required Apple to allow browser choice screens, independent browsers reported user spikes — the Cyprus-based Aloha Browser saw a 250% increase in EU users in March 2024 [22]. But browser choice is a simpler intervention than messaging interoperability or AI chatbot access.

The broader history of interoperability mandates in communications — from telephone number portability to email federation — suggests that mandated openness can produce competition, but typically over years rather than months, and often only when combined with strong enforcement and practical implementation support.

Meta's Calculation

Meta faces a strategic dilemma. Compliance with the interim order is straightforward in the narrow sense — restore Business API access under pre-October 2025 terms. The company already has the technical infrastructure; it simply needs to reverse a policy change [3].

But compliance also means subsidizing distribution for its AI competitors. Before the ban, OpenAI was effectively using WhatsApp's 3 billion-user platform as a free customer acquisition channel for ChatGPT, consuming API resources without contributing to WhatsApp's monetization framework [4]. Meta's October 2025 ban was commercially motivated: the company wanted to clear the field for Meta AI and capture the value of AI integration within its own ecosystem [9].

The fine for non-compliance — up to 10% of global turnover, or theoretically over $20 billion — makes defiance impractical [1][12]. Meta's stated strategy is to comply while appealing, a common approach in EU antitrust cases that can take years to resolve through the courts.

What Happens Next

The immediate timeline is clear: Meta has until approximately June 16, 2026, to restore free access for rival AI assistants to the WhatsApp Business API [1]. The underlying antitrust investigation has no formal deadline and could extend until 2029 [2].

The longer-term implications are broader. This case establishes that EU competition authorities view messaging platforms as critical infrastructure for AI distribution — and that self-preferencing an in-house AI assistant by locking out competitors can constitute an abuse of dominant position [8]. If the Commission's theory holds through litigation, it would create a precedent applicable to any dominant platform that bundles AI features while restricting third-party AI access.

For the AI industry, the order provides a partial answer to one of the central questions of the current moment: can the dominant platforms of the mobile era control the distribution of AI services? In Europe, at least, the Commission's answer is no — not without facing antitrust consequences.

The unresolved question is whether forced access translates into actual competition. WhatsApp's 50 million ChatGPT users represented less than 10% of the platform's estimated EU user base [5][13]. Whether those users — and new ones — will adopt AI assistants from smaller providers, or whether the restored access primarily reinforces the position of large incumbents like OpenAI and Google, will determine whether this intervention produces genuine market change or remains a regulatory exercise with limited practical impact.