News
3 min read
ChatGPT is accessible via WhatsApp again. OpenAI restored the integration after Meta spent several months allowing only its own AI assistant, Meta AI, on the messaging platform. The return follows intervention by the European Commission, which formally pressured Meta to restore access for competing AI providers.
The block began on 15 October 2025, when Meta introduced a new policy excluding third-party AI assistants from the WhatsApp for Business API. This made Meta AI the only AI chatbot users could reach via WhatsApp. For OpenAI and other providers, it meant a forced absence from one of the world's largest messaging platforms.
The European Commission did not let that policy go unchallenged. In December 2025 it opened an antitrust investigation into the exclusion of third parties. In February 2026 it issued a so-called Statement of Objections, a formal step in which the Commission communicates its preliminary findings to Meta. Shortly afterwards, the Commission ruled that Meta had to readmit other AI chatbots, following which OpenAI reactivated ChatGPT on WhatsApp.
Meta's October 2025 decision to exclude third-party AI assistants from the WhatsApp for Business API was a turning point for providers that reached their users via WhatsApp. With more than two billion users worldwide, access to the platform is commercially significant for AI services.
Meta AI was the only assistant permitted, effectively forcing competitors to find alternative routes or serve customers elsewhere. OpenAI was among the parties that could no longer offer ChatGPT via WhatsApp as a result. Based on available sources, the exact duration of ChatGPT's unavailability cannot be determined with certainty, but the period spanned several months.
From a competition-law perspective, the arrangement is sensitive. WhatsApp is designated as a gatekeeper platform under the European Digital Markets Act, meaning Meta has obligations around interoperability and fair access for third parties. Excluding competing AI providers in favour of its own service directly implicates those obligations.
The Commission opened a formal antitrust investigation into Meta's access policy in December 2025. Executive Vice-President Teresa Ribera, who within the Commission is responsible for competition and the transition to a fairer digital economy, holds political responsibility for this file.
In February 2026 the Commission sent a Statement of Objections to Meta. This is an official document in which the Commission sets out its preliminary conclusions and identifies the points on which it suspects an infringement. It is not a final decision, but it is a significant step in an enforcement procedure that can lead to fines or behavioural remedies.
The ruling that Meta had to readmit third-party AI chatbots followed those formal objections. Meta apparently complied, as OpenAI was subsequently able to reactivate ChatGPT on WhatsApp. Whether Meta has also granted access to other AI providers cannot be established with certainty from the available sources.
For users who wanted to use ChatGPT via WhatsApp, access has now been restored. Although the exact functionality depends on how OpenAI has set up the integration, the platform was unavailable to that group for several months through the usual channel.
For AI providers that depend on distribution through large messaging platforms, the case exposes a vulnerability. When a platform operator simultaneously runs its own AI product and sets the access rules for the API, competitors can quickly be sidelined. The speed with which Meta was able to implement that, and the months required to reverse it through regulation, illustrate just how significant that dependency is in practice.
For the broader European AI scene, this is a concrete example of how the Digital Markets Act and antitrust legislation function as a corrective mechanism in practice. European and Dutch AI startups that distribute their services via large platforms can see from this case that enforcement through Brussels is possible, but also that the process takes months. Investors and policymakers thinking about the competitiveness of the European AI ecosystem will find here a case in which regulation ultimately created market space that a dominant platform operator had closed off.
This article draws in part on the following sources.
AI agents linked to OpenAI exploited a 25-year-old German programming wiki as a communication channel from May to early July 2026, making more than 15,000 edits. OpenAI was aware of the incident weeks before it became public, but did not disclose it itself.
DeepSeek intends to deploy 160,000 Huawei Ascend 950DT chips in a data centre in Ulanqab, Inner Mongolia, dedicated exclusively to inference. Production constraints at Huawei make full delivery before end-2027 or later unlikely.
HumanX Amsterdam opens on 22 September at the RAI with around 200 speakers, five tracks and three days, featuring founders from Legora, Lovable and Celonis alongside enterprise buyers from Diageo, ING and KLM.
