Meta has joined Apple in delaying some of its most advanced artificial intelligence features for European Union users, turning regulatory uncertainty into a visible product divide. Meta said an upcoming multimodal Llama model would not launch in the bloc, even though a text-only version of Llama 3 was still expected to be available. The model was designed to work across video, audio, images and text and was intended for release under an open licence. Its absence therefore affects not only consumers, but also European developers and businesses that hoped to build products on top of the technology.
Why Meta Is Withholding Its Multimodal Llama Model
Meta attributed the decision to what it described as the unpredictable nature of the European regulatory environment. The company planned to release the multimodal Llama model in other markets over the following months, while excluding the EU. That creates an unusual situation for an open-model strategy: the underlying technology may be broadly accessible elsewhere, yet companies operating in Europe cannot rely on it for locally offered services. The restriction also reaches beyond a single model because Meta indicated that future multimodal releases would face the same regional limitation.
The distinction between text-only and multimodal systems matters. A conventional language model can generate and analyse text, but Meta’s forthcoming system was intended to combine several forms of information in a single interaction. That capability is central to products such as smart glasses, which need to understand images and spoken requests together. Meta planned to incorporate its multimodal technology into products including its Ray-Ban smart glasses. European users could consequently receive a materially narrower version of a product than customers in markets where image understanding and related functions are enabled.
Apple and Meta Create a Wider European AI Gap
Meta’s decision followed Apple’s announcement that Apple Intelligence would not initially be introduced in the EU. Taken together, the moves suggest that Europe’s regulatory framework is beginning to influence not just how AI products operate, but where they launch at all. A delayed release can become a competitive issue when AI capabilities are embedded into operating systems, devices and developer platforms. European companies may have to wait, adopt different models or maintain separate technical stacks for the EU and other markets.
The impact also extends to companies outside Europe. A developer that uses Meta’s multimodal model in a global product would be unable to offer the same service throughout one of the world’s largest economic regions. That can increase engineering costs and complicate product planning, because availability is no longer determined only by language support, infrastructure or demand. It is also determined by whether a provider believes the regulatory obligations are sufficiently clear. Regional product fragmentation may therefore become a significant consequence of the emerging AI rulebook.
The AI Act Raises Questions About Compliance and Copyright
Meta’s announcement came shortly after the EU finalised compliance deadlines for its AI Act. Companies were given until August 2026 to meet important requirements concerning matters such as transparency, copyright and certain high-risk uses of AI. The legislation aims to establish safeguards while giving businesses a common framework across the bloc. Yet Meta’s response shows that a detailed law does not automatically remove uncertainty for model developers, especially when obligations must still be interpreted and translated into product, data and licensing decisions.
Open models add another layer to the debate. When a company releases model weights under a permissive licence, independent developers can adapt and deploy them in ways the original provider does not directly control. Questions about responsibility, training data and downstream compliance may therefore be harder to resolve than for a centrally hosted service. Meta’s choice suggests it did not believe the commercial opportunity outweighed that uncertainty for its multimodal release. The result is a practical test of whether Europe can combine strong rules with timely access to frontier AI systems.
Europe Must Balance Protection With Access to Innovation
The EU’s approach is built around the idea that powerful technology should not enter society without meaningful protections. Rules on transparency, copyright and sensitive applications respond to real concerns about how AI is trained and used. At the same time, repeated exclusions by large technology companies create a cost of their own. If European consumers receive fewer capabilities and local developers cannot use models available to competitors elsewhere, the region risks weakening the businesses that are expected to build its next generation of AI products.
Meta and Apple may ultimately resolve their concerns and introduce the delayed features after adapting them to European requirements. However, the decisions establish an important precedent: companies are willing to use geographic availability as a compliance strategy. Regulators will need to decide whether this is a temporary phase while the rules mature or evidence that guidance needs to become more predictable. For European users and businesses, the central question is no longer simply whether AI will be regulated, but whether regulation and innovation can advance without creating a lasting technology gap.
The practical response should involve more than a contest of public statements. Model providers can explain which obligations they consider uncertain, while European institutions can provide implementation guidance that addresses realistic deployment and licensing scenarios. Developers and customers also need adequate notice when regional availability changes, because they may be designing products around promised capabilities. A predictable process would not guarantee that every model launches everywhere, but it would help companies distinguish genuine compliance barriers from ordinary commercial choices. That clarity is essential if the AI Act is to protect European citizens while leaving local innovators able to compete.
