Apple Holds Back AI Siri in the EU and China, Citing Regulatory Friction

Apple confirmed this week that its redesigned, AI-powered Siri will skip the EU and mainland China at launch, blaming regulatory requirements in both markets. The assistant, unveiled Tuesday, pulls context from a user's screen, messages, emails, and photos to answer questions, a shift from the command-and-response tool Siri has been for over a decade. An English-language beta is due later this year, following an initial developer preview.

The interesting part isn't the feature set. It's the map of where Apple is willing to ship it.

What the new Siri actually does

The pitch is personal context. Instead of treating each query as a standalone request, the upgraded Siri reads what is already on your device and reasons across it. Ask about a flight and it can reference the confirmation email. Ask who someone is and it can pull from your messages and photos. This is the on-device, personal-data version of the assistant idea that every large platform has been chasing, and it depends on deep, privileged access to the things a user already stores on their phone.

That access is exactly what makes the rollout politically complicated.

The EU problem: the DMA

In Europe, the friction traces back to the Digital Markets Act, the EU rulebook that designates large platforms as "gatekeepers" and forces them to open up. The DMA requires Apple to support alternative payment systems and third-party app stores, and it pushes for interoperability that cuts against Apple's tightly controlled model.

Apple has criticized parts of the DMA repeatedly, arguing that the openness mandates create privacy and security risks. The European Commission has not blinked. It said it would neither repeal nor amend the DMA in response to Apple's objections, which leaves Apple to do the adjusting if it wants to ship in the bloc. An assistant that reaches across apps and system data is precisely the kind of feature where interoperability rules and Apple's preferred architecture collide, so holding it back buys the company time to sort out compliance rather than launch something it might have to re-engineer.

The China problem looks different

China is a separate calculation. Apple has not detailed the specific requirements there, but the broader pattern is well known: generative AI services in mainland China face approval processes and content rules, and foreign firms typically need a local partner to supply the underlying models. Reporting has pointed to Apple working with domestic companies to meet those conditions. A delay here reads less like a standoff and more like an unfinished negotiation over who provides the intelligence layer and how it gets cleared.

Why this matters beyond Apple

For years the assumption was that a flagship feature ships globally on day one, with minor localization. That assumption is breaking. The most capable version of a consumer AI product now arrives first in the markets with the lightest platform regulation, and later, in modified form, everywhere else. Users in two of the world's largest economies are being told to wait, not because the technology isn't ready, but because the rules around platform power and data are still being settled.

That creates a real divergence in product experience tied to geography. It also hands a small opening to competitors and local players who are already operating inside those regulatory regimes. In China especially, every quarter Apple spends working out compliance is a quarter domestic assistants keep the field. The companies that figure out how to ship advanced AI features under strict platform rules, rather than around them, will have an advantage that has nothing to do with model quality.

Apple's bet is that the markets it is delaying are worth the wait, and that regulators will eventually accommodate a workable version. The European Commission's flat refusal to amend the DMA suggests the accommodation, if it comes, will be on the regulator's terms. [Reported by Jiwei, in Chinese.]