Europe's Digital Markets Act just landed three big blows on Apple and Google
In just over two weeks last July, the EU's tech law forced Apple to swallow a court defeat, ordered Google to share its search data with rivals and even AI chatbots, and hit Google with its first fine under the act. Here is what actually happened, why the three events are different, and why it could reshape how Big Tech operates.
The EU's Digital Markets Act (DMA) had a decisive July 2026. On 8 July, a court dismissed Apple's challenge to its "gatekeeper" status and set a precedent, constraining every gatekeeper, that they cannot fight the law's obligations in the abstract. On 16 July, the European Commission told Google exactly how to comply, including sharing anonymised search data with rival engines and AI chatbots, and opening 11 Android features to competing AI assistants. On 23 July, separately, it fined Google €890 million, its first DMA fine. Together they turn the DMA from a law on paper into an enforced regime, one starting to reshape how the two most consumer-facing platforms work, with a real edge for the AI race.
For a couple of years, the EU's Digital Markets Act was easy to dismiss as Brussels bureaucracy: a big law with grand ambitions and not much visible bite. July 2026 ended that. In the space of about two weeks, the DMA delivered a court win against Apple, a detailed order telling Google to open up its crown jewels, and a nine-figure fine. The three are genuinely different kinds of action and are easy to blur together, so here is the clean version of what happened and why it matters well beyond Europe.
What is the DMA, and who does it target?
The Digital Markets Act is a 2022 EU law (in force since late 2022) aimed at a handful of dominant tech firms it calls "gatekeepers." The idea is simple: a few platforms have become unavoidable gateways between businesses and customers, so the DMA imposes rules on them that ordinary competition law was too slow to enforce. Gatekeepers cannot favour their own services over rivals ("self-preferencing"), must let users and developers reach outside their walled gardens, and must open certain systems up so competitors can interoperate.
The original six gatekeepers, designated in September 2023, are Alphabet (Google), Amazon, Apple, ByteDance, Meta and Microsoft. July's actions hit the two biggest: Apple and Google.
Blow one: Apple lost its court challenge (8 July)
Apple had gone to court to fight its designation, arguing among other things that its App Store and iOS should not be treated as gatekeeper services. On 8 July 2026, the General Court of the EU dismissed the challenge and upheld both designations. (A separate Apple claim, over the Commission's decision not to designate iMessage, was thrown out as inadmissible.) Apple remains a gatekeeper, bound by the full weight of the DMA's obligations.
Two caveats keep this honest. First, the General Court is the EU's lower court; Apple can still appeal to the Court of Justice on points of law, so this is a heavy defeat but not the final word. Second, the ruling's most consequential part was procedural: the Court held that gatekeepers cannot challenge the DMA's obligations "in the abstract," and must instead wait for a specific enforcement action to contest it. That principle does not just bind Apple. As a practical matter it narrows the legal room every gatekeeper has to resist the law, which is why the whole industry read an Apple case closely.
Blow two: Google was told exactly how to comply (16 July)
A week later, on 16 July 2026, the European Commission issued two binding "specification" decisions telling Google precisely what compliance must look like. These are guidance, not punishments, and they are the ones with the biggest long-term consequences, especially for AI.
- Google must share its search data. Google has to give rival search engines, and pointedly AI chatbots that offer search functions, access to anonymised ranking, query, click and view data from its search results, on fair, reasonable and non-discriminatory (FRAND) terms. This is not Google's raw index or users' personal data; it is the aggregated signal that makes Google's results good, and sharing it could let competitors and AI answer engines close part of the quality gap. The Commission's phased timeline runs toward the data actually flowing around January 2027.
- Google must open Android to rival AI. Third-party AI assistants must get real access to 11 Android features, from voice activation ("Hey Google"-style wake words) to completing actions across apps and reaching device sensors and on-device AI models. In other words, a competing assistant should be able to do on your Android phone what Google's own can. Google must implement this in a coming Android release, by 1 August 2027 at the latest.
Read together, these are aimed squarely at the next platform war. The EU is trying to make sure Google cannot use its grip on search and Android to lock up the AI-assistant era before it starts.
Blow three: Google was fined €890 million (23 July)
Then, separately, on 23 July 2026, the Commission fined Google €890 million, its first fine under the DMA. This is a different instrument from the 16 July guidance: it punishes past conduct rather than specifying future compliance. The total splits into €460 million for self-preferencing in Google Search (favouring its own shopping, hotel, transport and sports results over rivals') and €430 million for Play Store rules that stopped developers steering users to cheaper deals outside the store. Google was ordered to stop the conduct within 60 days or face escalating penalties.
Keep the number in perspective in both directions. €890 million (about $1 billion) is a large, headline-grabbing fine, Google's first under the DMA and the largest the act has produced so far (the very first DMA fines hit Apple and Meta back in April 2025). It is also a small fraction of what the DMA in theory allows: penalties can reach 10% of a company's worldwide annual turnover, and 20% for repeat offenders, which for Alphabet would be on the order of tens of billions of dollars. The fine is a warning shot, not the ceiling.
Why it matters
Individually, each of these could be shrugged off. Together they mark the moment the DMA became a real, enforced regime rather than a threat, and the effects reach well past Europe.
The clearest stakes are in AI. Forcing Google to share search data with "AI chatbots" and to open Android to rival assistants is an explicit attempt to stop the company from carrying its search-and-mobile dominance straight into the assistant era. Whether it works will shape how competitive AI answer engines can be. Beyond that, the Apple ruling makes the law far harder to fight in court, the fine shows the Commission will punish as well as instruct, and because these platforms are global, changes forced in Europe often ripple outward. It all lands amid ongoing friction between Brussels and Washington over regulating American tech giants, which guarantees the fight is not over. But the direction is now unmistakable: in the EU, the terms on which Apple and Google operate are increasingly set in Brussels, not Cupertino or Mountain View. For more, see the Software section and our coverage of the AI race these rules are trying to keep open.
The DMA's July, at a glance
| Date | Action | What it does |
|---|---|---|
| 8 July 2026 | Apple loses court appeal (General Court) | Upholds Apple's gatekeeper status; rules gatekeepers can't fight the DMA "in the abstract" (appeal to the top court still possible) |
| 16 July 2026 | Two Google "specification" decisions (guidance) | Google must share anonymised search data with rivals and AI chatbots (~Jan 2027) and open 11 Android features to rival AI (by 1 Aug 2027) |
| 23 July 2026 | Google fined €890M (Google's first DMA fine, largest to date) | €460M for search self-preferencing + €430M for Play Store anti-steering; stop within 60 days |
'Vibe coding' let anyone build an app. Now the security bill is coming due
AI tools now let people describe an app in plain English and ship it without writing code, a practice nicknamed 'vibe coding.' It is genuinely empowering, and it is producing a wave of live apps riddled with basic, avoidable security holes. Here is what the scans are finding, why AI-written code fails this way, and what it does not mean.
Meta just entered the AI coding-agent wars. The big two are now a big three
With Muse Code, Meta has launched its own AI agent that writes, tests and ships software from your terminal, going straight at Anthropic's Claude Code and OpenAI's Codex. Its headline trick is running many AI coders in parallel without them tripping over each other. Here is what it is, what's genuinely new, and the honest caveats.
Even Linux is embracing AI now, and drowning in it at the same time
Linus Torvalds has made it official: arguably the world's most important open-source project 'is not one of those anti-AI projects.' At the same time, the volunteers who maintain the Linux kernel are being buried under a flood of AI-generated bug reports and patches. Both things are true, and together they are the story of open source in 2026.