The European Commission fined Google 890 million euros on 23 July, split into two decisions: 460 million for self-preferencing its own services in Google Search, and 430 million for restricting businesses from steering consumers toward cheaper purchase channels on Google Play. It is Google's first penalty under the Digital Markets Act and the largest issued under the regime to date.

Against Alphabet's balance sheet the sum is close to a rounding error. The company reported capital expenditure of roughly 44 billion dollars in a single recent quarter and has raised its 2026 spending guidance to between 195 and 205 billion dollars. A one-billion-dollar fine is roughly two days of capital expenditure.

Which is why reading this as a fine misses what happened. The financial penalty is the smallest instrument in the decision. The operative part is a 60-day compliance clock, backed by periodic penalty payments of up to 5 percent of Alphabet's average daily worldwide turnover for continued non-compliance. That mechanism does not have a ceiling in any practical sense. It is designed to make refusal more expensive than compliance, indefinitely.

What the Commission actually found

Two things, and the distinction between them matters for anyone assessing their own exposure.

On Search, the Commission found that Google gave preferential treatment to its own services in areas including shopping, hotels and transport, placing them more prominently than third-party equivalents and supporting them with enhanced visuals and filters. Under the DMA, designated gatekeepers may not rank their own services more favourably than those of third parties.

On Play, the Commission found that Google prevented app developers from promoting offers and entering into contracts with customers through alternative distribution channels. The rule the finding rests on is that developers must be able to inform users of alternative purchase channels free of charge.

Note the shape of both findings. Neither turns on proving consumer harm, market definition or the competitive effect of the conduct, which is the analytical machinery of traditional antitrust and the reason traditional antitrust cases take the better part of a decade. The DMA operates as a conduct code. The conduct is either compliant or it is not.

That is the structural innovation, and it explains the timeline. Google was designated a gatekeeper in 2023. The Commission opened its non-compliance investigation in 2024. The decision landed in 2026. Three years from designation to penalty is unrecognisably fast by the standards of the cases that preceded it.

The two arguments, stated fairly

Google's position, articulated by Kent Walker, its president of global affairs, is that compliance degrades the product. The company says it is having to strip real-time search features that European users value, including instant pricing and availability for hotels, flights and restaurants, and that removing Play Store restrictions on external steering introduces security risk by directing users to third-party sites. Google has also argued the search changes could harm travel businesses that acquire customers through its platform. It is reviewing the decisions and weighing an appeal.

The Commission's position, put by competition chief Teresa Ribera, is that the best products should succeed on merit rather than on ownership of the search engine that ranks them. Tech sovereignty commissioner Henna Virkkunen framed the decisions as safeguarding the business and innovation opportunities the DMA was designed to open.

Both positions contain a real claim. Integration genuinely does produce features that are hard to replicate across a market boundary, and instant hotel pricing inside a search result is a better user experience than a link to a page that loads one. It is also true that the entity deciding which results appear first has an interest in the outcome, and that no amount of good faith resolves the conflict structurally.

What is not yet established is which effect dominates in practice. The 60-day window will produce the first real evidence, because Europe is about to run the experiment on live traffic.

Why non-gatekeepers should read this closely

The DMA's obligations bind a designated list. Its logic does not stay there, for three reasons.

The conduct code is portable. Self-preferencing and anti-steering are not exotic behaviours confined to search engines and app stores. Any business operating a platform where third parties compete alongside the operator's own offering runs the same structure: marketplaces, booking systems, enterprise software with an app ecosystem, distributors with private label. The DMA has now produced an enforced definition of what those arrangements look like when they cross a line, and that definition will be cited in national competition proceedings, in commercial disputes and in contract negotiations by counterparties who now have language for what they object to.

Compliance is being built once and deployed unevenly. Google has said the changes apply in Europe. That is the standard posture and it rarely holds cleanly, because maintaining two materially different product architectures is expensive and because the engineering work, once done, is cheap to extend. Businesses whose customer acquisition depends on a gatekeeper's ranking should assume the European version of that ranking is a preview rather than a regional variant.

The dependency itself is the finding. Strip out the legal analysis and the decision documents something operational: a large number of businesses acquire customers through a channel whose ranking rules are set by a company competing with them, and those rules can change on a 60-day regulatory clock. Travel businesses are the named example, and Google's own defence rests partly on the disruption that changing the rules will cause them, which is an unusually direct admission of how much revenue sits on that dependency.

If a meaningful share of your pipeline arrives through a single platform's ranking, the DMA has not created your risk. It has made the timing of it legible.

What happens next

Three things are worth tracking over the coming quarter.

The compliance filings. What Google actually removes or restructures in European Search and Play by late September will define the practical content of these obligations far more than the decision text does. Watch whether the removed features are genuinely load-bearing or whether their removal functions as an argument.

The appeal. Google is evaluating one. An appeal does not suspend the compliance obligation, which means the company will likely comply and litigate simultaneously, and the European courts will take years to rule on a regime that will have moved on by then.

The trade dimension. The decisions drew immediate criticism from Washington, where the administration has previously threatened additional tariffs in response to European digital regulation. This lands during an already strained transatlantic trade relationship and alongside separate negotiations touching digital trade rules. Whether digital enforcement becomes a bargaining chip in a broader trade settlement is the genuine open question here, and it is a political one rather than a legal one.

The read

For Alphabet, this is a manageable cost and a product argument the company intends to keep making.

For everyone else, the useful takeaway has nothing to do with the money. A regulator has demonstrated that it can compel a specific architectural change in a dominant platform within roughly three years of designating it, using a conduct code rather than an effects analysis, enforced by a penalty mechanism with no practical ceiling.

That is a new capability, and it is being tested on the largest possible target first. The question for any business built on top of a platform is not whether Brussels wins this one. It is what your operating model looks like if the ranking rules underneath it change on someone else's schedule.

Share thisFacebookXLinkedIn