How the Digital Markets Act Is Unbundling Big Tech
For two decades, vertical integration was the whole strategy — hardware, OS, app store, payments, and search reinforcing each other. The DMA is an attempt to reverse-engineer that, and the platforms are already relocating the moat somewhere harder to regulate.

AI Overview
The European Union's Digital Markets Act regulates architecture rather than behaviour. Instead of proving harm case by case, it designates large platforms as gatekeepers by objective thresholds and imposes structural obligations in advance: users must be able to uninstall pre-installed apps and change defaults, platforms must not rank their own services above rivals, core hardware and software features such as NFC chips and messaging protocols must be opened to third parties, and personal data cannot be combined across separate services without explicit consent. Each obligation targets a specific mechanism of vertical integration — distribution through defaults, lock-in through walled gardens, and compounding advantage through pooled data. The measurable result so far is partial. Technical unbundling has occurred, but compliance complexity favours firms with large legal and engineering teams, and defensibility is migrating toward proprietary AI infrastructure and enterprise cloud integration, which interoperability rules do not reach.
In short
- What it is: a structural regime that regulates how large platforms are built, not just how they behave.
- Why it matters: it directly attacks defaults, bundling, and data pooling — the three mechanisms that made platform moats durable.
- Who benefits: developers and competing services gaining distribution access; consumers gaining lower switching costs.
- Biggest takeaway: the moat is relocating rather than disappearing, toward compute and AI infrastructure that unbundling cannot touch.
Key Facts
| Category | Economics |
| Difficulty | Intermediate |
| Read time | 7 minutes |
| Search intent | Informational |
| Updated | August 17, 2026 |
For over two decades, the core business model of large technology platforms relied on vertical integration and proprietary lock-in: tightly coupled ecosystems where hardware, operating systems, app stores, payment rails, and default search engines reinforced one another.
The Digital Markets Act is an attempt to reverse-engineer that architecture. Rather than litigating individual abuses after the fact, it designates gatekeepers by objective criteria and regulates the structure itself (Regulation (EU) 2022/1925).
Why it matters
Competition law spent thirty years struggling with services that are free at the point of use. A consumer-price test finds nothing wrong with a search engine that costs nothing, which is why traditional enforcement kept arriving years late and one market behind.
The DMA changes the question from did this firm harm consumers to is this firm structured in a way that forecloses competition. What it is really regulating is the durability of an economic moat built from distribution rather than from capability. That shift matters well beyond Europe, because compliance is largely architectural, and architecture is expensive to maintain in two versions.
1. Stripping away the default advantage
The most powerful defence a platform possesses is friction-free distribution through defaults. When an operating system routes a user automatically to its own browser, app store, payment processor, or search engine, it creates a structural advantage that superior competing products struggle to overcome.
The DMA forces gatekeepers to unbundle these layers through two mechanisms:
- Choice screens and uninstallation — users must be able to remove pre-installed apps and select alternative default browsers or search engines.
- Anti-self-preferencing — platforms may not rank their own services above competitors' in search results or marketplaces.
Removing the automatic gravity of defaults forces competition on the merit of the individual service rather than the weight of the operating system beneath it.
2. Fracturing the walled garden
Ecosystem defensibility depends on network effects and high switching costs. An iPhone user who leaves iOS loses compatibility with accessories, cloud storage, and messaging continuity — a cost paid at exit, which is precisely what makes it effective.
The DMA attacks this through technical unbundling: sideloading and alternative app stores break absolute control over software distribution, while interoperability requirements open underlying hardware and software features such as NFC chips, secure elements, and messaging protocols to third parties on equal terms. Apple's own compliance announcement documents the scope of the resulting changes to iOS, Safari, and the App Store (Apple Newsroom).
The strategic effect is on switching costs. When alternative stores can offer different developer terms, and messaging can cross network boundaries, the ecosystem stops functioning as a trap and starts functioning as a marketplace.
3. The data silo dilemma
Data network effects are the third pillar: more users generate more data, which trains better systems, which attracts more users. The DMA restricts how gatekeepers combine personal data across distinct core services — merging social network data with an advertising network or messaging service, for instance — without explicit, freely given consent.
This is the provision with the longest reach. It is aimed at preventing a firm from converting dominance in one market into an unassailable data advantage in an adjacent emerging one, which in practice means model training and localised advertising technology.
The Moat Relocation Model
The useful way to read the DMA's results is not as a scoreboard of compliance but as a map of where defensibility moves when a specific layer is regulated. Each mechanism the law dismantles has a natural destination.
| Layer under attack | DMA mechanism | Where defensibility relocates | Reachable by unbundling? |
|---|---|---|---|
| Distribution via defaults | Choice screens, uninstallation | Brand preference and habit | Partly |
| App store control | Sideloading, alternative stores | Security review, fee architecture | Partly |
| Hardware lock-in | NFC and secure-element access | Silicon design and performance | No |
| Pooled personal data | Consent requirements for combining | First-party data and licensed corpora | No |
| Integrated services | Anti-self-preferencing | Proprietary AI infrastructure, compute scale | No |
The pattern in the right-hand column is the finding. Obligations bite hardest on distribution layers, which are contractual and therefore legible to regulation, and barely at all on capability layers, which are capital-intensive and physical. A rule can compel a company to open an NFC chip. It cannot compel it to share a compute cluster or the engineering depth to use one.
That is why the compliance era has coincided with aggressive investment in exactly those layers. The moat did not evaporate; it moved to higher ground.
Market Analysis
The compliance paradox is the clearest result so far, and there is an irony in it. Complying with the DMA requires substantial engineering and legal capacity, which smaller firms do not have. Obligations written to constrain the largest companies are, in part, operationalised on their own terms — through fee structures, integrity requirements, and security guardrails that preserve much of the original economics while satisfying the text.
This is not evidence that the law failed. It is evidence that structural regulation of complex systems tends to produce complex compliance, and that complexity is itself a barrier to entry.
Limitations
Assessment here is necessarily provisional. Enforcement is ongoing, several specification decisions remain contested, and the counterfactual — what competition would look like without the DMA — is unobservable.
The moat relocation model above is an analytical framework rather than a measured finding. It organises observable behaviour into a pattern; it does not prove causation, and some of the investment it points to would have happened regardless.
The scope is also jurisdictional. Obligations apply to designated gatekeepers operating in the European Union, so global effects depend on whether firms maintain one architecture or two — a commercial decision, not a legal one.
References
- Digital Markets Act — European Commission — official hub, designations and enforcement
- Regulation (EU) 2022/1925 — EUR-Lex — the regulation's full legal text
- About the DMA — European Commission — scope, gatekeeper criteria, obligations
- Commission press material on DMA enforcement — non-compliance investigations
- Apple announces changes to iOS, Safari and the App Store in the EU — a gatekeeper's compliance in its own words
Explore related coverage
- Back to the economics hub
- The platform didn't beat the hotel — the balance sheet did
- The orchestration layer is the new moat
- The data moat in the AI era
- Top 10 chip bottlenecks and the AI moat
Final thoughts
The DMA treats digital platforms less like sovereign private kingdoms and more like public infrastructure. By forcing them to unbundle their stacks, the European Union is betting that contestability shifts competition away from who controls the gateway and toward who builds the best product.
The early evidence suggests a more complicated outcome. Competition at the gateway has genuinely opened, and simultaneously the decisive layer has moved somewhere the gateway rules do not apply. Regulating architecture works — but architecture is not fixed, and the firms being regulated redesign faster than the law that describes them.
What does the Digital Markets Act actually require gatekeepers to do?+
It imposes structural obligations rather than case-by-case findings. Designated gatekeepers must let users uninstall pre-installed apps and change defaults, must stop ranking their own services above competitors, must open core hardware and software features such as NFC and messaging to third parties on equal terms, and must not combine personal data across separate services without explicit consent.
Why is the DMA different from traditional antitrust enforcement?+
Traditional antitrust is retrospective and case-driven: a regulator proves harm after the fact, usually via consumer prices. The DMA is prospective and rule-based. It designates gatekeepers by objective thresholds and imposes obligations in advance, which sidesteps the problem that platform services are often free and therefore invisible to a price-centred test.
Has the DMA made it easier for smaller competitors to compete?+
Partially, and unevenly. Choice screens and alternative app stores lower distribution barriers, but compliance itself demands substantial legal and engineering capacity. Gatekeepers have also implemented obligations through fee structures and security requirements that preserve much of the original economics, so the practical gain for smaller firms has been narrower than the text implies.
Where is platform defensibility moving as software lock-in is dismantled?+
Toward layers that unbundling does not reach: proprietary model training and inference infrastructure, ownership of large compute clusters, and deep enterprise cloud integration. These are capital and capability moats rather than distribution moats, which makes them considerably harder to address with interoperability mandates.