Brussels Prices Platform Power by Adjacency

The General Court, on 9 September, dismissed Booking Holdings’ challenge to the Commission’s 2023 veto of its €1.63 billion purchase of Etraveli Group (T-1139/23, case M.10615). Seven days earlier it dismissed Opera Norway’s bid to put Microsoft Edge on the Digital Markets Act gatekeeper list (T-357/24). The wire read both as Europe versus platforms. That is the dominant file, and it is incomplete.

The residual is the causal chain in Booking, not a slogan about “ecosystems.” If the parties barely overlapped in hotel search, why was buying a flight OTA a significant impediment to effective competition rather than a convenient add-on?

Reverse leverage, not a hotel overlap

Booking was already dominant in EEA hotel online travel agencies. Etraveli was a flight specialist. The Commission’s harm was not that they competed head to head for the same rooms. It was that flights were one of the few customer-acquisition channels Booking did not yet dominate. Cross-selling rooms to flight customers — the “connected trip” — would grow hotel share through a channel rivals could not cheaply copy once Booking led both sides.

The Court allowed a “reverse” leveraging theory: a non-dominant flight position used to entrench an already dominant hotel position. The 2008 non-horizontal guidelines did not forbid that, it said, even if the Commission had little digital experience when it wrote them. It also said “competition on the merits” from Article 102 does not immunize a merger. A concentration is judged by structure, not by later courtesy.

The quantitative increment could be “a few tenths of a per cent.” The Court found the Commission’s share-growth arithmetic flawed, and it did not accept that hotels would dump rival OTAs or that commissions would rise. Strengthening dominance is not, by itself, a SIEC. What closed the case were two structural facts: strong network effects with a wide gap to the next firm, so a small increment still thickens the lead; and growth through a scarce channel that would make the travel stack hard to replicate, hitting OTAs that still matter to hotel competition. Efficiencies did not offset that.

Pre-installation is distribution, not a gateway

Edge already showed the other test before judgment week. Opera lost on 2 September. Edge met DMA size thresholds. The Court still accepted that it is not an “important gateway” for business users to reach end users. The Commission could rely on usage — about 5.8 percent of European page views in the data it used — and on Blink as the engine, which limits Microsoft’s autonomous control. Windows pre-installation and taskbar promotion did not convert into gateway status. Rivals without that placement matched or beat Edge’s scale. Thresholds are a presumption. Use can rebut them.

That is a different statute from merger control. DMA designation asks whether this core platform service is, now, a gateway. The Merger Regulation asks what acquiring an adjacent channel would do to contestability. Google’s DMA fight is about duties on a designated service, not about whether a default is a service. Treating both Luxembourg days as “scale” is a category error. Brussels is writing other rules without waiting for Washington. These two files are how it writes platform files.

The flight OTA was blocked because it was a customer channel into hotels. The browser walked because users did not walk through it.

What still mis-prices is a playbook that says dominant firms cannot buy adjacencies, or that Windows defaults are automatically DMA-bound. The test is dated and specific: whether the next deal is a scarce acquisition funnel under network effects — and whether the next CPS rebuttal shows actual use, not placement. Watch the revised Merger Guidelines’ “entrenchment” language this quarter, and any appeal in T-1139/23. Those, not a size league table, are the next constraint.