Skip to content
Social Rocket

Prebid Summit 2026: Why Ad Targeting Is Becoming an Infrastructure Question Again

Safari blocks, Google's Prebid obligations and new universal media platforms: the Prebid Summit shows why publishers need to reorganise their ad data and measurement.

From Social Rocket editorial team

This text was created with AI assistance and reviewed editorially before publication. All sources are linked below.

Server racks as a symbol of programmatic ad infrastructure
Photo: Kevin Ache / Unsplash

At the Prebid Summit on 13 October 2026, the question moved to the centre: who still controls audience and measurement signals on the open web at all? While browsers and operating systems restrict third-party calls, pressure is growing to organise auctions and permissible first-party data on the publisher side. At the same time, following the US adtech ruling, Google must support integrations with Prebid. For German publishers, this is a technical and economic question – not a return to unrestricted cookie tracking.

Three layers that the targeting debate often mixes up

Delivery first means that an ad is loaded. Addressability means whether an advertiser can identify a suitable audience based on permissible signals. Measurement finally asks whether a contact contributes to a result. If an identity provider fails, not every ad has to disappear; if a delivery endpoint is blocked, however, gaps in delivery are also conceivable.

The difference became visible through the Safari 27 case. There, blocking adtech domains triggered a new industry round on dependence on browser and operating system providers. Important: the specifically documented The Trade Desk bug was confirmed by WebKit as fixed on 9 October. The other tracking and privacy questions are thereby not resolved. Our updated report on the Safari 27 case and IAB Trusted Server separates both.

What publishers can actually change

The open header bidding project Prebid Server moves parts of the auction into a server environment. This creates different control options than an ad stack controlled purely by JavaScript in the browser. In addition, the official Prebid documentation on first-party data describes how publishers can provide contextual or user data only to certain bidders. Technically, however, control is not automatically the same everywhere: not every adapter supports the same conventions.

A further architectural approach is pursued by the IAB Tech Lab with Trusted Server: ad calls, parts of the auction and functions of various service providers are to come together in an open-source server runtime controlled by the publisher. The project had a prototype publishers could test in early 2026. The advantages – fewer client-side dependencies and more control – are goals and technical possibilities, not yet broadly proven additional revenue.

Server-to-server auctions do not resolve privacy requirements on their own. The Prebid privacy documentation provides, among other things, rules for consents, GPP and GDPR-related control for this. What remains decisive is which data may actually be transmitted.

Google must open up to Prebid – but impact takes time

A second development comes from competition law. The US Department of Justice confirmed in September that Google must provide integrations of AdX and DFP with Prebid as well as further data access and protection against self-preferencing. These are ordered obligations, not interfaces already fully implemented. The possible competitive effect can only be measured after implementation.

Our detailed classification of the Google adtech ruling and the Prebid obligations explains the legal framework. For Germany, European data protection and the separate EU competition proceedings apply independently of the US ruling.

"Universal Media Platform": a new word is not yet a new market

At the Summit, industry representatives also discussed, according to Digiday, new labels for large cross-channel buying platforms, such as "Universal Media Platform". The scepticism on the panel concerned less the name than the question of what these offers practically add compared with existing ad networks. Uniform buying across several channels guarantees neither independent measurement nor less dependence on proprietary data.

Advertisers should therefore ask specific questions: Can bids and fees be traced? Are there data exports and independent control groups? And are there differences between open-web inventory and walled gardens? The heightened debate around AI agents, advertising budgets and cost audits makes these questions relevant even without new platform labels.

What German teams should test now

For publishers, a limited server-side test makes more sense than an uncontrolled complete overhaul: separate inventory by browser and operating system, document technical dependencies, check consent and data flows, and then compare yield, latency, match rates, measurable advertising impact and operating costs. For advertisers, it also matters whether the measurement approach enables better selection – rather than merely delivering more identifiable IDs.

Context: The Prebid Summit did not present a new universal targeting cure-all. The measurable trend is rather that browser rules, publisher data sovereignty and adtech competition are all affecting the advertising stack at the same time. Those who separate these layers can assess investments more reliably.

Image note: Stock photo of a server room, not a photo from the Prebid Summit. Photo: Kevin Ache / Unsplash.

Sources: Digiday, 15.10.2026; Prebid Server: Architecture; Prebid: First-Party Data; IAB Tech Lab: Trusted Server; US Department of Justice, 16.09.2026.