Who owns the identity layer?

The identity layer is the connective tissue of programmatic advertising. Identity graphs resolve a person's devices, email addresses, and household into one addressable record, and every audience segment, match, and cross-device measurement pass runs through one of them. Whoever operates the graph decides how matching works, sets the terms of access, and sees a detailed picture of the plans that run through it.

For roughly a decade, the companies operating this layer positioned themselves as neutral intermediaries between agencies, platforms, and supply. Neutrality was the core of the offer. A graph owner with no media to sell and no competing agency to feed could be treated as disinterested infrastructure, and buyers planned accordingly.

That structure has changed. The largest identity assets now sit inside agency holding companies, credit bureaus, and media platforms, each of which operates other lines of business alongside the graph. This article maps the ownership as reported, as of mid-2026.

Four transactions define the consolidation. In 2019, Publicis Groupe acquired Epsilon for $4.4 billion, bringing roughly 2.3 billion consumer profiles inside a holding company (Publicis Groupe, 2019). In 2021, TransUnion acquired Neustar for $3.1 billion, moving a major ad-identity business inside a credit bureau (TransUnion, 2021).

The Omnicom and IPG merger, completed across 2024 to 2025, gives Omnicom the Acxiom business and roughly 2.6 billion identities (Omnicom/IPG, 2025). Acxiom arrived through a merger rather than a priced acquisition, so it carries no standalone price line. It is best read as scale.

The fourth move is pending. In 2025, Publicis agreed to acquire LiveRamp for roughly $2.2 billion (Publicis Groupe, 2025). LiveRamp operates the interoperability rail much of the open ecosystem matches through, and Publicis has stated the business will remain neutral. As of mid-2026 the deal has not closed; completion is expected before the end of 2026.

Publicis also folded Lotame into Epsilon in 2025. The combined graph holds around 4 billion profiles, near 91% of adult internet users, per Publicis and eMarketer (2025). The reach matters more than the purchase prices. Once the pending deal closes, one holding company would operate a graph of that scale alongside the media business it can be activated against.

The consolidation ran on two mechanisms. The first is acquisition. Holding companies bought graphs: Publicis with Epsilon, Lotame, and the pending LiveRamp agreement; Omnicom with Acxiom via the IPG merger; WPP with the InfoSum clean room (2025). Credit bureaus did the same, with TransUnion acquiring Neustar and Experian acquiring the Tapad cross-device graph. An acquired graph changes hands in public, and in principle it can change hands again.

The second mechanism is quieter. Comcast built FreeWheel's deterministic household graph on its set-top and broadband relationships, and Blockgraph, co-owned by Comcast, Charter, and Paramount, extends that into a shared cross-screen layer with reported reach of roughly 16 million TV subscribers (Blockgraph, reported; verified June 2026, directional). These graphs are anchored to subscriber relationships and screens that are not for sale, which makes built ownership the more durable kind.

Built graphs carry a governance question of their own. As of mid-2026, reported merger activity places Blockgraph's co-owners on opposite sides of a potential deal, with Paramount in a reported combination involving Warner Bros. Discovery and Comcast reported as a separate bidder (Reuters; AdExchanger). A graph governed jointly by companies engaged in merger negotiations has an unsettled ownership picture, and buyers relying on it should track how that resolves.

A middle of independents remains. The Trade Desk maintains UID2 as an open identity standard outside holding-company ownership, and Magnite, the largest independent sell-side platform, is building identity from the supply side. The independent list is shorter than it was in 2019.

The table below is the map worth keeping. When a plan activates a segment, the practical question is which parent the data asset reports to and what other lines of business that parent operates. Ownership does not make a data asset worse; it changes the incentive structure around how the asset is priced, accessed, and recommended.

Two rows carry caveats. LiveRamp appears under Publicis on the strength of a signed agreement that had not closed as of mid-2026, and Blockgraph's shared ownership is subject to the merger activity described above.

A buyer can surface most of this structure with five items of diligence, run the same way a careful buyer traces where streaming inventory comes from. None of them is a reason to walk away from a partner. They establish what is being bought alongside the data.

First, establish who ultimately owns the graph, traced past the product name to the parent company. Second, establish whether that owner also sells media the buyer purchases; a graph that sits inside a media seller sits next to an inventory business with its own targets.

Third, distinguish neutrality that is contractual from neutrality promised at announcement. A commitment that a business stays independent carries different weight in a signed contract than in a press release. Fourth, establish whose first-party data built the graph; provenance is a privacy question before it is a targeting one.

Fifth, and the item that decides the rest: establish who the graph serves if the parent's interest and the buyer's interest diverge. The answer to that question prices the ownership structure into the plan.

For independent agencies and in-house teams, the response is operational. There is also an added dimension for agencies specifically: the holding companies that own the largest graphs also own agencies that compete for the same accounts, so the ownership question sits closer to home. Four standing practices cover most of it.

Route plans toward supply paths with fewer intermediaries and clear data provenance. Require every segment to be named before activation, documenting whose graph built it and which parent that graph reports to, ahead of the buy rather than after the invoice.

Stay interoperable. Working across independent and holding-company graphs alike preserves leverage and avoids lock-in to a single rail or a single owner's data. Finally, measure independently, preferring holdout and clean-room methods over reporting produced by the same party that sold the media.

MediaPath runs programmatic as a managed service for independent agencies. The firm owns no identity graph and no media properties, so no plan it builds has a house asset waiting for the budget. That is a structural fact rather than a promise, and it is why the diligence above can be applied to every plan without exception.

In practice, this means naming whose graph each segment activates through before a plan runs, working across independent and holding-company rails as the plan requires, and reading results through independent measurement. The consolidation mapped above does not have to compromise a plan. It does require knowing whose data the plan activates, and as of 2026 that is part of the job.