AdExchanger published it at 12:35 in the morning, an hour normally reserved for confessions, DoorDash, and texts you regret. The column is called "Ad Tech's Last Neutral Party Just Picked A Side." The byline is Jonathon Shaevitz, co-founder of GrowthCode.

Every fact in it is right.

Publicis really is paying an enterprise value of $2.167 billion for LiveRamp at $38.50 a share, a 29.8% premium. Arthur Sadoun really did wave it off as a non-event because the technology is "neutral by design," a phrase with all the reassurance of a landlord telling you the building is fireproof.

John Wren really did yank his exit forward.

Cindy Rose really did answer "What do you think?" in Cannes, which is the most elegant public middle finger the agency business has produced in a decade.

Good column. Also a brochure with footnotes.

Here is the number that should have been the headline everywhere and was the headline nowhere.

On May 19, at the J.P. Morgan tech and media conference, an analyst asked Wren the obvious question: your data is about to sit inside a competitor, is that a problem? Wren did not give a statesman's answer. He gave an accountant's.

The LiveRamp contract, he said, is mutual and nets to zero. LiveRamp pays Omnicom roughly $50 million a year for data. Omnicom pays roughly $50 million a year back for services it could get somewhere else if it felt like it.

That is not a utility and a customer. That is two exes who still share a Costco membership because neither wants to be the one who cancels it.

Acxiom and LiveRamp were the same company until 2018, when Acxiom sold Marketing Solutions to IPG for $2.3 billion and spun LiveRamp out to fend for itself. Omnicom inherited the Acxiom half when it swallowed IPG. And Wren, entirely unprompted, explained why LiveRamp had to be "Switzerland" in the first place: because it could not be folded into Acxiom and still be worth anything to a client.

So neutrality was never a virtue LiveRamp practiced. It was a custody arrangement. Switzerland wasn't a country. It was the spare bedroom in the divorce decree, and everybody in the family knew it.

Acxiom spent five years quietly building Real ID, its own identity graph, which Omnicom says resolves around 2.6 billion people and 98% of addressable US adults. The LiveRamp contracts ran to the first quarter of 2028. The plan was to be fully gone by then. Wren pulled that in by about a year and said, in public, that he had simply accelerated something Omnicom was always going to do.

The column frames this as Omnicom "answering with actions." The action was editing a calendar invite.

Publicis did not start the clock. Publicis walked past a clock that had been ticking for five years and got billed for the noise.

The strongest passage in the column argues that once buying goes agentic, identity signal stops being a targeting input and becomes training data, and whoever owns the pipes owns the feedback loop. That is a genuinely sharp idea.

It is also the idea Omnicom's own CTO stepped on twenty minutes earlier. Paolo Yuvienco said flatly that Omnicom did not use LiveRamp for any of its agentic media buys. Didn't need it. Their entire thesis is shortening the supply chain, cutting hops, and pulling in whatever graph is useful that day.

Which means the largest agency group on the planet ran its robot-buying pilots without the neutral pipe, before the deal was even announced. If losing LiveRamp's neutrality is fatal to agentic media, someone should notify the people already doing agentic media, who appear to be entirely fine.

Yuvienco described Acxiom's pitch to clients: they should own their graph, own their data.

Shaevitz's column lands on this: the scarce asset is the signal you generate and control yourself, that nobody else can license.

Same sentence. Different letterhead. One of them has a graph with 2.6 billion people in it. The other has a pricing page.

GrowthCode sells publishers first-party identity graphs and bid enrichment. Its About page brags, accurately, that it has no holding company and no demand-side interest, and argues that neutral infrastructure is not a fact of nature but a choice backed by a business model. That is a defensible position. It is also the precise destination the column marches you toward, in somebody else's magazine, under a headline about somebody else's conflict of interest.

When the $2 billion acquirer and the publisher-side vendor independently arrive at identical advice, that is not two smart people converging on truth. That is two guys in a gold rush agreeing you should really look into buying a shovel.

One. The FTC still has to clear this, as Wren pointed out while allegedly staying out of everyone's business. Nobody in the trades has written what that review turns on, or what a neutrality remedy would even look like on paper. "We pinky swear the pipes stay open" is not a consent decree.

Two. LiveRamp booked $813 million in revenue for the year ended March 31. If a meaningful slice of that was a wash trade with a holdco now sprinting for the fire exit, what did Publicis actually buy for $2.167 billion? That is a revenue quality question. It is answerable from the filings. It has been asked by nobody.

Three. Everybody is telling independents and mid-market publishers to build and own their graph. Nobody prints the invoice. What it costs, how long it takes, and how much of your glorious new independence gets quietly recaptured by the guy who sold you the infrastructure. The people telling you to stop renting are, in every single case, landlords.

LiveRamp's neutrality did not die in May. It died at a closing table in 2018, and has been propped up in the corner at industry parties ever since while everyone nods at it politely on the way to the bar. The new part isn't the death.

The new part is that the replacements now have marketing budgets, a column slot, and the nerve to deliver the eulogy at half past midnight.

How we reported this: This is an opinion column built on primary documents, covering May 17 through August 10, 2026. Every Wren and Yuvienco quote comes from Omnicom's corrected transcript of the May 19 J.P. Morgan Global Technology, Media and Communications Conference, read in full rather than lifted from someone else's summary. Deal terms come from the Publicis and LiveRamp announcements of May 17 and from LiveRamp's fiscal 2026 filings. Sadoun's remarks come from Publicis earnings coverage, the Shaevitz column from AdExchanger, and GrowthCode's positioning from its own published materials. Two limits, stated plainly: the $50 million figures are Wren's characterization on a public stage, not audited contract values, and Real ID's reach numbers are Omnicom's own claims. LiveRamp does not break out its Omnicom exposure in public filings, so the revenue quality point above is a question, not a finding. ADOTAT has no financial relationship with any company named here, takes no paid editorial, and is funded by subscriptions and readers.