A monopoly finding without a breakup
On September 2, 2026, US District Judge Leonie Brinkema in the Eastern District of Virginia declined to order Alphabet to divest its AdX ad exchange or open-source the auction logic behind its DFP publisher ad server, rejecting the Department of Justice's core structural remedies in United States v. Google LLC (Al Jazeera). Instead, she is requiring Google to adopt most of the parties' proposed behavioral fixes — including sharing real-time bid data for open-web display ads with rival ad servers, deprecating Unified Pricing Rules so publishers can set different price floors for different bidders, and ending the "first look" and "last look" auction privileges that let Google's exchange see and react to competitors' bids before finalizing its own (AdExchanger). The parties now have until October 2, 2026 — 30 days from the order — to jointly submit a final judgment translating those behavioral requirements into enforceable language.
This is a remedies-stage decision, not a reversal of liability. Brinkema already found, in an April 17, 2025 ruling, that Google illegally monopolized two distinct ad tech markets — publisher ad servers and ad exchanges — through more than a decade of acquisitions and tying conduct that locked publishers into Google's stack (DOJ press release). The Antitrust Division, joined by Virginia and other states in a January 2023 complaint, asked her to follow that finding with a structural cure: force AdX out of Google's hands entirely. She declined.
The case for divestiture, stated fairly
The DOJ's position deserves a fair hearing rather than a caricature. Behavioral remedies against a company with Google's engineering and legal resources are notoriously hard to monitor and easy to erode at the margins — a modified fee schedule here, a reinterpreted "first look" exception there — while a court is left refereeing compliance disputes for years. Public Knowledge's policy director Lisa Macpherson made the strongest version of this argument days after the ruling: because Google still owns both the ad server publishers use and the exchange that clears their inventory, the fundamental conflict of interest — Google grading its own auction — persists, and the order only "changes how the exchange can operate," not who operates it (Tech Policy Press). That is a real structural weakness, not a talking point. The DOJ's Microsoft-era instinct — that a monopolist which built its dominance through integration will find new ways to leverage that integration unless the pieces are separated — is grounded in a documented pattern, not paranoia.
Why behavioral remedies are still the better fit here
But the ad tech market is not the desktop operating system market, and the differences matter. Brinkema herself reportedly favored behavioral fixes partly because they arrive faster than a divestiture that would trigger a multi-year appeals fight over how to actually sever a codebase and customer relationships built around AdX and DFP operating together — during which nothing would change for publishers (AdExchanger). That timing argument is not merely procedural convenience. Digital advertising infrastructure evolves quickly, and a structural remedy engineered around today's auction mechanics — first-price versus second-price bidding, header bidding workarounds, server-side auctions — risks becoming obsolete or exploitable in new ways before a divestiture is even completed.
The specific remedies Brinkema ordered target the actual conduct her 2025 opinion found unlawful, rather than reaching for Google's market share as a proxy for harm. Ending "first look" and "last look" removes the informational asymmetry that let Google's exchange peek at rival bids before committing its own — the precise mechanism through which integration became advantage. Publisher-side floor-price flexibility addresses the specific complaint that Unified Pricing Rules stripped publishers of the ability to negotiate. These are narrower, falsifiable obligations a court can actually assess for compliance, rather than a one-time structural change whose competitive effects would only become visible years later. As the Computer & Communications Industry Association put it, the ruling confirms that "antitrust remedies should be narrowly tailored to address specific identified harms" rather than exceeding a court's own findings (CCIA) — a standard the DOJ's own divestiture ask arguably strained against, since Brinkema's 2025 opinion had already rejected the government's claim that a distinct "open-web display advertiser ad networks" market existed at all.
The real test is enforcement, not the order itself
None of this means the remedies will work. Behavioral orders live or die on enforcement rigor, and the DOJ's own institutional memory here is instructive: the 2001 Microsoft consent decree was widely criticized in its early years for toothless monitoring before later amendments tightened it. If the October 2 joint final judgment lacks concrete audit mechanisms, reporting cadences, and a credible path back to structural remedies if Google doesn't comply, Macpherson's skepticism will be vindicated regardless of how well-designed the underlying obligations are on paper. Congress and the FTC, not just this court, retain the tools to revisit the question if behavioral compliance proves illusory. But the remedy Brinkema chose reflects the specific violation she found, arrives without a multi-year implementation delay, and leaves room for correction — that is a reasonable trade-off, not a capitulation.