In his August 2024 ruling in the Google Search antitrust litigation, U.S. District Court Judge Amit Mehta found that Google’s default-distribution agreements—through which the company paid Apple, Mozilla, and others to make Google the preloaded search engine—were exclusionary under Section 2 of the Sherman Act. The court’s rationale focused on “default bias” and scale effects; by securing key default placements, Google purportedly obtained an insurmountable advantage from users’ reluctance to switch away from the default provider, the effect of which was to deny rivals the user data necessary to refine their own search engines, making it significantly harder for them to catch up.
Attention now turns to remedies. Under former President Joe Biden, U.S. Justice Department (DOJ) leadership advocated aggressive measures: banning Google from paying for default status, restricting its forays into artificial intelligence (AI), mandating data sharing with rivals, and even breaking off Google’s Chrome browser.
But the current DOJ leadership under President Donald Trump should not replicate their predecessors’ overreach. Rather, they should adopt more carefully tailored remedies that address any exclusionary effects, without harming broader innovation or saddling the court with the role of market designer.
The Biden DOJ’s proposed remedies fail to meet antitrust’s requirement of a tight causal connection between offense and relief, and risk imposing significant costs. The new DOJ should consider narrower solutions—like limiting exclusive terms—rather than imposing risky structural or behavioral fixes that could jeopardize browser competition, chill AI investment, and create more problems than they solve.
Lack of Causal Connection: Remedies Must Address the Proven Harm
Antitrust remedies must fix the violation the court found. Here, the judgment was that Google’s default deals foreclosed rivals’ access to scale. Yet as I have previously explained, a core criticism of the case is that it failed to establish a direct causal link between Google’s default deals and any lasting competitive harm. Yes, the court found that those deals “foreclosed” a chunk of the market, but did it truly prove that, “but for” those contracts, a competitor would meaningfully erode Google’s market share? I would say not.
As I have also discussed, the court’s liability analysis leaned heavily on behavioral assumptions of “default bias,” even though real-world evidence—like Firefox’s switch to Yahoo in December 2014—showed only a modest shift in user behavior subject to different defaults. While, as the court notes (p. 116), there was a small shift in usage from Google to Yahoo on Firefox, it was hardly enough (20 percentage points) to constitute market foreclosure.
And more importantly, the overall market effect was negligible. Most users appear to have changed the default or simply navigated to google.com (or perhaps switched to other browsers). Google’s overall market share barely budged following the switch (and nor did Yahoo’s) and, by May 2015 (six months later), Google’s share was above the pre-switch level and Yahoo’s was below it. This suggests that consumer preference for Google Search, rather than exclusive-default deals, might be the major driver of Google’s success.
As the court itself discusses, another default-switching experiment by Firefox yielded similar results:
In a 2016 experiment, Mozilla switched the default [general search engine] on both new and existing users from Google to Bing. By the twelfth day, Bing had kept only 42% of the search volume. After some additional time, those numbers dropped to 20–35%…. (Google Search decision, p. 117 (emphasis added))
And in still another experiment, “Mozilla… switch[ed] the default [general search engine] to Yahoo. Yahoo only retained 16.5% of the total search volume.” (Google Search decision, p. 117 (emphasis added))
Given the weak evidence on actual market foreclosure from default distribution, it’s unclear whether banning default contracts would really spur competition. If the court’s liability ruling was based on assuming, rather than proving, that these deals caused harm, the decision effectively “reliev[ed] the plaintiffs of having to show but-for causation,” making it easier to brand the deals as anticompetitive without proving they caused Google’s enduring market position. If prohibiting Google from paying for default status wouldn’t actually change market outcomes, a ban on default contracts could be an intrusive, costly remedy with little competitive benefit.
That’s a problem, as Google notes in its proposed final judgment (PFJ):
A mere inference of causation cannot support the imposition of drastic remedies expressly designed to intervene in the market and prevent Google from competing on the merits for a decade—and eliminate incentives for Google (and its rivals) to innovate and improve search and search ads technologies during the term of any final judgment (p. 2).
Antitrust law requires remedies that address the harm shown in court, not general corporate punishment for being “too big.” Overly broad relief could ultimately hamper market-driven outcomes and harm consumers—especially if Google’s popularity is mostly due to quality, not default status.
A more precise remedy would limit the duration of default agreements—giving rivals periodic chances to secure distribution—but not ban them outright. If such contracts were really a barrier, the expanded opportunities would allow user share to shift; if not, we’ll see Google remain on top, because users prefer it.
Problematic as it may be (see more on this below), at least prohibiting Google from paying for default-search placement is arguably connected to the violation found by the court. But as Ben Thompson keenly notes, the same cannot be said of the rest of the Biden DOJ’s far-reaching proposed remedies.
[An injunction against Google paying for search default distribution] is at least straightforward and directly connected to the violation at hand. That, though, is apparently not enough for the Justice Department: instead of identifying a problem and seeking to solve it, they are interpreting this case as an invitation to carve up a company they don’t like. (Emphasis added)
The new DOJ leadership should insist on a closer nexus between violation and remedy.
Prohibiting Default-Distribution Payments: Don’t Cripple Browser Competition
As noted, the prior DOJ team advocated barring Google from paying for default-search placement in browsers and operating systems. On its face, this might sound reasonable: it would stop the behavior deemed exclusionary. But beyond the problems discussed above, such a measure would ignore a crucial side effect: it could gut the revenue of independent browser developers like Mozilla, which relies on its search deals with Google for roughly 86% ($510 million in 2021-22) of its funding. Mozilla stated that the previous DOJ’s proposals would force it to “fundamentally reexamine [its] entire operating model,” constituting an existential threat. This would potentially crush the very browser competition that antitrust authorities should want to preserve.
Why Browser Competition Matters: Preserving Technical Diversity
A competitive browser market is important in its own right. If Firefox is weakened or exits the market, consumers will have fewer choices beyond the Google and Apple ecosystem. Perhaps more importantly, Firefox is now one of the last browsers not built on Google’s Chromium code base. Mozilla’s demise would imperil Gecko, the only major rendering browser engine not controlled by either Google (Blink) or Apple (WebKit).
Browser engines form the backbone of the user internet experience; they determine which new web standards get implemented and how securely browsers handle malicious sites. As the UK Competition and Markets Authority put it (in this case, with respect to mobile browsers):
Mobile browser engines play an important role in the user experience of mobile browsing, as they can impact factors such as speed, stability, and compatibility with different types of web content and websites. Different browser engines may also offer different levels of support for web standards, features and technologies, which can impact the types of web content that can be displayed on a particular mobile browser. (p. 9)
A world with just one or two engines is more brittle and less innovative, and Blink could become a near-monoculture on non-Apple platforms. Gecko’s presence helps to keep web standards a bit more open and pressures Google’s Chrome team to consider privacy and user-centric features.
Ironically, a remedy aimed at curbing Google could simultaneously strengthen the company’s power in the browser market. Yet there is no clear reason to do so, given that the alleged harm is not that browsers benefit from default deals, but rather that exclusive (or overly long) agreements might foreclose search rivals.
Of course, you’d also be hurting a company that is not a monopolist. Firefox has only a single-digit percentage of browser share globally—it’s hardly dominating anything. It would be a bitter irony if, in trying to punish Google’s dominance in search, we accidentally help entrench Google’s dominance in browser technology by tanking a smaller browser rival that relies on Google’s payments to survive.
The new DOJ team should take this as a cautionary tale. Remedies don’t exist in isolation and adjacent markets and technologies could be affected. In this case, a balance must be struck so that Google’s search contracts are reformed without inadvertently snuffing out one of the last independent browser engines.
Moreover, this is another practical reason to allow procompetitive default deals to continue. If Google can pay Mozilla for default placement without eviscerating competition, Mozilla gets funding to keep Gecko strong, and Google’s competitors still have pathways to users. It’s a win-win compared to a blanket ban.
The new DOJ leadership should prioritize a remedy that fosters multiple forms of competition—both in search and in the underlying browser tech—rather than trading one for the other.
Divesting Chrome: Not as Simple as It May Seem
Some have urged that the new DOJ should continue to pursue the Biden DOJ’s proposal that Google be forced to divest its Chrome browser. But this aggressive remedy has several significant problems.
To begin, the DOJ’s case never alleged that Chrome itself was acquired or maintained illegally. Chrome grew—indeed, overtook established competitors—by being a popular product. Demanding Google sell or separate Chrome would be a huge step beyond addressing the default contracts at-issue in the case. As Gus Hurwitz has noted, “[c]ourts expect any remedy to have a causal connection to the underlying antitrust concern. Divesting Chrome does absolutely nothing to address this concern.”
Forcing Google to sell off Chrome also risks creating more problems than it solves—for the browser ecosystem, for innovation, and even for Chrome’s own viability.
Chrome is much more than a standalone piece of software. It’s deeply intertwined with Google’s development infrastructure, security and anti-malware tools, and revenue model (search ads and related services). These synergies allow Google to invest heavily in Chrome—funding rapid iteration, security patches, speed improvements, and experimental features like privacy sandboxes. It also allows Chrome to remain free for users while leveraging revenue from Google Search deals.
A divested Chrome would lack those integrated revenue streams and supportive resources. Instead of drawing on Google’s massive R&D budget and expertise, an independent Chrome might need to implement a number of problematic correctives.
Among other things, it would need a source of revenue. But if the court simultaneously limits or bans certain default-search contracts, Chrome under new ownership may struggle to secure the funding once provided by Google’s search integration. New monetization models might compromise user experience (e.g., introducing ads into the browser interface) or force paywalls that deter users, making the browser less valuable for consumers and therefore less competitive.
In other words, splitting off Chrome does not guarantee a strong independent competitor; it may simply spawn another underfunded browser entity fighting to stay afloat.
Some have suggested that these problems could be avoided or ameliorated if a large firm (like OpenAI) could buy Chrome. But that hardly solves the problem of ensuring effective browser competition.
OpenAI excels in AI-based large language models (LLMs) and chatbots, not in the day-to-day realities of secure and standards-compliant browser development. The synergy that works for Google (tying search-ad revenues to browser adoption) isn’t obviously replicable for OpenAI, whose revenue model is based on selling AI services and API access. And OpenAI would need enormous new teams and continuous investment in user-interface development and web-rendering technology, which could distract it from its core AI mission.
At the same time, selling Chrome to a different large corporation might simply shift the “vertical integration” concern from Google to the new owner. If a buyer with vast resources (say, Microsoft) acquired Chrome, we might end up back in a scenario where the browser is entangled with the new parent company’s products, raising similar competition questions. Forcing a divestiture of Chrome could result in an outcome that does nothing to improve consumer choice or foster innovation. It might just reshuffle which big firm holds Chrome’s keys, while introducing friction and overhead that reduce Chrome’s performance and long-term viability.
Divestiture’s Woeful Track Record
Such a costly outcome would not be an aberration—indeed, it would be consistent with the abysmal history of divestiture remedies in U.S. antitrust cases, which suggests that forced breakups rarely help consumers.
In his masterful 2001 assessment of divestiture remedies, Bob Crandall finds virtually nothing positive to say:
[R]eview of the major Section 2 Sherman Act cases won by the government or ending in consent decrees provides remarkably little evidence that these cases and the relief that emanated from them had a positive effect on competition and consumer welfare. (p. 37 (emphasis added))
Further analysis undertaken by Crandall with Ken Elzinga offers an even more dire assessment:
In the best of circumstances, the behavioral relief obtained is simply irrelevant and has no economic consequence other than the cost of the litigation and any costs of compliance….
…The bottom line is that we cannot discern positive results for conduct remedies in the subsequent behavior of prices in any of these case studies…. (pp. 99-100 (emphasis added))
Broad remedies often fail because policymakers or courts misunderstand underlying economic forces. Divestiture may merely reshuffle control of assets without stimulating real competition. Vertical separation, in particular, tends to reduce efficiencies.
Copious empirical evidence also finds that, in most circumstances, vertical integration benefits consumers through lower costs and better product quality.
[T]he weight of the evidence is telling us… that, under most circumstances, profit-maximizing vertical-integration decisions are efficient, not just from the firms’ but also from the consumers’ points of view…. We therefore conclude that, faced with a vertical arrangement, the burden of evidence should be placed on competition authorities to demonstrate that that arrangement is harmful before the practice is attacked. (p. 680 (emphasis added))
Breaking up Chrome and Google Search purely because they are both successful might reduce innovation and degrade Chrome’s security, features, or synergy with related Google offerings. It is unlikely to help consumers.
In short, there is no basis for diving headlong into such a radical remedy, particularly where the proposed remedy has virtually no connection to the harms at issue. As Crandall and Elzinga put it (and they could well have been talking to the new DOJ antitrust leadership):
Given the difficulty in perceiving the causes of market changes, the DOJ should be reluctant to try to design complex remedies to remedy conduct that it does not fully understand in markets where the trajectory of change is difficult to predict. To do so may impede technological progress without providing any offsetting consumer benefits. (p. 100 (emphasis added))
The new DOJ leadership should acknowledge history and swiftly abandon the aggressive notion of a divestiture remedy.
Forcing Google to Share Data: The Pitfalls of a ‘Duty to Deal’
It isn’t only through proposed divestitures that the previous DOJ would dramatically exceed the boundaries of a sensible remedy in the Google Search case. One of the more expansive proposals pushed by the Biden DOJ was mandatory data sharing with rivals, on the theory that smaller search engines need access to Google’s usage data to refine their algorithms.
But forced “duty to deal” remedies are notoriously complex. And under U.S. antitrust law (the Supreme Court’s decision in Trinko), there is generally no obligation for a firm—even a monopolist—to share its proprietary facilities or data with competitors.
In Trinko, the Supreme Court rejected the idea that antitrust law generally requires a monopolist to affirmatively aid its competitors by providing access to its infrastructure. Writing for a unanimous Court, the late Justice Antonin Scalia warned that imposing a duty to deal “may lessen the incentive for the monopolist, the rival, or both to invest in… valuable innovations.”
Forcing Google to share its search index, real-time search data, or user signals would be even more invasive than the network-sharing obligations at-issue in Trinko. Google’s search algorithms and user data are highly proprietary, the product of billions in R&D over two decades. Notably, one of the Microsoft decisions (Massachusetts v. Microsoft) rejected forced, royalty-free sharing with rivals of extensive technical information and the open-sourcing of Internet Explorer (among other things) precisely because those proposed remedies would have deprived Microsoft of valuable intellectual property (holding that “confiscat[ing] much of the value of Microsoft’s [intellectual property] investment” was an “analogous form of structural relief” to divestiture (pp. 1230-31)).
And such a remedy would certainly turn the court into the sort of “central planner” the Trinko Court warned against.
Enforced sharing also requires antitrust courts to act as central planners, identifying the proper price, quantity, and other terms of dealing—a role for which they are ill-suited. (p. 408)
Search engines are not built for seamless interoperability. Compelling Google to share real-time signals or indexing technology would require detailed and ongoing supervision (e.g., how often to update data, in what format, and at what cost?); the creation and oversight of a complex technological apparatus (as even the Biden DOJ recognized when it noted that Google would need months to “implement the technology necessary to comply with this Section.” (DOJ PFJ, p. 13)); and, ultimately, price controls. As I’ve written elsewhere (with Josh Wright):
[M]andatory access must be regulated on economically bene?cial terms. The antitrust regulator becomes a systematic price regulator by virtue of accepting the theoretical claim that certain market dynamics can, in certain circumstances, extend an incumbent’s dominance beyond a theoretically optimal point. The information required to make this determination, however, is forever unavailable to the regulator. (pp. 176-77)
Indeed, Judge Mehta explicitly dismissed other “refusal to deal” claims in this very case on exactly such grounds, writing that:
Plaintiff States seek to bypass the “no duty to deal” doctrine entirely… Their claim requires grappling with a host of questions that the court is ill-equipped to handle…. And those thorny questions foreshadow the challenges the court would face in administering a remedy…. A favorable outcome for Plaintiff States thus would mire the court in Google’s day-to-day operations…. The court has learned a lot about Google, but it is “ill-suited” for that role. (Google Search decision, pp. 268-69 (emphasis added)).
Moreover, a mandatory data-sharing remedy is disconnected from the court’s actual finding: the default contracts were the alleged problem, not a refusal to deal. Rather, forced sharing addresses a different alleged problem—lack of scale or “data advantage”—that the court didn’t specifically deem a violation. If a remedy aims to fix the contracting practice (e.g., limit exclusivity), that doesn’t require telling Google to open its trove of user data to Bing. As a matter of legal tailoring, it’s out of scope.
Rather, imposing a forced-access remedy crosses the line into industrial policy, where the government aims to “level the playing field” by redistributing competitive assets. Under U.S. antitrust principles, that’s a bridge too far, unless it directly remedies specific exclusionary conduct.
Restricting AI Investment: Unintended Consequences in Emerging Markets
The prior DOJ’s proposed restrictions on Google’s AI investments—such as forbidding acquisitions of AI startups and forcing divestitures of Google’s existing AI technologies—could backfire by dissuading venture capital in AI more broadly. If founders and investors worry that a deal with Google will be impossible or unwound, they may avoid building or funding new AI innovations, ironically reducing competition.
The Biden DOJ’s proposed AI remedy is based on its claim (in the executive summary of its PFJ) that:
Google’s financial entanglements with current or future rivals risk compromising the proposed remedy. Investments in or acquisitions of potential rivals would stifle emerging competition or reduce their incentives to challenge Google. Such arrangements frustrate the PFJ’s remedial goals of fostering innovation and transforming the general search… market[] over the next decade. (p. 7)
But would they? Large firms like Google often acquire startups, not to eliminate a competitor, but to incorporate new technology and talent, helping the startup’s ideas integrate with other products and reach a larger scale. A blanket ban eliminates that pathway. And when a remedy casts such a wide net, it can entangle beneficial acquisitions that have nothing to do with search dominance.
Consider the message this sends to the AI entrepreneurship community: If you take investment from Google or get acquired by Google, the government might unravel the deal or force a sell-off later. That uncertainty can be a strong disincentive for investors and startups. If Google is taken out of the acquisition equation by law, it could dry up funding or lower valuations for AI ventures. As I’ve discussed elsewhere, “acquisitions by large incumbents are known to provide a crucial channel for liquidity in the venture capital and startup communities….”
Or as investor and serial entrepreneur Leonard Speiser put it:
The proposed prohibition on Google acquiring any “query-based AI product” is extremely broad. It wouldn’t just stop Google from buying an AI search engine; depending on interpretation, it could stop Google from buying any AI-related company that deals with answering queries. As Daniel Castro notes:
[F]ew companies that Google might invest in are going to be completely untouched by AI. In essence, Google would be unable to incorporate outside AI innovations into its products and services through a typical acquisition or strategic partnership.
As he notes, these restrictions could even prevent Google from acquiring an AI-enabled cybersecurity product that could improve user security—a completely unrelated area—simply because the wording might cover any AI that involves queries.
For a company like Google, which has invested some $300 billion in product innovations since 2010, such a restriction could dry up the well of consumer benefit that those investments have engendered.
And of course, such a heavy-handed intervention in AI markets could have a chilling effect on AI research more broadly. Google is one of the world’s biggest AI investors. If it has to think twice about integrating a new AI feature into search (for fear of regulatory scrutiny or violation of a remedy), progress could slow. Startups, too, might avoid building certain cutting-edge capabilities if they suspect those will be off-limits for one of their best use cases.
In accordance with the Trump administration’s stated AI goals, the new DOJ leadership should be very skeptical of the Biden DOJ’s AI remedy proposals. Trying to micromanage Google’s investment decisions could discourage exactly the investment needed for AI startups to flourish.
Conclusion: Avoiding Overreach
Antitrust law, at its best, is a scalpel, not a sledgehammer. It aims to remedy specific anticompetitive conduct, not to reshape industries (“transform the market” in the Biden DOJ’s words) to fit someone’s imagined ideal. The court should not become a “central planner” by rearranging conglomerate companies’ constituent parts or dictating business models. As one court has put it:
[A] finding of an offense under the antitrust laws does not invest a court with a license to embark upon a general program of comprehensive control of the defendants’ business…. To range far beyond the hard facts of the questions actually tried in the case is to invite the errors of a broad rule which proves unjust in the light of unforeseen circumstances.” (p. 355)
Antitrust remedies should be tailored to address proven exclusionary conduct, not to comprehensively redesign the tech industry.
Overly aggressive remedies risk punishing success, rather than misconduct. Google’s dominance in search is partly a result of building a great search engine that users choose in droves. The antitrust case is about specific contracts that allegedly crossed the line. Any remedy should aim to ensure opportunities for competition, not to handicap Google in unrelated areas or diminish the rewards of its innovations. As Judge Learned Hand famously warned, “the successful competitor, having been urged to compete, must not be turned upon when he wins.” (p. 430)
It is important to note, as Google does in its PFJ, that “[t]he Court did not find that Google’s payment for non-exclusive distribution or promotion of Google Search would have been anticompetitive or otherwise unlawful.” (p. 3) Leaving aside the arguable non-exclusivity of default deals, there is no basis in the decision to prohibit clearly non-exclusive arrangements.
That means there is no basis to prohibit search-distribution deals that don’t preclude partners from installing and promoting competitors, that don’t require partners to use Google Search in order to get access to other Google products, that aren’t locked in place by excessively long terms, and the like. There is certainly no basis to force Google to divest Chrome, share proprietary data, or forswear all AI investments.
By restricting itself to addressing the adjudicated violation, the new DOJ can avoid entangling the court in complicated oversight and inadvertently harming consumers. By contrast, the Biden DOJ’s wish list of forced data sharing, browser divestitures, and AI restrictions resembles industrial policy—a plan to engineer a particular market structure—rather than a remedy tied tightly to the alleged harm of exclusionary default agreements.
There is a fine line between promoting competition and trying to engineer a competitive outcome. The former is legitimate and desirable; the latter risks making courts (or regulators) into central planners, which historically hasn’t worked out well for consumers or for innovation. The new DOJ leadership should be mindful of this distinction. They have an opportunity to recalibrate the case toward measured, effective relief that addresses Google’s exclusionary conduct without overshooting and harming the broader tech ecosystem.
