It’s been a seismic few days in tech antitrust, with three major announcements that have changed the landscape for the months ahead. For proponents of a fairer and more competitive Internet it’s fair to say that there have been highs and lows, but there is still everything to play for.
Search case
First, on the 2nd September, we heard from Judge Mehta, who was presiding over the Google Search case. This was brought by the DoJ and a group of US states claiming that Google was exploiting a monopoly in the search market. The early stages of the case looked promising for Google’s critics (MOW included) – Mehta Found in his August 5th 2025 ruling on liability that Google is an illegal monopolist acting in breach of US law. Google’s monopoly over search was confirmed and the issue that was then left open was the potential remedies. The US government and third parties in industry focused around forcing Google to share its data horde with competitors and a possible divestiture of the Chrome browser. This would address the key components of an antitrust remedy: the “rectification of harm to competition” and “denying the monopolist the fruits of its illegality”.
The judgement, when it came on 2 September 2026, was a major disappointment. The judge focused on the way that Google went about its monopolization, such as entering exclusive agreements with OEM suppliers such as Samsung, and distributors like Verizon and AT&T and competitors such as Apple. He struck those exclusivity provisions down. Divestiture is now off the table and while data sharing has been put in place to “share the fruits” of monopoly with rivals, there is now a major concern that Google will get fresh and up-to-the-second data and third parties will get delayed, out-of-date data.
As judge Mehta himself observed, the training of an AIO takes time. So, AIs trained on Google’s search histories and data horde have a problem with being up to date. For example, in response to the question “Is Taylor Swift single?” the AI trained on Google’s data from last year would return a response that Taylor Swift is single – but that is based on out-of-date data from last year. To complete the picture Google uses an internal product that it calls: “Fast Search” – an add-on that is used to “ground” the results in more up-to-date reality with fresh data. The 2025 position is that Taylor Swift has recently got engaged so Google’s AI would produce that result. Since Fast Search isn’t part of the remedy, and third parties are only given a subset of the Fast Search ranked results, third parties will not easily be able to compete. Any competitor using Google data to train a competing AI will have a lot more work to do to build a competing AI that would say that Taylor Swift is engaged.
Since Google’s monopoly has created a 2 billion user base with huge interaction data, its difficult to see where third parties will be able obtain equivalent information that is gathered every second by Google.
As a general matter the provision of data to Google’s own internal systems will be fresher and refreshed more frequently than provided to third parties. The DOJ thought this could be remedied overtime and overseen by a Technical Committee that could determine what is used by Google and hence what third parties can get. The judgement does have a Technical Committee but its powers have been pared back severely. This means the third parties will likely be out of date.
Rather than real time sharing of all of the data required to enable the development of a competing search engine, Mehta has proposed only the one time sharing of distorted and limited data, of limited use to anyone looking to swiftly compete with Google in search or using AI chatbots in competition with Google’s Gemini ( since Google’s Gemini will get up to the second fresh data inputs from Fast Search and other sources controlled by Google that are unavailable to rivals). Moreover, proposed language that would have prevented Google from circumventing even these weak remedies was removed.
As the final output of a five year process this was a huge disappointment to many. A once-in-a-lifetime chance to rebalance the web and curb Google’s excessive power had been thrown away. However, we believe that this is not the end of the case. The DoJ was, without doubt, keen to achieve a far stronger result than the one that was delivered and is being encouraged to appeal the decision. More importantly, the arguments given by judge Mehta to justify his decision appear to be inconsistent both within the document but also with some significant Supreme Court precedents.
We would hope that an appeal would lead to a reconsideration of this decision with the potential for more meaningful remedies to emerge from a rethink.
EC AdTech decision
Whilst the market was digesting the impact of Mehta’s flawed thinking, the EC decided to follow up on the evening of Friday 5th Sept with a bombshell of its own.
Following an extensive but largely behind-the-scenes process that was started by a complaint from MOW, the Commission announced the results of their own investigation into Google’s monopoly in AdTech. The Commission has decided to impose a nearly $3 billion fine and has asked Google to propose effective remedies to the findings of abuse. These relate to its “first look” and “last look” at bid prices that it sees in its ads exchange which it manipulates for its own benefit.
The ability to benefit its own business and discriminate against rivals arises from the vertically integrated nature of Google which is both the buyer and seller of ads and the owner of the exchange. It has been likened to the stockbroker owning the stock exchange and being able to manipulate the bids for stocks to its own benefit. Breaking off the exchange would solve the problem, as was proposed by Commissioner Vestager some time ago. The new Commissioner for Competition, Commissioner Ribera, appears to have taken the same view. The Friday Sept 5th press release raises the need for significant divestitures to remedy the situation.
Whilst one can debate the impact of a $3bn fine that amounts to three days’ of Google income, the contrast between the EU’s announcement after Judge Mehta’s weak ruling is clear. The EC has been willing to stand up to Google where Judge Mehta has not. The Commission cogently described the nature of Google’s monopoly and its impact on the market and suggested break up to remove the monopoly and prevent its re-emergence. In the court of public opinion no one is expecting much to change after Judge Mehta’s ruling and Google’s share price went up by 9% on the day.
Google now has 60 days to respond to the EC with its proposals of how the situation should be remedied. It remains to be seen what Google will suggest, but the divestiture of some parts of Google’s AdTech empire seems inevitable.
The fine proposed raised dark threats from Donald Trump of additional sanctions on Europe. We doubt it would come to that. The timing of the 60 days period for Google to respond will co-ordinate neatly with the final phases of Judge Brinkema’s AdTech trial in the US (see below) so it would seem likely that Google’s proposals to the EC will look to address both cases at once, removing any risk that the EC can be seen to be unilaterally punishing a US business.
AdTech case
With two such significant announcements in a few days, you’d think that the tech antitrust world would be sated but – sadly for the sleep patterns of those who follow this space – that was not the case. On Friday night the DoJ also posted its proposed remedies in the ongoing case under Judge Brinkema.
The proposed remedies are from the Trump appointed Department of Justice on behalf of the USA and deserve close reading. They also analyse the antitrust issues in much the same way as the EU Decision on Ad Tech and call for meaningful divestitures by Google as well as behavioural remedies to prevent it from self-preferencing in the ad auction process in the future. We know that the EU and DOJ have been talking. Now we know what they have been talking about.
These are welcome and necessary actions. However, there a number of areas remain where Google could potentially circumvent the US’s proposed remedies, such as by leveraging its existing first party data, by going around the exchange to sign direct deals with publishers and potentially even by removing the exchange from the value chain completely with on-device monetisation via the browser in something like Privacy Sandbox. The EU Decision is expected to address on-going harm and cover such circumventions. The US court process prevents such a finding and the risk will need to be closed off by other means.
What’s next?
For those of us who have been advocating for the dissolution of Google’s monopoly power last week was a roller coaster. Shock at the weakness of Mehta’s ruling was replaced by joy at the EC’s decisive action, finally balanced with qualified enthusiasm in the US proposals for remedies in Brinkema’s case.
What’s clear is that it’s more vital than ever that anyone with an interest in the outcome of these cases gets involved however they can, either on their own or as a part of organisation like MOW. If you want to be a part of changing the future of the Internet, email us at join@movementforanopenweb.com.