When it rains it pours and for Google, it has been a torrential downpour of biblical proportions. Within the past 12 months, they got clobbered by the DOJ in two major antitrust trials that has the potential to impact almost three quarters of their revenues (depending on how you look at their service lines) and the DOJ just opened a new probe in Google’s $2.7 Billion Character.ai deal from last year.

However, it is the Search remedies trial that has everyone’s attention as it goes into closing arguments next week (think of it as the sentencing hearing after a guilty verdict). Over the past several months, the DOJ made numerous far-reaching recommendations to Judge Amit Mehta of the United States District Court for the District of Columbia. Judge Mehta will need to strike a precise and delicate balance between restitution and overreach. Anything short of this incurs the risk of a reversal on appeal, like the Microsoft break-up order overturned in 2001 by a federal appeals court.

Is the problem as draconian as the proposed remedy?

The DOJ has proposed several remedies to end Google’s chokehold on the Search market, but the one that has drawn the greatest interest of industry players and spectators alike is the divestiture of Chrome.

DOJ’s claim is that this divestiture is vital to unfetter search markets by ending Google’s control of a critical search access point. 90% of Google’s search volume flows through Chrome, choke that off and voila, no monopoly. One can certainly see the appeal the DOJ sees in this. Google of course is vehemently opposed to this, claiming adverse impact to consumer privacy and data security.

Sold to the highest bidder?

Counter to Google’s hopes, there are potential buyers able afford to pay fair market value for Chrome, plus staff the hundreds of engineers necessary to keep it updated to a changing tech landscape. Undaunted by valuations ranging upwards of $20 billion (more likely 2X that), a line-up of suitors has emerged including ChatGPT owner OpenAI, AI Search engine Perplexity, Yahoo! And video-sharing platform Rumble. However, except for two, the rest may not have access to swing a $20-$40 billion Chrome buyout in the near-term. Keep in mind that any buyer is essentially paying the money to acquire Chrome users, the product itself has no revenue. Perplexity has a current valuation of only about $8-10 billion and has raised about $915 million to date (although they do have Bezos as an investor and therefore access to his formidable war-chest). Rumble has a bit over 100 million cash on hand. That realistically leaves OpenAI and Yahoo! (although Apollo Global Management, Yahoo!’s 90% owner may just have enough ready cash on hand to come in at the lower end of the buying price range).

So, that leaves OpenAI as the most likely contender, barring a wildcard entry into any bidding scenario. For instance, xAI’s Grok can certainly outbid all of the above players, and it will certainly catapult Grok from AI underdog with a current <10 million users, to topdog with Chrome’s 3.5 billion users.

Separately, if OpenAI doesn’t evolve to a for-profit entity by end of year, the $30 billion from Softbank it would need to make the Chrome acquisition reduces drastically to $20 billion. A “for-profit” OpenAI dramatically changes the optics on how Chrome fits into its roadmap as we will touch upon further down.

Bottomline- there are potentially one or two serious contenders to acquiring Chrome. And by finding a new home for Chrome, the DOJ hopes to reverse the harm from limiting of consumer choice created by Google’s default deals.

Sometimes the cure is worse than the disease

Before we put Chrome up for adoption, let’s take a look at its genesis. When Chrome launched in 2008, the slow, clunky and crash-prone Internet Explorer was reigning supreme in the web browser game. Chrome came along offering a significantly better user experience- faster start-ups and page rendering, clean user interface, combined search and address bar, one tab not crashing all others, automatic malicious content warnings, cross-device syncing and so on. It was a game-changer in web experience for users.

Since then, Chrome has become the world’s most popular browser. And this has a lot more to do with feature dominance than default agreements. And it is a direct result of the sheer amount of support behind the Chromium open-source project that is the backbone of Chrome’s codebase. In 2024, the Chromium project had over 100,000 commits (significant code updates), and according to Chromium Blog, 94% of these updates came from Google. Other browsers like Microsoft Edge are also built on Chromium. Chromium’s Blink is the primary engine used by developers to test how their web apps will work across Chromium-based browsers. All of this could evaporate depending on what agenda Chrome’s new owners want to pursue.

Chrome is an independently superior browser, no doubt. And currently, Google Search is the best search engine (although AI driven-search is changing this dynamic as we speak). To quote Judge Mehta in his ruling, Google is “the industry’s highest quality search engine”. It is highly convenient for consumers today to use these two products together (Google Search function and Chrome).

A hasty divestiture can significantly disrupt consumer use of the product. Any buyer that hopes to retain and convert Chrome users, needs to offer the same seamless user experience within their product stack, as it does currently within the Google stack.

And decoupling Chrome from the Google product stack is far from a trivial effort; single-sign on across Google products gets disabled, password manager gets disabled and users need to bulk-download current passwords, users then create a login within the new stack and opt into any products with a compelling enough value proposition as well as user experience benefit to integrate into Chrome.

Needless to say, user attrition will follow- the $20-$40 billion dollar question is how much?

And if Chrome finds a new home in an AI-first stack, then who ensures that the user data that opens up is not abused? We are still in fairly nascent stages of the Generative AI evolution and while discussions have started around how Generative AI uses the vast amounts of consumer data it is trained on, regulatory oversight is woefully lagging its rapid ascendancy in the marketplace. The sheer amount of data from Chrome userbase into a profit-motivated AI startup will likely put Generative AI growth on steroids, without the pervasive privacy regulation it currently has under Google. Very subtle tweaks in the underlying LLM models that power Generative AI models, through either training data or hyperparameter tuning, can amplify existing data biases or introduce new ones. Seven years on from the Cambridge Analytica scandal and we still do not have a comprehensive all-encompassing federal privacy law and the private sector Generative AI development lead over Governmental regulatory oversight is widening every moment.

So, the true question for Judge Mehta is not who can buy Chrome, but rather who should? The wrong buyer will likely do worse consumer harm than what was caused by default agreements by substantially undermining a highly functional consumer product, and setting up the potential for an adverse consumer privavcy event of monumental proportions.

Getting to the heart of the matter

Splitting up firms should always be a last remedy as it has the greatest potential for unintended and permanent consequences in comparison to behavior-corrective remedies. To quote legal scholar and leading antitrust expert, Professor Herbert Hovenkamp, “When considering the appropriate remedy for a proven antitrust violation, courts should always look first to injunctions, which can be more effectively targeted at the harm and are likely to do less collateral damage”.

Google’s ability to unfairly capture more than their share of search ads through monopolization (and continue doing so) is made possible, predominantly by the default deals, and the immense scale they have achieved, partly through those deals. Three of the other remedies the DOJ is asking for already address these two issues.

First, ending competitor blocking by terminating existing third-party default deals and prohibiting future ones will be huge in restoring immediate competitive balance in the short term. To put this in perspective, per Google’s own expert testimony, 36% of its $175 Billion paid search revenue in 2023 came through Apple’s Safari browser and per their internal modeling 60-80% of that opens to competition once default deals are eliminated, injecting much needed inventory liquidity. Of course, Google may still be able to retain some of those users through branding ads targeting iPhone users, but that is fair game.

Second, providing competitive access to Google’s vast Search Index intellectual property is life-blood for medium term competition. It will afford the benefits of Google’s scale to smaller competitors to develop their own rival offerings without the massive computing outlay entailed in crawling and indexing billions of sites on the ever-expanding web.

Third, enabling industry access to a syndicated and privacy-safe version of Google’s organic search insights has really the potential to be a game-changer for the long-term. For instance, this allows competitors to refine AI-enabled search overviews to far greater levels of precision than seen today. Indexing and retrieving specific content users are looking for becomes exponentially complex as the world wide web grows by 402 million terabytes of data every day. Rival Search engines can take a quantum leap forward by leveraging Google’s data with the sheer power of generative AI to assimilate, evaluate and summarize usable information. Google’s already doing this through Gemini powered Search overviews.

Of course, there are logistical challenges in implementing 2 & 3 above and the DOJ’s six-month assumption for Google to accomplish all of this is naïve. This will likely need to be implemented via Google’s existing clean room architecture that services their Ads Data Hub product (which took over 2 years to build and needed a fraction of the scale required for Search data). This system will need to be partitioned from Google’s ads unit to ensure adequate separation to prevent any “insider” privileges for Google Search Ads.

Finally, the oversight needed to ensure that the implemented solution is fit for purpose and to manage it, will likely need to be far more sophisticated than the five-person expert Technical Committee envisioned by the DOJ. This needs the resources of an independent technology watchdog commissioned for the oversight role that can deploy different areas of expertise as and when needed. That will likely need an act of Congress to be established

In any case, I believe Judge Amit Mehta has an incredible opportunity to be the catalyst for a comprehensive set of remedies that will course-correct the Search industry and set the tempo for data-driven collaborative innovation that will have ripple effects beyond Search into the broader future of Generative AI.