What could’ve been Google’s worst year turned into one of its best

2025 could have been a pretty disastrous year for Google. The company came in knowing it had three major legal battles to fight, increasing competition in the burgeoning AI market, and a new US president to adapt to. It looked possible — perhaps likely — that the company might be broken up and forced to sell off Chrome or part of its ad tech business, leaving it forever smaller.

But that’s not what happened. Instead, we’re looking back at a year where Google set profit records, persuaded the courts to let it keep the status quo going at least a little longer in search and advertising, and established itself as one of AI’s few unambiguous success stories. What changed?

First, it’s worth remembering how bad things looked like they might get for Google 12 months ago. Its biggest problems lay in the courtroom: it had been declared a monopoly and was staring down the barrel of remedies in its Search antitrust case, including the potential of a forced sale of its Chrome browser; it had to defend itself in a separate ad tech antitrust case that threatened the possible break up of its other big moneymaker; and it was trying to appeal its defeat in court against Epic Games, which could have seen it forced to slash its fees on Android apps and open up to alternative payment methods and third-party app stores. And in those first two cases, it would be fighting the Department of Justice under President Donald Trump, a loud and frequent Google critic.

That’s all before you get to the fact that in addition to competing across all its usual product sectors — cloud computing, mobile, smart home, YouTube, and more — it had AI to worry about. Google couldn’t afford to miss out on tech that at bare minimum has the potential to upend its lucrative search business and could do a lot more than that, according to its proponents. But that meant fighting off old rivals like Microsoft and Meta, alongside new upstarts including OpenAI and Anthropic, all while pumping tens of billions into R&D and data centers just to keep up with the Joneses. Like every AI company, Google has had to take a very expensive gamble that this tech is ultimately going to pay dividends and that it will be one of the companies to profit when it does so. Google’s bet is even bigger than most, since its AI efforts are cannibalizing the golden goose, hastening Google Zero and eating into the search business that’s underpinned Google’s balance sheet for decades.

You can see why its executives might have had some nervous moments in January. But for the most part, Google’s worst fears haven’t come to pass.

The biggest bullet Google dodged was being forced to sell its browser Chrome. Judge Amit Mehta ruled that Google was a monopolist in search last August, but didn’t back the most headline-grabbing of the Department of Justice’s proposed fixes. The DOJ had argued that separating Google from its browser would remove one of the main ways it captures users and funnels them toward its search engine, prompting interest from suitors including OpenAI, Perplexity, and Yahoo, all of which would have quite liked their own way to funnel users to their search tools, thank you very much.

Google’s lawyers successfully convinced Mehta that splitting the business wasn’t the neatest solution, the judge ruling that “It would be incredibly messy and highly risky.” He also ruled that Google could keep making payments to Apple and other companies to secure prominent placement of its Search and AI products, one of the few cases where Google’s increased competition from OpenAI and other AI upstarts has done it a favor — Mehta argued that “these companies already are in a better position, both financially and technologically, to compete with Google than any traditional search company has been in decades (except perhaps Microsoft).”

That’s not to say that Google got off without consequence. In fact, the main remedy Mehta did propose — that the company must sell its search data to rivals at a marginal cost, to help them build search competitors — is one that Google had fought against tooth and nail, arguing that it would “deeply undermine user trust” in search and put data in the hands of smaller, less secure companies. But even here, Google got most of its own way. It’s only required to sell a subset of its search data, not the whole lot, and only once, rather than regularly supplying rivals with its latest data.

Taken together, the remedies are fairly toothless, as Google’s critics have been quick to point out. A process that could have seen Google sell off one of its greatest assets instead ended with a required sale of search data that is unlikely to give rivals a major leg up. (Google has also stated plans to appeal the underlying monopoly ruling, so even that requirement isn’t a sure thing.) So it’s business as usual for Google.

It’s too early to say if the same will happen in the company’s other antitrust trial, which Google lost in April. The DOJ wants to split Google up here too, forcing it to sell both its Ad Exchange marketplace and its Ad Manager tool, but there are already signs that it won’t get its way. Judge Leonie Brinkema has hinted pretty heavily that she’d rather the two parties settle than force her to determine remedies and suggested that behavioral changes to Google might have more short-term impact than structural ones would, since they’re slow to take effect and likely to be even slower if Google appeals. It adds to the feeling that the impetus to break Big Tech up may be fading away just at the right time for Google to get away scot-free.

Finally, there’s the lawsuit with Epic over Android’s app store and payment systems. Google lost that suit two years ago, and this summer it lost all over again in its appeal, leaving Epic to declare “total victory.” Google is appealing all over again, but in the meantime it’s been forced by the court to allow alternative payment methods in Play Store apps and will eventually have to list alternative app stores in its own storefront, too.

Those are big changes, but they’re ones Google still hopes it won’t have to make — at least not entirely. It negotiated a settlement with Epic that would instead see it reduce its app store fees and introduce a new class of “Registered App Stores.” Crucially, the settlement would apply worldwide, not just in the US, softening the blow by spreading it out. Just as crucially, the judge in the case still needs to approve the deal, with a hearing next month where the companies will plead their case together. If they succeed, Android will still be changed, but Google will keep control of it — more crucial than ever, as it tees up the launch of an Android-powered PC OS in 2026.

In the backdrop of all this legal wrangling sits the Trump administration. The president is no fan of Google by default, and his relationship with Big Tech has waxed and waned over the years, but he’s certainly proved amenable to persuasion. Earlier this year, YouTube paid a $22 million settlement to Trump to end a lawsuit over Trump’s suspended account, money which Trump promptly directed be spent on the White House’s new ballroom. Google itself was separately reported to have paid $5 million into the ballroom fund, following $1 million to Trump’s inauguration in January. None of that’s related to Google’s ongoing antitrust cases, of course, but a happy president never hurts.

In court, Google has spent most of this year running defense, trying to hold off the worst of the DOJ’s threats and maintain the status quo as much as it can. But elsewhere, it’s used 2025 as a chance to push forward.

You can see that in some of the small wins across its consumer hardware portfolio. The Pixel 10 phones are the first major Android flagships to support Qi2 charging, while the 10 Pro Fold was the first IP68 foldable. For perhaps the first time since the early Pixel cameras blew us away, Google is pushing Android hardware forward, rather than waiting for other manufacturers to get there first.

But Pixel phones don’t really drive Google’s revenue, and never will. AI? It might, or at least the company’s executives clearly think so. And here, Google’s had a good year. Google’s AI models have enjoyed a series of wins over the past 12 months. Veo 3 was the first video generator to dominate social media, long before the Sora app went viral; Nano Banana Pro has become perhaps the most convincing image generator on the market; and the flagship Gemini 3 launch went so well that it apparently sent Sam Altman’s OpenAI into a “code red” panic to make sure ChatGPT could keep up.

That’s all well and good, but for now at least, it’s not actually where the money is. Model-making rivals like OpenAI and Anthropic are burning through investors’ money in the hope of future returns, but it’s the hardware companies — and none more so than Nvidia — turning the biggest profits so far.

The advantage Google has over some of its rivals is that it can play both roles and is bolstered by steady income streams across the company as it builds up on both fronts. Steady is actually underselling it: in October, Google posted its first-ever quarter with over $100 billion of revenue, with $31 billion of income, giving it a sizable buffer to offset its investment in AI. The majority of that money still comes from ad operations, but the $15 billion of revenue from Google Cloud is the clearest sign of a positive impact from AI on Google’s bottom line.

And there’s good reason to believe substantially more growth is possible thanks to the hardware that underpins Cloud. Google has designed its own Tensor Processing Units (TPUs) to power Google Cloud for years and since 2018 has sold access to them through its cloud platform. But its seventh generation chip, Ironwood, will be the first it sells directly to other companies, with Anthropic already planning to buy a million of the chips to power its AI Claude and Meta reportedly close to signing a deal worth billions, too. Google isn’t yet a meaningful rival for Nvidia’s GPU market dominance, but this year it took the first steps to becoming one.

A lot remains uncertain. None of Google’s major legal battles are entirely settled, with appeals, remedies, and hearings still to come in each, which could go in Google’s favor or against it. It could still be forced to split its ad business in half, an existential threat to its largest revenue stream. And as the AI industry as a whole begins to look increasingly bubbly, you might worry that Google’s prospects there aren’t the surest guarantee of success.

But for a year that began with the prospect of Google being split apart, and has ended with it in one piece and making more money than ever, you can’t imagine CEO Sundar Pichai will have much to complain about right now. For now, that’ll have to do, and 2026 will be next year’s problem.