Google is winning the AI search race (and it hasn't played its best card yet)

The dominant narrative for the past few years has been simple: ChatGPT is the face of AI, and Google got caught sleeping. ChatGPT went mainstream first, it became a verb in some circles, and every "Google killer" headline you've read since has been built on that assumption.

The data tells a different story. A new Tensor Tower report has been making the rounds, and its main headline is that ChatGPT's market share among AI apps has dipped below 50% for the first time. Not because ChatGPT is shrinking, but because everyone else is growing faster. And that report doesn't even capture Google's biggest AI surface: the billions of searches that now run through AI overviews and AI mode.

In this post, I'll walk you through the numbers, and what Google's position means for e-commerce and the shopping journeys of your customers.

ChatGPT dipped below 50% market share

Let's start with the report itself. Tensor Tower tracks unique monthly users across mobile and web applications, and their latest data shows a clear trend: the AI market is no longer a one-horse race.

To be clear, ChatGPT is not losing users. Quite the opposite. Their user base has kept climbing and recently crossed the 1 billion monthly users mark.

The real story is relative growth. Other AI platforms are gaining users at a faster pace, especially Google Gemini and Claude.

And that faster growth shows in the market share numbers: for the first time since the report started, ChatGPT now accounts for less than half of all AI app users.

One important caveat: this report only measures unique monthly users of mobile and web applications. I believe Google's actual share is underreported here. Lots of searchers interact with AI overviews and AI mode directly in Google's search results, and as far as I can tell, those interactions are not included in this data. If they were, this picture would look even more lopsided in Google's favor.

ChatGPT might have been first to go mainstream, but it's becoming clear that other AI providers are doing better in a lot of ways. Claude has a firm grip on the developer and enterprise market (I personally use Claude most as well). And Google is poised to take the consumer market.

Follow the money

Market share is one thing. Sustainability is another. And this is where I believe ChatGPT's position gets shaky.

OpenAI has never turned a profit. In fact, it's burning cash at a horrifying rate, with losses reportedly surging to 21 billion dollars. OpenAI's only real product is ChatGPT, and it's not making them money. You can't subsidize a billion free users forever.

Compare that to Anthropic, whose only real product is Claude, and who is in a much better position: they're projected to have hit a profitable quarter in Q2. Now, this could all change in a couple of months again, but there's no denying that Anthropic is currently doing better than OpenAI in a lot of ways.

And then there's Google. Google has been one of the most profitable companies in the world for decades. They're heavily diversified, and they have a tremendous cash cow with their dominance in search and ads. They can easily carry losses in this AI race for years, and they don't have to worry about the same funding pressures that Anthropic or OpenAI do.

So we have three very different players: one burning billions with no clear path to profit, one that seems to have found a sustainable niche, and one that can outspend and outlast everybody. Keep that in mind for what follows.

Google hasn't played its biggest card yet

Google has been very clear about the future: Search will be an AI experience.

ChatGPT came out with nothing to lose, and it was the first company to define what that AI experience might look like. Google, on the other hand, had everything to lose. It couldn't simply copy that approach overnight without putting its ad revenue at risk. But make no mistake, Google has been moving in that direction rapidly.

AI overviews rolled out fast. What started as the Search Generative Experience (SGE) has become the standard answer that sits on top of the SERP (below the ads, of course).

Then Google launched AI mode as a standalone experience, and over time it has been integrated deeper and deeper into Search. AI overviews now transition seamlessly into the AI mode experience. AI mode ships on every phone that runs Android. AI mode is heavily embedded in Google Lens. Most SEOs believe AI mode will eventually become the default search experience, once Google is confident enough that its cash cow (ads) will perform well in that new environment.

And here's what makes this uncomfortable to admit: AI mode is very good. There's a huge ethical issue with how Google first copied public content and is now taking clicks away from that very same content. Less than one third of Google searches still send a click, and that number keeps dropping. But if you set that discussion aside for a moment and look purely at the user experience, AI mode delivers. You don't have to sift through countless websites searching for an answer. AI mode does it for you, and the results are pretty damn good.

So eventually, when Google makes AI mode the default, I believe the AI race for the consumer market will effectively be over. Google currently handles roughly 10 to 15 billion searches per day. Once AI mode becomes the default, all those searches will siphon straight into Google's AI ecosystem. No app download required, no habit change needed. That's the card Google hasn't played yet.

The place where people actually buy

There's one more area where Google's position is stronger than any AI app: shopping.

Earlier this year, I experienced this firsthand. I broke my phone during a late-night run, asked Claude for replacement recommendations, got a solid suggestion, and then did what everyone does when they're actually ready to buy: I went to Google. The product carousel showed me the same phone, 50 euros cheaper, with next-day delivery and reviews I could check before clicking. I've shared more examples like this in my post on organic product SEO and bottom-of-funnel visibility.

That buying journey illustrates my point: even with other AI tools claiming their stake, Google has maintained its position as the place where people make purchases. You may have done your research in ChatGPT or Claude, but when you're ready to buy, you will still google your product, find the best price, and click on either a Google ad or a Google search result.

Now, with Google getting better and better, Google aims to help you through that entire process.

Shortly after I published that post, Google announced the Universal Commerce Protocol (UCP). And it has been gaining ground fast, as I covered in my UCP rollout update. If you sell online, it's worth learning how to prepare for the UCP sooner rather than later.

Google can also rely on its Shopping Graph, a bet they made a couple of years ago by going all-in on e-commerce data, and it's currently paying off. If you're not familiar with it, I explained how Google's Shopping Graph works in an earlier post. Since the UCP rollout, the size of the Shopping Graph has increased massively as well.

Google still has a lot in store, and even without pushing full throttle (remember, AI mode is still not the default), they're catching up fast and look poised to overtake OpenAI. In fact, OpenAI has been found to scrape Google's results in an effort to improve its own AI answers. I documented the whole saga on LinkedIn, from the polite API access request that Google declined (now a DOJ antitrust exhibit) to the evidence that ChatGPT pulls encoded Google Shopping queries to populate its product carousels.

You can read the full breakdown of how ChatGPT uses Google Shopping data on this blog. The short version: the supposed Google killer depends on Google's data to answer product queries. That should tell you a lot about who's actually winning here.

What this means for SEO and AI search

Let me restate my position: I strongly believe Google will win the AI consumer market. Anthropic may well hang on to the developer and enterprise markets (seriously, they are really good), but on the consumer side, I expect Google to come out on top.

Marketers need to prepare for a future where Google is the dominant player in the AI landscape, especially if and when AI mode becomes the default. With UCP rolling out, Google wants consumers to complete their purchase without ever leaving Google. The research happens in Google, the comparing happens in Google, the checkout happens in Google. The user may never visit the seller's site at all.

All of that is powered by the Shopping Graph. It currently powers the product carousels that dominate traditional SERPs, and it powers the product results you see in the Gemini app and within AI mode.

That's exactly why I built Productrise: to help you optimize for Google's Shopping Graph. And in that sense, if you're already optimizing your product data for the Shopping Graph, nothing really changes. To be found in Google, AI mode, product carousels, Google Lens, hell, even ChatGPT, you need to make sure your products are properly included in Google's Shopping Graph, and from there you start optimizing. A good first step is auditing your product feed. That's the entire challenge Productrise is built to solve.

Want to see where your products currently stand? You can create a free Productrise account and start tracking your visibility without swiping your credit card. The AI search race may still be playing out, but the infrastructure it runs on is already clear. Optimize accordingly.