Google’s new phones aren’t really about the phones

At an event Tuesday afternoon, Google unveiled the ninth generation of its Pixel smartphone line. These are the phones that Google designs fully in-house, which typically draw high praise from reviewers but generate little in the way of actual sales. Google’s Pixel phones accounted for just under 1% of global smartphone shipments last year and the first half of this year, according to data from Counterpoint Research.

But the new phones are giving Google and its parent, Alphabet, a chance to lean even harder into generative artificial intelligence. New AI features, such as an image-generating app called Pixel Studio and another app that scans screenshots for content, took up the majority of time at Tuesday’s event—far more than any hardware design elements. And some new features won’t even be exclusive to the Pixel, as Google wants its AI tools widely dispersed across other phones running on its Android mobile operating system. Gemini Live—a conversational AI chatbot powered by voice—is rolling out to all compatible Android phones in English on Tuesday, and Google said a version for Apple’s iOS devices is also coming soon.

The catch? Gemini Live and other AI features are only available to subscribers of Google’s Gemini Advanced plan. That plan costs $20 a month, though Google is throwing in a year of free access to buyers of its higher-end Pixel Pro, XL and Fold devices.

That might seem steep—especially since Android users are already accustomed to having Google’s well-known digital assistant free of charge. But Google has actually proven more adept at upselling once-free services than it has in selling premium smartphones. Non-advertising revenue from YouTube—much of which comes from viewers wishing to avoid ads—totaled $11.9 billion last year, which is 20% higher than Google’s total hardware revenue for the year, according to consensus estimates from Visible Alpha. YouTube’s non-ad revenue has also averaged 52% annual growth over the past four years, compared with 16% for all Google hardware, according to those same estimates.

Upselling AI on a mobile device is still no sure thing, though. Particularly given some high-profile stumbles by Google over the past 18 months, as it has raced to stay competitive with Microsoft and its anointed AI partner, OpenAI. And Google now faces the added challenge of competing with Apple’s generative AI debut. Apple Intelligence—the iPhone maker’s moniker for a set of new AI tools designed for its devices—is set to launch this fall.

But while Google may be a bit player in hardware, its products like search, Gmail and Android are effectively the world’s largest distribution network for new technologies such as generative AI. And its Pixel event Tuesday proved a bit of a flex in that regard, coming a month ahead of when Apple typically introduces its new iPhones. Rick Osterloh, who runs Google’s device business, emphasized repeatedly Tuesday that the AI services shown at the event are ready for launch—another subtle dig at the more gradual rollout expected for Apple Intelligence.

Google’s little phone business is clearly still looking to punch above its weight.

Write to Dan Gallagher at dan.gallagher@wsj.com