For Google, the Economics of AI Mode Look Shaky

When a company publishes its quarterly earnings statement, it can only hope for the best. The numbers are the numbers, and in those first minutes after release, there’s nothing the executives can do to influence what happens to the stock. There is, however, an opportunity to control the damage of an unpopular report using the earnings call. There is also the potential to make things worse, as was the case this week for Alphabet Inc.

Shares in Google’s parent company started tumbling soon after Alphabet reported its third-quarter results on Oct. 28, with both revenue and profit missing analysts’ targets. But the mood on Wall Street really darkened toward the end of the earnings call, when Chief Business Officer Philipp Schindler offered a glimpse into the changing economics of Google Search. “It’s worth noting,” he said, “that the strong growth in AI Mode queries had an impact because of the lower monetization we were seeing.”

What investors apparently took away from his comment highlights an unwanted paradox at the heart of Google’s business right now. The company remade its search engine by promoting AI Mode at the top of results and other artificial-intelligence-driven features such as a dedicated tab where people can converse with a chatbot. This has helped keep Google relevant as more people exhibit an interest in using AI instead of traditional search engines. But at least so far, these features haven’t delivered strong returns for Google. On Thursday the stock fell 13%, erasing more than $500 billion in market value.

Ever since the introduction of ChatGPT, investors have worried about the toll AI would take on Google’s business model. When they work well, chatbots deliver clean answers to users’ questions, meaning they don’t need to click on links — or ads. And then there’s the matter of volume. Google has sought to populate AI Mode with more ads, but that hasn’t matched the number of advertisements it can show on its main search page — or the revenue it generates from that page.

Google enjoyed a reprieve from these questions while reporting its quarterly financial results last year, when it delivered steady search growth. That was, in part, a quirk of the technology: The back-and-forth nature of conversations with chatbots like Google’s AI Mode meant more queries from users, according to Gene Munster, managing partner of investment firm Deepwater Asset Management. The latest report has shown the harsher economics of AI search.

To regain its footing, Google needs to shift from answering people’s questions to building agents that navigate the web for them and complete a whole host of other tasks, Munster says. Although modern AI including ChatGPT is built on breakthrough research by Google, the company often seems to be playing catch-up. OpenAI, Meta Platforms, Anthropic and SpaceX are all promoting their agentic software at the moment, but Google doesn’t yet have a comparable product that’s widely available.

These tools, including Dots, Duet, Grok Bot and Muse, offer promises for monetization through subscriptions or by taking a cut of transactions the bots facilitate. Even if Google did have such a product and deployed it, it wouldn’t generate enough revenue to make up for what the company’s core business could lose from a swift migration to AI.

Last year Google “pulled a rabbit out of the hat,” Munster says. It may need a lot more tricks to secure its future.Read next: Meta and OpenAI Use Cutesy Agents to Lure Early Adopters