This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
Every number on an earnings call is aimed at one audience. When Sundar Pichai told Alphabet’s investors this afternoon that AI Overviews now has more than 2 billion monthly users, the audience was shareholders and the message was demand. The same number reads differently if you are one of the companies those overviews describe.
The figures
In his prepared remarks for the second quarter, Pichai said Search delivered double-digit revenue growth and that AI Overviews has over 2 billion monthly users “across more than 200 countries and territories and 40 languages.” AI Mode, launched in the U.S. and India and still rolling out, has more than 100 million monthly active users. He added that AI Overviews are driving over 10 percent more queries globally “for the types of queries that show them,” that the effect is growing over time, and that the feature now runs on Gemini 2.5.
In April, the message was that AI Overviews monetize at about the same rate as before. This quarter the emphasis is scale and usage. Both are statements about Google’s business. Neither is a statement about accuracy, and an earnings call is not where anyone should expect one.
Read it as a denominator
For a communications or investor relations team, the useful move is to treat 2 billion as a denominator. Any error rate gets multiplied by it. For a company that people search often, even a rare mistake in its overview is a recurring one, repeated to readers who will never report it and whom the company will never see.
The query figure is more revealing than the user count. More queries “for the types of queries that show them” means people are asking more of the questions that trigger a generated answer. My guess is that these lean toward the explanatory and comparative: is this company stable, why did the shares fall, what happened with the lawsuit. Those are the questions where a company gets characterized rather than merely located.
Then there is geography. Forty languages and more than 200 countries means a company’s description is being generated in markets and languages where its source material may be thin or old, and where the company may not be looking.
Why issuers should care
Investors are a small slice of 2 billion people, but it is a slice with money attached. A retail investor checking a ticker, an analyst getting up to speed on a new name, a counterparty doing light diligence before a call: many of them now meet a generated paragraph before they open a filing. I argued in 2023 that markets move on filings but reputation hardens on the results page. The paragraph at the top of that page is now written by Google, at a scale its own CEO is happy to quantify.
Three things follow. Treat the overview for your company name, and for the obvious questions about your stock, as part of the disclosure environment you watch around earnings, alongside wire coverage and analyst notes. Make the facts an answer engine needs easy to find on your own investor pages: segment names, current leadership, the dates and outcomes of past events, in plain sentences rather than only in PDFs. And check the major non-English versions, because those 40 languages are part of the same exposure.
Alphabet’s shareholders heard a growth story today, and on the numbers given it is one. Everyone else should hear the same figure as a count of how many people are now introduced to companies by a paragraph the companies did not write.