Why ChatGPT's User Count Is Still Up for Debate Today

Six hundred million. Simple. Bold. That's the figure research firm Similarweb raised this week for monthly active users of the ChatGPT app — and it landed like a splash in a calm pond. Reuters had earlier published a far loftier number: one billion. Which one is right?

Numbers tell stories, but they don't always tell the same one. Part of the disagreement comes down to definitions. What counts as an "active user"? An app install that opens once? A signed-in account that exchanges a single prompt? Or repeat engagement across devices? Different firms answer those questions differently, and the resulting totals can swing dramatically.

Methodology matters. Similarweb leans on panel data, clickstream partners and inferred traffic patterns to estimate app usage. It's a well-regarded approach for gauging audience behavior outside a publisher's own analytics, but it builds from samples and models. Other reports — like the Reuters piece citing a billion users — may draw on different sources, internal disclosures, or broader interpretations that include web sessions, API-driven activity and enterprise accounts. Apples and oranges, often at scale.

Then there are technical confounders. Bots and automated traffic can inflate counts. Shared or corporate accounts obscure unique-user math. Cross-device sessions complicate deduplication. The pace of change matters too: AI services can spike and plateau quickly as new features, pricing shifts or platform launches alter habits. Snapshots taken days or weeks apart won't always line up.

The debate isn't just pedantic. Advertisers, investors and regulators pay attention to these metrics. A larger user base can justify higher valuations, attract more partners, and trigger closer scrutiny from competition and privacy watchdogs. Equally, an overhyped headline can mislead market expectations and erode trust when reality reveals a different picture.

At the end of the day, both figures point to enormous global reach — but they describe different slices of the same phenomenon. Look past headlines. Ask how the data was gathered. And maybe, ask what we really mean when we measure influence in the age of AI.