From Zero to a Billion People: How Fast ChatGPT Really Grew

Every technology has a launch story. The telephone took decades to reach a majority of homes. The web needed years of cables, browsers and dial-up modems. ChatGPT needed a web page. This post looks at how quickly it spread, what the numbers can and cannot tell us, and why the speed itself is part of the story.

The 100 million race

The comparison comes from a UBS analysis that Reuters reported in February 2023. It called ChatGPT the fastest-growing consumer application in history, reaching an estimated 100 million monthly active users about two months after its launch in late November 2022. The same note put TikTok at around nine months to reach that number and Instagram at about two and a half years.

It is worth saying plainly that these are analyst estimates, not audited figures. They are widely quoted because nothing else measured the same thing in the same way. I treat them as a good indication of scale rather than a precise ranking, and I would encourage you to do the same with any league table that compares products from different eras.

What made the launch so different

Most consumer apps need something from you before they work: a download, an account, a friend who is already on it. ChatGPT asked for almost nothing. You typed a sentence and got a reply that looked like it came from a patient expert. There was no learning curve to climb, which is rare for a powerful tool.

A few other ingredients mattered:

• It was free. People could try it without a budget or an approval.
• It worked in the browser and on phones. No new hardware, no company IT department in the way.
• The output was shareable. A funny poem or a clever answer made a good screenshot, and screenshots spread by themselves.
• Everyone had a different use. Students used it for essays, programmers for code, managers for emails. Each group told the next group.

From 100 million to 900 million to 1.2 billion

The part people forget is what happened after the first spike. Three reported data points tell the story:

Again, I only plotted points that were reported. I did not draw a smooth line between them, because that would imply data I do not have.

Be careful with the word "users"

Notice that the early comparison uses monthly users and the later numbers use weekly users. Weekly is a stricter measure, so the later growth is, if anything, more impressive than the headline suggests. Still, a user who opens the app once is counted as a user. Reach is not the same as value, and it is not the same as trust.

Three questions help whenever you see an adoption chart:

1. Who measured it? An independent analyst, or the company itself?
2. What was measured? Sign-ups, monthly users, weekly users or paying customers?
3. What is it compared with? Apps from different eras, markets and device generations are not truly comparable.

What I take from it

• Distribution beats invention. This technology spread through ordinary phones and browsers, so no one needed a new device or a company licence. That is why it moved so fast.
• Free access matters. Millions of students tried it before any university wrote a policy. Rules are now catching up with habits, not the other way around.
• Norms are being written right now. Your habits as a student, analyst or manager today will shape what is normal in your workplace tomorrow.
• Speed is a risk as well as an opportunity. When a tool reaches a billion people in a few years, mistakes, misinformation and over-trust reach a billion people too. Later posts in this series look at exactly that.

If you want to see how we got to a world where this was even possible, read the short history from Turing to the Transformer. And if you want the bigger picture of what the numbers mean for you personally, start with You Are Not Early and You Are Not Late.

Keep reading

• You Are Not Early and You Are Not Late: The Numbers Behind the AI Boom
• Then Machines Started to Create: GPT-3, ChatGPT, GPT-4 and the Pope in a Puffer Jacket
• Hundreds of Billions In: The Numbers Behind the AI Spending Boom

See the complete AI Era series

---