Who Is Using AI to Write Code? A Map of 1,659 Cities

"Who is actually using AI to write code?" is one of those questions everyone has an opinion on and almost nobody has a map of. The default mental picture is San Francisco: a few thousand engineers in a handful of ZIP codes, quietly reshaping the industry. That picture is wrong — or at least badly incomplete.

So we built the map. Who Is Using AI? is a live world map that estimates AI-coding-tool adoption across 1,659 cities in 197 countries, covering roughly 5.79 million tracked developers. It's a playful public-data estimate, not a census — more on exactly what that means below. But even as an estimate, it upends the Silicon Valley story almost immediately.

Here's what it reveals.

India, not the US, leads by developer volume

The single biggest finding: by total tracked developers, India is out in front.

India's lead isn't a rounding-error squeaker. It's roughly 170,000 developers ahead of the US, and it's built on genuine depth rather than one mega-hub. Bangalore, Pune, Hyderabad, Delhi, Chennai, Mumbai, and Noida all show up as serious concentrations. When people say "AI coding is a global phenomenon," this is the concrete shape of that sentence.

The US remains enormous — second overall, and still home to the densest cluster of frontier labs and tooling companies. But "AI-assisted development is a Bay Area thing" is a story that survives contact with exactly zero of this data.

Beijing and Bangalore run hottest by adoption index

Raw developer counts tell you where the people are. They don't tell you where AI-assisted coding is most intense relative to the local developer base. For that we use an adoption index — a 0–100 score blending developer density with how strongly a place searches for AI coding tools.

On that index, two cities sit at the very top:

Then a tight chasing pack: London (98), Shanghai (98), and Pune (97).

This is the more interesting metric, because it's normalized. A high index means AI coding tools aren't just present in a city — they're a live, actively-searched part of how that city's developers work. Beijing topping the chart with Bangalore a hair behind is a neat summary of where the center of gravity actually sits: split between East Asia and South Asia, with the West very much in the conversation but not owning it.

You can hover any city on the live map to see its index and estimated developer count for yourself.

London has the most raw developers of any single city

Here's a fact that surprises almost everyone: the single city with the most tracked developers isn't San Francisco, New York, or Bangalore. It's London, with ~147,775 developers and an index of 98.

The top of the city leaderboard:

São Paulo at number three is its own quiet headline — Latin America shows up far higher than the usual coverage would suggest, with Brazil ranking fourth among all countries. The San Francisco Bay Area is present and strong, but it sits below several cities that rarely get named in the AI-coding conversation. Density of influence and density of headcount are simply not the same thing, and the map makes that gap impossible to ignore.

The real story is how distributed it is

Zoom out and the pattern is unmistakable: AI-assisted coding is not concentrated in one country, one continent, or one time zone.

The top-25 countries by developer count include Bangladesh (~110k), Pakistan (~110k), Indonesia (~109k), Nigeria (~57k), Ukraine, Colombia, and Turkey — right alongside the US, Germany, and Japan. Dhaka alone shows ~81,165 developers, more than Seattle, Berlin, Tokyo, or Paris carry on raw count.

If you had to compress the whole map into one sentence: the future of software is being written everywhere, and the tooling is spreading fastest in exactly the places legacy tech coverage tends to skip.

That's the reason we made this thing public and free. A map is the honest way to show a distributed phenomenon — a bar chart of five countries would have hidden the actual story.

How we built it — and what it isn't

We're upfront about methodology because a stat you can't interrogate isn't worth much. The index blends two public signals:

1. Developer density — public, aggregate GitHub activity per city, used to estimate how many active developers a place has.
2. AI-tool interest — Google Trends interest in AI coding tools like GitHub Copilot and Cursor, used as a proxy for how strongly a region is reaching for these tools.

Those two signals are combined into the 0–100 adoption index and refreshed daily, so the map reflects a moving picture rather than a one-time snapshot.

Now the honest caveats, because they matter:

• It's an estimate, not a headcount. We are not reading anyone's editor, keystrokes, or private repositories. Nothing here is surveillance — it's public, aggregate signal only.
• Search interest is a proxy, not proof of usage. High Trends interest suggests appetite; it doesn't confirm every searcher became a daily user.
• Public GitHub data undercounts some regions. Places where developers lean on private repos, GitLab, or region-specific platforms will read lower than reality. If anything, the true global spread is wider than what you see, not narrower.

We think an honest estimate you can explore beats a precise-sounding number nobody can check. You can read the full breakdown of sources and limitations on our about page.

Find your own city on the map

Country-level numbers are a headline. The fun starts when it gets personal.

Search your own city on the live map. See its estimated developer count, its adoption index, and how it stacks up against Bangalore, Beijing, London, and São Paulo. Most people are surprised where their hometown lands.

The one-sentence answer to "who is using AI to write code?" used to be a shrug pointed vaguely at California. The data gives a much better one: millions of developers across nearly 200 countries — led by India in raw volume, topped by Beijing and Bangalore in intensity, anchored by London's sheer developer headcount, and growing fastest in the cities the usual maps leave blank.

It's a global map now. Go find your spot on it.