OpenAI just put a number on something a lot of Indian developers already felt was happening. Weekly active users of its AI coding tool Codex in India have grown 27 times since the start of 2026. The company shared the figure at Mumbai Tech Week in late May, and it places India among the top five countries in the world for Codex adoption and the top 10 for overall engagement.
Here’s the short version if you only read one paragraph: OpenAI Codex in India isn’t growing because a few thousand engineers found a new toy. Daily interactions with the tool climbed more than 20x by late April, and over a quarter of all requests now have nothing to do with writing code. Students, founders, and freelancers are using it too. That mix is what makes this interesting.
So let’s break down what Codex is, why the numbers look like this, and what it actually means if you’re learning to code or trying to earn from it.
What is OpenAI Codex?
If you tuned out the name “Codex” because you remember an old OpenAI code tool that got shut down years ago, fair enough. This is a different product.
Codex is an AI coding agent today. You give it a task in plain language, and it reads your project, writes or edits the code, runs the tests, and hands you back something you can review. It launched as a command-line tool in April 2025 and has since spread across the ChatGPT app, a desktop app, browser extensions, and editor plugins. In May 2026, OpenAI even added it to the ChatGPT mobile app, so you can approve an agent’s work on your phone while a task runs in the cloud.
The current version runs on GPT-5.5, OpenAI’s agent-focused model. Globally, roughly 4 million developers use Codex every week.
A few things it actually does:
- Writes features and fixes bugs across an entire codebase rather than one file at a time
- Explains existing code in plain terms, which helps when you’re learning
- Runs tests and proposes changes you can accept or reject
- Handles longer tasks step by step instead of one prompt at a time
OpenAI has been clear about where it’s headed. The company describes Codex as something that helps people go from an idea to a working result, with software being the first big use case rather than the only one.
Why are Codex users in India growing so fast?
A 27x jump in five months is steep even for a fast-moving market. So what’s behind it?
Start with the raw size of the talent pool. India crossed 27 million developers on GitHub in 2026, up from 21.9 million in October 2025. More than two million of those signed up this year alone, which works out to roughly one in seven new developers worldwide. India is now GitHub’s fastest-growing developer community, and on current trends, it’s set to become the largest anywhere by 2028.
The depth matters as much as the raw count. India is the second largest source of open-source contributions in the world after the United States, and the second largest contributor to open-source AI projects, with more than 7.5 million contributions logged. So you’ve got a huge population of people who already write code and already work with AI. Codex walked into a room that was primed for it.
Then there’s the startup layer. Indian founders are under real pressure to ship products quickly with small teams, and an agent that can knock out features and maintenance work fits that exactly. Enterprises are moving too. TCS, Infosys, and Razorpay have all started deploying Codex across engineering and business workflows.
And students are a big part of the story. AI tools are cheap to start with, often free at the entry tier, and a college student in Pune can now get the same coding assistant a senior engineer in San Francisco uses. That access gap, the one that used to slow Indian learners down, has mostly closed.
OpenAI itself frames India as one of its fastest-scaling markets, and earlier data from its Signals report found that coding-related Codex use in India ran about three times the global average. The 27x figure continues a trend that’s been building all year.
Codex use in India is moving beyond code
This is the part most coverage glosses over, and it’s the most telling.
More than 25% of Codex requests from India are now for non-coding work. Drafting documents. Running research. Organising workflows. Pulling scattered information into something usable. People are using a coding agent as a general execution tool.
Thomas Jeng, who heads startups for APAC at OpenAI, put it plainly: adoption isn’t limited to software engineers. Founders, operators, researchers, students, and business teams are using Codex to turn ideas into finished work faster. He pointed out that while Codex started as a coding product, people keep stretching it into other parts of their jobs.
Why does that matter? Because it changes who the tool is for. A researcher with no engineering background can hand Codex a messy pile of notes and get back a structured summary. A small business owner can automate a reporting task they’d normally pay someone to do. The coding skill becomes optional. That widens the market far past the 27 million developers, and it’s probably the real reason the India numbers are climbing the way they are.
What this means for students
If you’re a student learning to code, this is good news, with one catch.
The good part is speed. Codex can explain a confusing chunk of code in language you understand, walk you through why a bug is happening, and help you build actual projects instead of toy exercises. Debugging, which used to eat entire evenings, gets a lot faster. You can ship a working app for your portfolio in a fraction of the time it took a few years ago.
Use AI to learn, not to skip learning
Here’s the catch, and it’s a big one.
An AI that writes correct code for you is dangerous if you never understand what it wrote. Copy-pasting Codex output into your assignment without reading it teaches you nothing, and it falls apart the moment you’re in an interview or a job where you have to explain your own work.
Treat Codex like a sharp tutor, not a vending machine. Ask it why it chose an approach. Read the code it gives you. Break it on purpose and watch what happens. The students who get the most out of these tools are the ones who already understand the fundamentals, or are actively building that understanding while they use them. The concept matters more than the output. Don’t trade one for the other.
What this means for freelancers
For freelancers, the picture is mostly upside-down, as long as you adapt.
Faster delivery is the obvious win. A web development project that took two weeks can shrink considerably when an agent handles boilerplate, fixes, and repetitive setup. That means you can take on more clients, or charge by outcome instead of hours.
A few practical angles worth thinking about:
- Automation as a service. Plenty of small businesses want repetitive tasks handled and don’t know how. If you can wire up Codex to do that, you’re selling a service most of your clients can’t build themselves.
- Faster site and app builds. Landing pages, dashboards, small tools, all of these come together quicker, which lets you say yes to work you’d have turned down before.
- Research and content workflows. Since a chunk of Codex use is now non-coding, you can lean on it for the research and drafting parts of a project, well before you write any code.
The freelancers who’ll struggle are the ones charging purely for tasks a tool can now do in minutes. The ones who’ll do well are those who move up the chain, toward judgment, client relationships, and putting the pieces together. If you’re building a freelance career around AI tools, that shift is worth planning for now.
Is India becoming an AI builder nation?
For a long time, the worry went like this. India would end up consuming AI built elsewhere and never make much of its own.
The Codex data complicates that story. Indians are using AI to build and ship real products at scale, across startups, enterprises, students, and solo workers. Globally, Codex crossed two million weekly users by March 2026, and India’s growth has outpaced that broader trend by a wide margin.
Pair that with the GitHub numbers, the open-source AI contributions, and the enterprise deployments, and a clearer shape appears. India has the developers, the startup pressure to ship fast, and now the tools to do it cheaply. Whether that fully translates into a builder economy depends on what gets built over the next few years. But the raw inputs are all there, and they’re growing faster than almost anywhere else.
FAQ: Why are Indian users adopting Codex so fast?
It comes down to three things stacking up at the same time.
First, scale. India has 27 million developers and the fastest-growing developer base in the world, so even modest adoption rates produce huge absolute numbers.
Second, access. AI coding tools are cheap or free to start, which removes the barrier that used to hold Indian learners and small teams back. A student now reaches for the same assistant a funded startup uses.
Third, breadth. Codex stopped being only a coding tool. With a quarter of Indian requests going to research, drafting, and workflow tasks, the user base extends well past engineers into founders, students, and business teams.
Put those together, and a 27x jump in 2026 looks less like a surprise and more like what happens when a ready market meets a tool that finally fits it.
Published May 2026. Figures are from OpenAI data shared at Mumbai Tech Week and GitHub’s 2026 developer statistics. AI tools change fast, so check current pricing and features before relying on them.
