OpenAI’s Codex has recorded a 27-fold jump in weekly active users in India since the start of 2026, placing the country among the company’s five largest global markets for adoption of the AI coding agent.
Daily interactions with Codex in the country rose more than 20 times by late April, underscoring a sharp acceleration in demand for agentic software tools among developers, start-ups, enterprises, researchers and non-technical teams. India also ranks among the top ten markets for Codex engagement, strengthening its position as one of the world’s most active advanced AI user bases.
The figures, shared around Mumbai Tech Week on May 28, point to a widening shift in how AI coding agents are being used. Codex was built primarily to assist with software engineering, including writing code, fixing bugs, analysing repositories, running tests and proposing changes for review. Its use in India, however, is moving beyond traditional programming work. More than a quarter of Codex requests from users in the country are now for non-coding tasks such as synthesising information, drafting documents, automating research routines and organising workflows.
The expansion reflects a broader movement in artificial intelligence from chat-based assistance towards agents that can complete defined tasks with greater autonomy. Codex is designed to interpret natural language instructions, work inside development environments and help users move from an idea to a working output faster. For software teams, that can mean faster debugging, documentation, testing and prototype development. For founders and business teams, it can reduce the technical barrier to building internal tools, websites and applications.
OpenAI’s Asia-Pacific start-up leadership has described the country’s adoption pattern as notable because usage is not confined to software engineers. Founders, operators, researchers, students and business teams are increasingly using Codex to convert plans into working outcomes. That widening user base is important for a market where digital services, outsourcing, enterprise technology and start-up formation already have deep roots.
Earlier OpenAI data showed Codex usage for coding tasks in India at about three times the global average, while coding-related questions were nearly three times the global median. Those figures suggest that the country’s developer ecosystem is not merely experimenting with AI tools but integrating them into daily work. The rise also fits with the large concentration of engineering talent across Bengaluru, Hyderabad, Pune, Chennai, Delhi-NCR and Mumbai, where software services, financial technology, cloud operations and product engineering remain central to employment and investment.
Codex has also become part of OpenAI’s enterprise push in the country. The company has announced Codex-related collaborations with Tata Consultancy Services, Infosys and Razorpay, covering software engineering and enterprise workflow use cases. Such partnerships could deepen adoption among large technology services firms and digital businesses, especially where teams need to manage code quality, delivery timelines and productivity across distributed workforces.
The rapid uptake comes as competition in AI coding tools intensifies. OpenAI faces strong pressure from rival developer agents and assistants, including Anthropic’s Claude Code, GitHub Copilot, Cursor and other tools aimed at automating parts of software development. For enterprises, the choice of coding agent increasingly depends on reliability, security controls, integration with existing repositories, auditability, data governance and the ability to produce code that can pass internal review.
That caution matters because faster code generation does not remove the need for human oversight. AI-generated code can introduce errors, security weaknesses, licensing concerns or architecture choices that may not suit production systems. Companies adopting Codex at scale are likely to pair the tool with review workflows, automated testing, secure development policies and clear rules on sensitive data. For students and early-stage founders, the tool can speed up learning and prototyping, but it may also create dependence if users do not understand the code being produced.
The growth in non-coding use also shows how the definition of a coding agent is changing. Codex is increasingly being used as a general productivity layer for structured work: reading material, summarising requirements, preparing drafts, organising tasks and connecting technical and non-technical teams. That could make the tool relevant beyond engineering departments, particularly in operations, product management, research, consulting and small business automation.
