Google has revealed that 75% of all new code written at the company is now generated by artificial intelligence, with human engineers reviewing and accepting the output rather than writing from scratch. The disclosure, made by Google CEO Sundar Pichai, is the most specific and striking public data point yet on how thoroughly AI has already transformed software development at one of the world’s largest and most sophisticated technology organisations.
The number is not a projection or a target. It is the current reality at Google today.
What 75% Actually Means
To understand the significance of 75% AI-generated code at Google, it helps to understand what Google’s engineering operation looks like. Google employs tens of thousands of software engineers across its Search, Cloud, YouTube, Android, Maps, DeepMind and other divisions — some of the most highly compensated and technically accomplished developers in the world. These are not junior programmers writing simple scripts. They are engineers building the infrastructure that serves billions of users globally every day.
When Pichai says 75% of new code is AI-generated, he means that across this entire operation, for every four lines of new code added to Google’s codebase, three are now produced by an AI model rather than typed by a human engineer. The human engineers are still in the loop — they review, evaluate, modify and accept or reject what the AI generates. But the act of writing code — the actual production of the text that becomes the programme — has already shifted to AI for the majority of Google’s new software output.
Why This Matters Beyond Google
Google’s disclosure is significant not just for what it says about Google but for what it signals about the direction of the entire software industry. Google is not a company that adopted AI coding tools experimentally or partially. It is a company that built the foundational research behind modern large language models, that created the Transformer architecture that underpins virtually every major AI model in use today, and that has been integrating AI into its engineering workflows longer and more deeply than almost any other organisation.
If 75% of Google’s new code is AI-generated today, the question for every other technology company — and every company that employs software developers — is where they are on the same trajectory and how fast they are moving along it. Microsoft, Amazon, Meta and other major technology employers have all disclosed significant AI coding adoption through GitHub Copilot and internal tools. But 75% is a number that places Google ahead of any publicly disclosed figure from a comparable organisation.
What Human Engineers Are Now Doing
The natural question that follows from 75% AI-generated code is: what are the human engineers doing? The answer, based on how AI coding tools work in practice, is that they are doing the work that AI cannot yet do reliably — system architecture decisions, complex debugging across large codebases, security and reliability review, understanding business requirements and translating them into engineering specifications, and the review and quality control of the AI’s output itself.
Reviewing AI-generated code is not a passive activity. It requires understanding what the code is supposed to do, whether it does it correctly, whether it introduces security vulnerabilities, whether it is efficient, whether it integrates cleanly with existing systems, and whether it will be maintainable by future engineers. These are high-judgment tasks that currently require human expertise. But the volume of code that engineers are reviewing has increased dramatically as AI generates more of it, changing the nature of the job from production to evaluation.
The India Implication — The World’s Largest Developer Workforce
India has the largest software developer workforce in the world, with estimates ranging from 5 to 6 million working developers and a pipeline of hundreds of thousands graduating from engineering programmes annually. India’s technology services industry — TCS, Infosys, Wipro, HCL and hundreds of smaller firms — is built substantially on the human capital of this developer workforce providing software development, testing, maintenance and support services to global clients.
Google’s 75% figure lands in this context with particular force. If the world’s most sophisticated technology company has already automated three quarters of its new code production, the implications for an industry whose primary value proposition is human software development capacity are significant and immediate. The transition is not theoretical or distant. It is happening now, at scale, inside one of the most technically capable organisations on earth.
The Indian IT industry’s response to this transition — through AI upskilling, shifting toward higher-judgment engineering roles, building AI-native service offerings and repositioning the value of human engineers from code production to AI supervision and system design — will be one of the defining strategic questions of the next decade.
What Comes Next
Pichai did not say 100% of code will be AI-generated. The 75% figure implies that 25% is still human-written — suggesting there are categories of code where AI either cannot yet perform adequately or where human judgment remains essential. The direction of travel is clear, however. The number was lower last year than it is today. It will be higher next year than it is today.
For developers, the Google disclosure is both a warning and a reframing. The engineers who remain most valuable in a world of 75% AI-generated code are those who understand systems deeply enough to judge whether the AI’s output is correct, secure and appropriate — not those whose primary skill is the mechanical act of writing code. The craft is shifting from production to judgment.
Google just told the world where software engineering is going. The 75% number is not a ceiling. It is a waypoint.