AI writes 75% of Google’s new code, says Sundar Pichai: What it means for engineers

A major shift is underway at Google team, where artificial intelligence is now doing most of the coding work. Sundar Pichai said around 75 per cent of new code at the company is now generated using AI systems. The update was shared during Google’s Cloud Next 2026 event in Las Vegas, highlighting how quickly AI is changing software development.

This marks a sharp rise from 50 per cent just six months ago, showing how rapidly AI tools are being adopted within the company.

AI writes the code, engineers stay in control

Even though AI is generating most of the code, human engineers remain an essential part of the process.

Pichai explained that every piece of AI-generated code is reviewed and approved by engineers. This means AI acts more like a fast assistant, creating drafts, while humans ensure accuracy and reliability. This approach is helping teams work faster without compromising quality. Engineers are no longer writing every line themselves but are focusing more on reviewing, guiding, and improving outputs.

What are ‘agentic workflows’?

Google is moving towards what Pichai calls “agentic workflows.”

In simple terms, this means engineers now use multiple AI agents that can:

Work independently on tasks

Collaborate with each other

Improve solutions step by step

Pichai described engineers as people who “orchestrate digital task forces” made up of AI agents. One example shared by the company showed that a large coding project was completed six times faster than similar work done a year earlier.

Faster apps and product development

AI is also speeding up how Google builds products.

Pichai pointed to the development of the Gemini app for macOS, where teams reportedly built a working application in just a few days using an internal system called Antigravity. Google is testing these tools internally first, acting as what Pichai called a “customer zero” before offering them to businesses.

This helps the company understand how well the technology works at scale.

What this means for the tech industry

Google’s shift reflects a broader trend across the tech world.

AI tools are increasingly:

Writing code

Finding errors

Suggesting improvements

Sam Altman has also noted how quickly software development is evolving, with AI systems now able to perform tasks in seconds that once took hours.

This does not mean engineers are becoming less important. Instead, their role is changing.

They are now focusing more on:
Decision-making
Oversight
System design
A new way of coding

The rise of AI in coding is not just about speed. It is changing how software is built. At Google, the process is now less about writing code line by line and more about guiding intelligent systems to produce the right results. As AI continues to improve, this model could become the standard across the industry, reshaping the future of work for software engineers worldwide.