Google CEO Says AI Generates 75% of Company Code, Launches Deep Research

Google AI Generates Three-Quarters of Internal Code

Google CEO Sundar Pichai revealed that artificial intelligence now generates 75% of the code written at the search giant, marking a dramatic shift in how one of the world’s largest technology companies develops software. The disclosure came as Google announced significant upgrades to its autonomous research capabilities through new Deep Research and Deep Research Max agents.

According to The Times of India, Pichai emphasized that this AI-generated code undergoes human review and approval before implementation. The milestone represents Google’s aggressive internal adoption of AI tools to accelerate development cycles and increase programmer productivity across its vast engineering organization.

Deep Research Agents Target Enterprise Workflows

Google on Monday launched Deep Research and Deep Research Max, two autonomous research agents built on the Gemini 3.1 Pro model that can fuse open web data with proprietary enterprise information through a single API call. The agents can produce native charts and infographics inside research reports and connect to third-party data sources through the Model Context Protocol (MCP).

“We are launching two powerful updates to Deep Research in the Gemini API, now with better quality, MCP support, and native chart/infographics generation,” Pichai wrote on X. The release targets enterprise research workflows in finance, life sciences, and market intelligence — industries where information accuracy carries high stakes.

VentureBeat reported that the upgrade represents Google’s clearest bid to position its AI infrastructure as the backbone for enterprise research workflows, allowing developers to conduct exhaustive, multi-source research that traditionally consumed hours or days of human analyst time.

DeepMind Spinoff Prepares AI-Designed Drug Trials

Isomorphic Labs, the UK-based biotech spinoff of Google DeepMind, will soon begin human trials of drugs designed by its Nobel Prize-winning AlphaFold AI technology. Company president Max Jaderberg announced at WIRED Health in London that the company is “gearing up to go into the clinic.”

According to Wired, the timeline represents a delay from CEO Demis Hassabis’s earlier prediction that AI-designed drugs would enter clinical trials by the end of 2025. Isomorphic Labs, founded in 2021, uses DeepMind’s AlphaFold platform that predicts protein structures for drug discovery.

The company released AlphaFold 3 in 2024, which advanced beyond modeling proteins in isolation to predicting other important molecules like DNA and RNA. This progression from basic protein structure prediction to drug design represents a significant milestone in applying AI to pharmaceutical development.

AlphaFold’s Evolution from Research to Medicine

DeepMind’s AlphaFold solved a decades-old challenge in biology by accurately predicting protein structures. Researchers had attempted this since the 1970s, but the astronomically high number of possible protein shapes made manual prediction painstaking.

AlphaFold 2’s breakthrough in 2020 used deep-learning techniques to achieve stunning accuracy. Google released an open-source version in 2021, making the technology available to researchers worldwide. The platform has already revolutionized scientists’ understanding of proteins, with its drug design capabilities now facing real-world testing.

Enterprise AI Adoption Accelerates Across Google

The 75% AI-generated code figure illustrates Google’s comprehensive internal AI adoption beyond consumer-facing products. This internal transformation spans software development, research automation, and drug discovery through its various subsidiaries and partnerships.

Google’s approach contrasts with competitors who primarily focus on external AI products. By deeply integrating AI into its own operations first, Google gains practical experience with AI limitations and capabilities before scaling solutions to enterprise customers.

The company’s dual strategy of internal AI adoption and external enterprise tools positions it to compete with Microsoft’s Copilot suite and other enterprise AI platforms. The Deep Research agents specifically target knowledge work that requires synthesizing information from multiple sources — a core enterprise use case.

What This Means

Google’s disclosure that AI generates 75% of its internal code represents a watershed moment for enterprise AI adoption. When one of the world’s most sophisticated technology companies relies so heavily on AI for core development work, it signals that AI-assisted programming has moved beyond experimental to production-critical.

The simultaneous launch of Deep Research agents and progress on AI-designed drugs shows Google’s strategy to lead in both horizontal AI tools and vertical applications. By proving AI’s value internally first, Google builds credibility for selling similar capabilities to enterprises.

The timeline delays for Isomorphic Labs’ drug trials highlight the gap between AI breakthroughs and real-world implementation, particularly in regulated industries like pharmaceuticals. Success in human trials would validate AI’s potential to accelerate drug discovery beyond just research efficiency gains.

FAQ

How does Google ensure quality when 75% of code is AI-generated?
Google maintains human review and approval processes for all AI-generated code before implementation. The company treats AI as an accelerant for human developers rather than a replacement, with engineers responsible for validating and integrating AI suggestions.

What makes Deep Research different from existing AI research tools?
Deep Research agents can combine open web data with proprietary enterprise information through a single API call, generate native charts and infographics, and connect to third-party data sources. This integration capability distinguishes it from tools that only access public information.

When will Isomorphic Labs’ AI-designed drugs reach patients?
Isomorphic Labs plans to begin human clinical trials soon, though specific timelines weren’t disclosed. The company previously targeted trials by end of 2025, but drug development timelines often extend due to regulatory requirements and safety protocols.