Google CEO Says AI Generates 75% of Company Code as Deep Research Agents Launch

Google CEO Sundar Pichai revealed that AI now generates 75% of the code at the tech giant, coinciding with the launch of two new autonomous research agents that can search both web data and private enterprise information. The milestone underscores Google’s aggressive internal adoption of AI tools while positioning the company as a leader in AI-powered enterprise workflows.

According to multiple reports, Pichai disclosed the 75% figure during recent public statements, marking a significant jump from earlier estimates. The statistic reflects Google’s comprehensive integration of AI coding assistants across its development teams, from search algorithms to cloud infrastructure.

Deep Research Agents Target Enterprise Markets

Google on Monday launched Deep Research and Deep Research Max, two new AI agents built on the Gemini 3.1 Pro model that can conduct autonomous research across multiple data sources. According to VentureBeat, the agents represent “the most significant upgrade to its autonomous research agent capabilities since the product’s debut.”

The new agents can fuse open web data with proprietary enterprise information through a single API call, produce native charts and infographics inside research reports, and connect to third-party data sources through the Model Context Protocol (MCP). Pichai announced on X that the agents offer “better quality, MCP support, and native chart/infographics generation.”

Key capabilities include:

• Multi-source research: Combining public web data with private enterprise databases
• Native visualization: Generating charts and infographics within reports
• Third-party integration: Connecting to external data sources via MCP
• API accessibility: Available through the Gemini API for developer integration

AlphaFold Technology Advances to Human Drug Trials

Isomorphic Labs, Google DeepMind’s biotech spinoff, is preparing to begin human trials of drugs designed using its Nobel Prize-winning AlphaFold technology. According to Wired, Isomorphic Labs president Max Jaderberg confirmed the company is “gearing up to go into the clinic” at WIRED Health in London.

The trials represent a crucial test of AlphaFold’s ability to design safe and effective drugs beyond its proven success in protein structure prediction. AlphaFold revolutionized protein research by solving the decades-old challenge of predicting how amino acid chains fold into three-dimensional structures.

Isomorphic Labs was founded in 2021 as an Alphabet subsidiary focused on drug discovery. The company released AlphaFold 3 in 2024, which expanded beyond protein modeling to predict DNA, RNA, and other molecular interactions. CEO Demis Hassabis had previously targeted clinical trials by the end of 2025, though Jaderberg did not provide an updated timeline.

Enterprise AI Integration Accelerates

Google’s 75% AI code generation figure signals a broader trend of enterprise AI adoption across the technology sector. The statistic suggests that three-quarters of Google’s software development now relies on AI assistance, from initial code generation to debugging and optimization.

This level of integration extends beyond simple code completion tools. Google’s AI systems likely handle complex tasks including:

• Architecture design: Suggesting optimal system designs and data structures
• Code review: Identifying bugs, security vulnerabilities, and performance issues
• Testing automation: Generating comprehensive test suites and edge cases
• Documentation: Creating technical specifications and user guides

The Deep Research agents target industries where comprehensive analysis is critical, including finance, life sciences, and market intelligence. VentureBeat noted that these are “industries where the stakes of getting information wrong are extraordinarily high.”

Government and Academic Partnerships Expand

Google DeepMind’s influence extends into government circles, with Financial Times reporting that Morgan McSweeney, a senior UK political figure, held discussions with the AI research lab about potential projects. While details remain limited due to paywall restrictions, the talks suggest growing government interest in leveraging DeepMind’s capabilities for public sector applications.

Academic partnerships have also accelerated since AlphaFold’s open-source release in 2021. The platform has enabled researchers worldwide to access protein structure predictions that previously required months of computational work. Over 200 million protein structures are now available through the AlphaFold database, supporting research in drug discovery, agricultural science, and environmental biology.

What This Means

Google’s 75% AI code generation milestone represents a fundamental shift in software development practices at one of the world’s largest technology companies. This level of AI integration suggests that human developers are transitioning from writing code directly to orchestrating AI systems that generate, review, and optimize software.

The simultaneous launch of Deep Research agents positions Google to capture enterprise customers seeking comprehensive AI-powered analysis tools. By combining web search capabilities with private data access, Google is addressing a key limitation of existing AI research tools that operate in isolation from proprietary information.

Isomorphic Labs’ progression to human trials marks a critical validation point for AI-designed drugs. Success in clinical trials could accelerate pharmaceutical industry adoption of AI-driven drug discovery, potentially reducing the typical 10-15 year development timeline and multi-billion dollar costs.

FAQ

What does “75% AI-generated code” mean for Google’s development process?
This statistic indicates that three-quarters of Google’s software code is now written by AI systems rather than human developers directly. Developers likely focus on high-level design, problem-solving, and AI system orchestration while AI handles routine coding tasks, debugging, and optimization.

How do Deep Research agents differ from existing AI research tools?
Deep Research and Deep Research Max can access both public web data and private enterprise databases through a single API call, generate native visualizations within reports, and connect to third-party data sources. Most existing AI research tools are limited to either public or private data, not both simultaneously.

When will AI-designed drugs from Isomorphic Labs reach the market?
Isomorphic Labs is preparing for human clinical trials but has not announced specific timelines. Even successful trials would require additional phases of testing and regulatory approval, typically taking several years before reaching patients. The company previously targeted clinical trials by end-2025 but appears to be running behind that schedule.