Google reveals AI now generates about three-quarters of its new internal code, with human engineers reviewing the output, marking a rapid rise from 50% late last year and 25% in October 2024.

CEO Sundar Pichai described a move toward truly agentic workflows where engineers collaborate with more autonomous AI systems capable of handling broader engineering tasks beyond code completion.

Overall, Google is accelerating a company‑wide shift to AI‑generated code and AI‑assisted work processes that empower engineers across the organization.

Alphabet signaled that AI investments are translating into business performance, projecting capital expenditures of $175–$185 billion for 2026 after $91.4 billion in 2025.

The disclosures precede Google Cloud Next 2026, where Gemini Enterprise was spotlighted as a platform to support AI agents across organizations.

At Cloud Next 2026, Google pushed monetization of AI through enterprise offerings, unveiling Gemini Enterprise and a revamped Vertex AI with new governance and security features for production‑ready AI agents.

Pichai acknowledged the mounting challenge of coordinating large networks of AI agents, stressing the need for oversight as adoption scales beyond individual systems.

Google presents its internal AI‑driven approach as a signal of readiness for broader commercial use of agent‑based tools, expanding enterprise use from traditional ML to rapid deployment of custom AI agents.

The company indicated its internal AI adoption will influence external product strategy, with further announcements planned for I/O in May.

Google is promoting widespread use of AI tools beyond coding to accelerate workflows and boost productivity across tasks.

Industry dynamics show Google facing competition in AI coding tools, with Bloomberg noting attempts to unify coding initiatives as Anthropic gains ground, and Business Insider reporting friction from DeepMind teams using Claude Code alongside Gemini tooling.

Some DeepMind staff have been allowed to use Anthropic’s Claude Code, fueling internal tensions as Google strives to prove its coding stack can compete externally.