Synopsis
Google plans a massive $185 billion investment in artificial intelligence infrastructure this year. CEO Sundar Pichai highlighted AI's growing role in software development and security. New platforms like Gemini Enterprise Agent and advanced chips like TPU 8t and TPU 8i are being introduced. This significant spending signals Google's strong focus on enterprise AI solutions and future growth.Listen to this article in summarized format
Addressing the Google Cloud Next 2026 summit here, Pichai said more than 50% of Google's machine learning compute resources will be dedicated specifically to its cloud business, signaling a massive prioritisation of enterprise customers.
The company's ballooning capital expenditure, mostly to harness artificial intelligence (AI) capabilities, will double this year from $91.4 billion in 2025, representing a nearly sixfold increase over its $31 billion investment just four years ago in 2022. Alphabet's stock had fallen by as much as 5% after the announced spending plan blew past forecasts, with investors questioning the scale and sustainability of big tech's investment plans. However, the stock has recovered since then amid the tumult of the Iran conflict.
But signalling the company's increasing reliance on AI resources, Pichai revealed to the annual gathering that nearly 75% of all new code at Google is now AI-generated and subsequently approved by engineers. In a techtonic shift in software development, this figure has jumped from just 15% in 2023.
Meanwhile, using AI agents to triage tens of thousands of unstructured threat reports each month, Google’s security teams have reduced the time required for threat mitigation by over 90%, he said.
The CEO said Google’s internal marketing teams also used Gemini and Chrome to generate thousands of creative variations, resulting in 70% faster turnaround times and a 20% increase in conversion rates through large-scale personalisation.
Major announcements
Pichai also announced the Gemini Enterprise Agent Platform, a new enterprise agent platform to address the complexity of managing "thousands" of AI agents. The new developer platform is an evolution of Google's Vertex AI. It brings together a full suite of models, development, and tuning services, with new features for businesses to build, scale, govern, and optimize agents that can work autonomously to execute complex business workflows.
First launched last year, Pichai said Gemini Enterprise saw 40% growth in paid monthly active users quarter-over-quarter in Q1.
The company also announced the eighth generation of Google's custom Tensor Processor Unit (TPU), TPU 8t and TPU 8i. Amin Vahdat, chief technologist, AI and Infrastructure at Google Cloud, said the two chips are designed to power Google's custom-built supercomputers, drive cutting-edge model training and agent development, and work through massive inference workloads.
Google also introduced the Agentic Data Cloud, an AI-native architecture that evolves the enterprise data platform from a static repository into a dynamic reasoning engine. Major global firms such as Vodafone, American Express and Virgin Voyages are already using the cloud to manage and run thousands of AI agents.
Leveraging the company's largest ever acquisition, that of cloud security firm Wiz for a reported $32 billion, Google has also announced three new agents In Google Security Operations can help hunt threats, engineer detections, and provide context on third parties.
(The writer is at Google Cloud Next 2025 in Las Vegas at the invitation of Google.)
