Google is pitching Cloud Next '26 as the moment the enterprise stack resets around agentic AI.
The company's core argument: only Google owns the full stack from custom silicon to productivity suite, and that vertical integration is the only way to scale AI beyond pilots.
CEO Sundar Pichai opened the keynote with a data point meant to signal internal conviction: roughly 75% of all new Google code is now AI-generated, up from about 25% a year ago. Google Cloud CEO Thomas Kurian followed by declaring "the era of the pilot is over, the era of the agent is here," framing the company's pitch as a unified stack with chips designed for models, models grounded in customer data, and agents secured by Google's infrastructure. The centerpiece is a full rebranding. Vertex AI is now the Gemini Enterprise Agent Platform.
Agentspace has been absorbed into a unified Gemini Enterprise product. Workspace Studio gives business users a no-code agent builder across Gmail, Docs, Sheets, Drive, Meet, and Chat, type "every Friday, ping me to update my tracker" and Gemini creates the automation. The Model Garden hosts more than 200 models including Anthropic's Claude and open-weight alternatives.
But analysts who watched the keynote say the vision comes with real ambiguity. HyperFRAME Research's Stephanie Walter told CIO.com that "Google announced a lot at once, and the way the AI product portfolio fits together is still somewhat unclear." CIOs will like the ambition while still asking for "a cleaner map of where Gemini Enterprise, the Agent Platform, the Application, and the data layer begin and end." The competitive market makes that confusion costly. AWS Bedrock with its Agents framework is maturing fast.
Microsoft Copilot sits inside virtually every Fortune 500 company already. OpenAI's enterprise revenue now accounts for 40% of its total, with three million weekly users on Codex alone.
Anthropic's Model Context Protocol has reached 10,000 servers and 97 million monthly SDK downloads.
Google Cloud holds roughly 11% of cloud infrastructure market share against AWS at 31% and Azure at 25%. But it exited Q4 2025 with 50% year-on-year growth, the fastest of the three hyperscalers.
Andi Gutmans, who runs Google Cloud's data business, argued in an interview with The Register that no competitor combines cloud infrastructure, frontier models, and a data platform under one roof. "If you think about AWS and Azure," he said, "they've got the infrastructure, they don't have the model." Data providers have platforms but need infrastructure from others.
"The AI model providers just do the AI model."
Gutmans said Google has roughly 80 data-related announcements at the conference this week and has re-engineered every agent in its data portfolio over the past year after Gemini 2.5 crossed a reasoning threshold that made older approaches obsolete. On hardware, Google announced eighth-generation TPUs: TPU 8t delivering 3x performance per pod with support for up to 9,600 TPUs via 3D torus topology, alongside TPU 8i for inference workloads. Anthropic has committed to up to one million Ironwood units from Google's previous-gen TPU line.
The most strategically significant announcement may be invisible to end users: A2A (Agent2Agent) protocol has reached 150 organizations in production, not pilot. Microsoft, AWS, Salesforce, SAP, and ServiceNow are all running A2A in production environments routing tasks between agents built on different platforms. The protocol is now governed by the Linux Foundation's Agentic AI Foundation at version 1.2.
SiliconANGLE's John Furrier framed what's really at stake in a pre-conference analysis: this is not a model war but a control plane war for how enterprise work gets done. When agents become users instead of tools, SaaS pricing built around human logins breaks down. The app becomes a feature rather than a product.
But execution risk remains high. Google's enterprise go-to-market has historically lagged its technology. And as CIOs evaluate pitches from AWS, Microsoft, and Google that sound strikingly similar on paper, unified stacks for agentic AI, Walter warned that underlying trade-offs in pricing models, ecosystem depth, and interoperability remain difficult to parse at an operational level.
Ashish Chaturvedi of HFS Research told CIO.com that most enterprises are drowning in integration tax compounded by fragmented stacks. Google's pitch directly addresses that pain point. But he added that enterprises will need stronger FinOps disciplines to regain control over increasingly complex AI spending.
