top of page

Inside Google Next: Agents, AI, and What’s Coming Next for the Cloud

  • RedCloud
  • 3 days ago
  • 2 min read


Attending Google Next this year felt less like a tech conference and more like a glimpse into the operating system of the future enterprise. With major product launches and a clear roadmap for AI integration, our clients and friends at Google made one thing abundantly clear: they’re betting big—and smart—on the future of cloud and AI.


$75 Billion in Cloud Investment

Google is investing $75 billion in capex through 2025 to expand its cloud servers and data centers. That level of commitment reinforces Google Cloud’s central role in scaling secure, high-performance AI workloads.


New TPU: Ironwood

The latest TPU (Tensor Processing Unit), Ironwood, doubles the performance per watt of the previous gen (Trillium) and offers 6x more high bandwidth memory. This upgrade supports more powerful, efficient AI model training, critical for enterprise-scale use cases.


Security-First AI

A consistent theme throughout the conference was Google’s focus on security and compliance-first AI. From model design to deployment, trust and privacy were front and center, especially important for highly regulated industries.


Agents Everywhere

"Agent" was the word of the week. These aren't just models—they're task-completing AI entities. Google announced:


  • AgentSpace: A new platform to embed agents in enterprise apps to complete tasks like scheduling meetings or writing content.

  • Data Agents: Handle metadata generation, pipeline creation, and natural language analysis.

  • Looker Agents: Carry out follow-up queries, summarize data in slide decks or even podcasts.


Google also launched an Agent Development Kit, including a new Agent2Agent protocol, allowing AI agents to work together across systems.


Smarter AI Starts with Better Data

Sessions emphasized that AI quality hinges on solid data prep and governance. Our key takeaways include:


  • Automate governance with Gen AI for ambient oversight

  • Treat governance as a dedicated platform

  • Standardize data inputs and monitor model outputs continuously


This is increasingly central to how we approach enterprise data strategy at RedCloud.


Gemini Everywhere

Google’s latest reasoning model, Gemini 2.5, is now integrated across tools. It’s powering smarter interactions—from summarizing content to generating insights—without the need for advanced prompting.


Relevant Product Updates

Several product upgrades stood out for our consulting work:


  • BigQuery Dataframes (Bigframes): A Python-to-SQL converter that lets large datasets run efficiently on BigQuery’s compute power.

  • Multimodal in BigQuery: Combine structured and unstructured data in Pandas-like workflows.

  • AppSheet + Gemini: Adds image and PDF processing to low-code environments.

  • Looker + Gemini: Enhanced semantic analysis with natural language prompts. Future capabilities will let users build LookML models with AI.


Final Take

Google Next confirmed the direction is clear: AI agents, secure data infrastructure, and smart governance are redefining how enterprises operate. For us at RedCloud, the tools and strategies shared provide a practical path forward as we help clients embrace this AI-first era, with speed, responsibility, and scale.


Partner with us to develop the right strategy, governance, and architecture, unlocking the transformative potential of AI and turning your data into your organization’s most valuable asset. Learn more at: redcloudconsulting.com/data-and-ai.

bottom of page