This Week in Cloud — May 29, 2025
Welcome back to The Cloud Cover, your weekly briefing on the most critical developments across the cloud computing landscape. This week, the "Age of AI Agents" wasn't just a theme at developer conferences; it was a reality hitting service catalogs. Buckle up – the bots are building.
🤖 The Agents Are Here. Are You Ready?
Forget chatbots fetching weather reports. This week, all major cloud providers unleashed or significantly upgraded AI "agents" designed for heavy lifting in development and operations. AWS launched its Transform service, using AI agents to automate complex tasks like migrating VMware and mainframe applications or refactoring .NET code to Linux, promising massive speed boosts and cost cuts.
Meanwhile, Microsoft, still buzzing from Build 2025, pushed hard on its agentic vision. It introduced Microsoft Entra Agent ID to give these AI workers unique identities for better governance, rolled out Copilot Tuning to let businesses train agents on their own data, and even proposed "NLWeb" as an open protocol for agents to communicate. Google wasn't sitting still either, showcasing Gemini Code Assist Agents that can handle tasks from code migration to test generation and co-launching an "Agent2Agent Protocol" with over 50 partners to foster interoperability.
The Takeaway: This isn't just about faster coding. It's a fundamental shift towards AI-driven software development, deployment, and even scientific discovery. For architects and IT leaders, the implications are huge: How do you manage, secure, and govern these autonomous agents? How do you re-skill teams? And how do you navigate the inevitable interoperability challenges? The age of agents is here, and it demands new strategies, now.
🔍 The Rundown
New Database Options: Amazon Aurora DSQL, a serverless, active-active multi-Region SQL database, hit general availability, promising strong consistency and high uptime.
Strategic Partnerships: AI-focused collaborations were announced with FICO (AI-driven decisioning) and Elastic (AI search platform integration with Bedrock)
More MCP: Amazon Neptune launched an open-source Model Context Protocol (MCP) server, aiming to make it easier to connect graph databases with generative AI tools.
New AI Models and Tools: AI Foundry expanded, adding xAI's Grok models and new tools like a Model Leaderboard and Model Router
Multicloud AI: Azure API Management's AI Gateway now supports AWS Bedrock endpoints, letting Azure customers manage Bedrock calls with Azure policies.
Security Scare: A critical security incident hit Commvault's Metallic SaaS platform, hosted on Azure, highlighting supply chain risks and the need for tight access controls.
New AI Offerings: Google I/O brought new generative AI models to Vertex AI: Veo 3 (video), Imagen 4 (image), and Lyria 2 (music).
New Deals: Travel-tech provider Amadeus signed a strategic deal to move parts of its platform to GCP, leveraging its infrastructure and AI services
📈 Trending Now: The Great GPU Grab Heats Up
If AI agents are the new workforce, the cloud providers are in a frantic race to build their digital factories. This week hammered home the intensity of the AI Infrastructure Race. AWS announced its Blackwell-powered P6-B200 instances. Google countered by highlighting its 7th-gen Ironwood TPUs and being the first to offer both A4 and A4X VMs with NVIDIA's latest Blackwell GPUs, even promising future Vera Rubin GPUs. Microsoft, meanwhile, reported massive Q3 capital expenditures (~$20 billion) primarily aimed at AI and cloud infrastructure.
This isn't just about bragging rights; it's a high-stakes gamble on who can provide the most power, most efficiently, for the exploding demand for AI training and inference. For users, this means more options but also potential complexity in choosing the right chip/VM for the job and navigating availability. Keep an eye on how these hardware wars impact pricing, performance benchmarks, and the strategic partnerships (like Google's deep dive with NVIDIA) that will shape the AI landscape.
📅 Event Radar
29-31
Last week! See the sessions on YouTube.
9-11
Cybersecurity, AI, data security, and risk management.
10-11
Session catalog now available
20
Free registration
💼 Job Spotlight
Field Solutions Architect I, Generative AI at Google Cloud
$123,000-$176,000 | Multiple US
Build and prototype cutting-edge Generative AI applications for top Google Cloud customers, influencing product direction and scaling AI innovation across industries.
Solutions Architect, Alexa+ at Amazon
$118,200-$204,300 | Multiple, US
Architect next-gen AI experiences by integrating top-tier brands with Alexa+, Amazon’s generative AI-powered personal assistant built from the ground up for voice and multimodal interactions.
👋 Until Next Week
This week felt like a significant acceleration, especially in the practical application of AI through agents. Until next week, keep building, keep securing, and maybe keep an eye on your AI assistants.
Do you enjoy these emails? Your friends and colleagues might, too! Help us grow the cloud community by sharing the newsletter with others.