AI News Of The Week (1st May, 2026)
TLDR: Key AI Developments This Week
OpenAI and Microsoft rewrote one of AI's most important contracts, ending Microsoft's exclusive resale rights while keeping them as primary cloud partner through 2032. OpenAI then expanded to AWS, launching GPT-5.5, Codex, and Managed Agents on Amazon Bedrock. Google made Gemini much more useful by adding file generation for PDFs, DOCX, XLSX, and Google Workspace formats directly from prompts. Anthropic launched Claude for Creative Work with connectors for Adobe, Blender, Ableton, and other creative tools. Microsoft committed to $190 billion in 2026 capex with Azure growing 40% and Copilot reaching 20 million paid users. Meta raised its AI infrastructure forecast to $145 billion and sold $25 billion in bonds. The Pentagon signed classified AI deals with seven companies, notably excluding Anthropic over supply-chain concerns.
OpenAI and Microsoft Rewrite One of AI's Most Important Contracts
On April 27, OpenAI and Microsoft announced an amended agreement that keeps Microsoft as OpenAI's primary cloud partner but ends Microsoft's exclusive rights to resell OpenAI products. The new terms also let OpenAI serve customers across other cloud providers, keep Microsoft's license to OpenAI IP through 2032 on a non-exclusive basis, and preserve OpenAI revenue-share payments to Microsoft through 2030 under a cap. Read more
OpenAI Brings GPT-5.5, Codex, and Managed Agents to AWS
On April 28, OpenAI and AWS expanded their partnership to launch OpenAI models including GPT-5.5 on Amazon Bedrock, Codex on AWS, and Amazon Bedrock Managed Agents powered by OpenAI in limited preview. The move matters because it turns OpenAI into a much more cloud-agnostic platform right after the Microsoft contract rewrite, giving enterprises new ways to buy and run frontier models inside AWS security, compliance, and procurement workflows. Read more
Google Makes Gemini Much More Useful as a Work Output Tool
On April 29, Google rolled out file generation in the Gemini app, letting users create PDFs, DOCX, XLSX, Google Docs, Sheets, Slides, CSV, Markdown, and other formats directly from a prompt. Rather than stopping at chat answers, Gemini can now jump straight to a shareable artifact, which makes the product look more like a real productivity workspace than just an assistant. Read more
Anthropic Pushes Claude Into Creative Software Workflows
On April 28, Anthropic launched Claude for Creative Work, adding connectors for tools including Adobe, Ableton, Affinity by Canva, Autodesk Fusion, Blender, Resolume, SketchUp, and Splice. Anthropic's framing was clear: Claude is being positioned not just as a chat assistant for ideas, but as something that can sit inside real creative pipelines and help with repetitive production work, documentation, scripting, and tool-heavy workflows. Read more
Microsoft Shows AI Demand Is Big Enough to Justify a $190 Billion Spend Plan
In results published April 29, Microsoft said it expects $190 billion in calendar 2026 capital spending, above analyst expectations, and reported Azure revenue growth of 40%. Reuters also reported that Microsoft's paid Microsoft 365 Copilot user base rose to 20 million from 15 million in January, showing that the AI race is increasingly being measured in cloud capex and paid seat growth, not only in model launches. Read more
Meta Doubles Down on AI Infrastructure and Sells $25 Billion in Bonds
Meta raised its 2026 capital expenditure forecast to $125 billion to $145 billion for AI infrastructure and followed with a $25 billion bond sale reported on April 30. Reuters described the financing as part of Meta's effort to fund its AI buildout, underlining how even the biggest tech companies are leaning harder on debt markets as AI infrastructure costs keep climbing. Read more
The Pentagon Signs New Classified AI Deals and Anthropic Is Left Out
On May 1, the Pentagon said it had reached agreements with seven AI companies—SpaceX, OpenAI, Google, NVIDIA, Reflection AI, Microsoft, and Amazon Web Services—to bring their tools into classified Defense Department networks. Reuters reported that Anthropic was excluded because of an earlier supply-chain risk ruling, making this one of the clearest signs yet that access to government AI work is starting to split the frontier labs into those willing to accept national-security deployment terms and those pushing back on them. Read more
