Urgent.News

What's breaking now, across thousands of outlets.

AI

Every Cloud Now Sells Claude, GPT, and Grok. Who Runs Them Is Where They Differ.

Take Maya, the ML platform lead at a US health insurer. Her shortlist of three models clears every cloud's frontier catalog in an afternoon, then breaks on the cheapest model on the list and on a footnote about who actually runs Claude. In her brief, engineering wants Claude Fable 5.2 for hard coding work. The claims team has built on OpenAI's GPT-6.2 Sol. A document-triage pipeline needs a cheap…

Maya, an ML platform lead at a US health insurer, has three models on her shortlist from the "every cloud" frontier catalog: Claude, GPT, and Grok. Depending on the application, her engineering team prefers Claude, while the claims team uses OpenAI's GPT-6.2 and a document-triage pipeline needs a cheap open-weight model like Kimi or Qwen.

The key point of contention is that Claude can be billed separately from the cloud provider. Anthropic-run Claude is processed outside the cloud's boundary, providing more control and compliance with US data residency requirements. Microsoft's license to OpenAI's models became non-exclusive in April, allowing OpenAI to serve its products on any cloud. Google has launched its own Gemini platform, with Gemini 4 Neon, Hydrogen, and Carbon in preview, while Bedrock and Foundry do not offer Gemini models.

For Maya, Kimi is available on Bedrock and Foundry, but each offers different terms and processing locations. Bedrock runs its open-weight catalog itself, while Foundry runs the models on Azure infrastructure. Google offers serverless GPU endpoints on its own platform, where users manage the infrastructure. The choice of cloud and model provider significantly impacts privacy, performance, and compliance considerations for AI applications.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

How 5 Companies Use AI Coding Agents and Assistants Across the SDLC

An agent-generated pull request has a green build and a full set of tests. A reviewer compares it with the ticket and discovers that the agent guessed how duplicate requests should behave.

  • AI coding agents assist at each SDLC stage from requirements to merge
  • Agents prepare tasks, engineers verify implementations, reviewers ensure test coverage
  • Approved plan guides process; engineers decide on failures and merge owner

More from Thursday 19 November →