What happened
According to a post on the Strands Agents blog, Amazon moved production workloads from Anthropic's Sonnet 4.6 to its own Strands Decider solution, saving millions of dollars in the process.
The story drew attention on Hacker News, reaching 200 points. The discussion centers on what a large cloud provider shifting from an external model to its own system means for cost and dependency.
Why it matters
Major cloud companies treat model choice as a direct cost line in AI workloads. Amazon's move to an in-house solution can be read as an example of the push to reduce reliance on external model providers and lower inference costs.
Limited information
The source text consists only of a headline; the exact savings, the technical details of Strands Decider and which workloads were affected are not shared. The figures and scope should therefore not be treated as verified.
