Amazon Bedrock brings in-region Claude 5 inference to London
AWS has added in-region Amazon Bedrock support for Claude Opus 5.5 and Sonnet 5 in London, keeping inference requests and data inside the called UK Region.

Key takeaways · 3
- 01
UK Bedrock users can invoke Claude Opus 5.5 and Sonnet 5 while keeping inference in the London Region.
- 02
Teams evaluating generated prose should test for model-specific patterns rather than relying on older, generic AI-writing tells.
- 03
Data residency and output quality require separate controls because regional processing does not remove linguistic model signatures.
Claude inference stays in-region
Amazon Bedrock now supports Anthropic’s Claude Opus 5.5 and Claude Sonnet 5 with in-region inference in the UK (London) Region. [1] Customers can invoke both models through the bedrock-runtime endpoint in London’s eu-west-2 Region. [1] AWS says inference requests and data are processed within the called Region and never leave it, allowing customers with UK data-processing requirements to keep inference in-country. [1]
Opus retains writing tells
Graphite used 10,000 articles published before ChatGPT as a human control group, then had AI models rewrite the articles from summaries to reduce source bias. [2] The study identified 13,000 phrases that appeared at least twice as often in AI content as in human content. [2] For Claude Opus 5.5, “dependable” appeared 23 times more often, “this matters” appeared 116 times more often, and “why X matters” appeared 92 times more often than in human samples. [2] Graphite chief AI officer Greg Druck said Claude models are getting closer to the human word distribution over time, while GPT models are getting further away. [2]
What it means
Together, the sources highlight two distinct considerations for Claude Opus 5.5 adopters: Bedrock can keep inference inside the London Region, while generated prose can retain statistically detectable habits. Graphite’s reported direction also separates Claude from GPT models: Claude’s word distribution is moving closer to human writing over time, while GPT’s is moving further away. Deployment location and output review therefore remain separate evaluation questions. What the sources don't address: whether London-hosted model performance, pricing, or writing patterns differ from deployments in other Bedrock regions.
The Bedrock expansion gives UK organizations another way to use frontier Claude models while keeping inference processing inside the country. Graphite’s findings also show why output review should account for model-specific language patterns rather than depending on familiar but outdated AI-writing indicators.
Why it matters
Put this to work — one session a day, built for your industry.
Create a free account for a daily session — eight questions and one real-work challenge, on the news that affects your role.
Start freeHow this developed
2 October 2026
Amazon Bedrock brings in-region Claude 5 inference to London
2 October 2026
Event created from source cluster.