Anthropic brings full Claude platform to AWS with native API parity

If you build on AWS, using Anthropic's frontier models has always felt like a game of trade-offs: wait weeks for Bedrock to roll out the newest native features, or juggle a second vendor contract and set of API keys directly with Anthropic.

That friction just vanished. Anthropic officially launched the Claude Platform on AWS, making it generally available today. It basically pipes the full, native Claude platform directly into your AWS account, complete with IAM authentication, CloudTrail audit trails, and unified AWS Marketplace billing that burns down your existing enterprise spend commitments.

Day-one parity with zero wait time

Unlike traditional cloud platform integrations that lag behind official releases, this rollout guarantees day-one parity with Anthropic's native API. Whenever Anthropic ships a new model or beta feature, AWS customers get it instantly.

The initial feature lineup includes Claude Managed Agents, the advisor strategy, web search, sandboxed Python code execution, Skills, and the remote MCP connector. Developer staples like Prompt Caching, Citations, batch processing, and the Files API are supported out of the box across Claude Opus 4.7, Sonnet 4.6, and Haiku 4.5. Developers also get direct access to the Claude Console, and tools like Claude Code and Claude Cowork can point to an AWS workspace in just a few lines of configuration.

One crucial difference: the data boundary

Naturally, you might wonder where this leaves Amazon Bedrock. The real distinction comes down to who handles your data.

On Amazon Bedrock, AWS acts as the data processor, meaning your requests never leave AWS security and regional boundaries. If you operate in healthcare, government, or heavily regulated finance with strict data residency rules, Bedrock remains your home.

In contrast, the Claude Platform on AWS is operated directly by Anthropic. Your requests are processed outside AWS's perimeter. You get the operational simplicity of AWS billing and IAM credentials, paired with the raw speed of Anthropic's platform. One financial catch: existing Bedrock private pricing discounts will not automatically or retroactively apply, so enterprise teams need to sync with their account reps before switching over.

Why it matters

For engineering teams, this removes the single biggest roadblock to adopting Anthropic's agentic ecosystem. You no longer have to beg procurement for a new software vendor approval just to test Claude Code or build MCP workflows; your team can spin up workspaces under existing AWS guardrails this afternoon.

Strategically, Anthropic pulled off a brilliant maneuver. Instead of letting a hyperscaler relegate Claude to a commoditized API dropdown inside a cloud console, Anthropic effectively reduced AWS to an identity and billing layer while retaining direct control over the developer experience and telemetry.

This move puts immediate pressure on rival foundation model labs and traditional model hubs. If developers can get bleeding-edge agents inside their main cloud bill on day one, slow-moving managed cloud catalogs will look increasingly stale. The only real limitation is compliance: for companies that legally cannot allow data to cross outside AWS fences, the wait for Bedrock updates continues.