From the AWS edition of November 10, 2026
AWS details Moonshot AI's Kimi K3.1 Pro Highspeed on Amazon Bedrock, with data-boundary claims and US-only routing option
Amazon Web Services / Moonshot AI · Cross-SLED AI platforms / model procurement · Global commercial AWS Regions (US and India geographic routing options)
- Publisher
- AWS Artificial Intelligence Blog
- Original publication
- 2026-11-10
- Source retrieved
- 2026-11-11
- Event date
- 2026-11-06
What happened
AWS's November 10, 2026 launch post says Kimi K3.1 Pro Highspeed, an open-weight model from China-based Moonshot AI with a 1-million-token context window, is on Amazon Bedrock (AWS's What's New entry is dated November 6). AWS states customer data is processed within the AWS data boundary, is not shared with Moonshot AI, is not used for training, and that zero data retention is always on. Access is through cross-Region inference: a global profile (cheapest, any commercial Region), a US profile and an India profile. A Flex tier is offered at half the Standard price. Performance claims are attributed to Moonshot AI; no benchmarks or GovCloud availability are given.
Why it matters
Bedrock is a common route for state agencies and universities to reach many models through one contract. A Chinese-origin model now appears in that catalog, which matters because several states restrict PRC-affiliated AI vendors (see the Texas item in this edition). SLED cloud teams need to decide whether hosting by AWS changes how those rules apply.
Evidence and measured results
Vendor blog. Data-handling statements are AWS claims; capability claims are attributed to Moonshot AI. No independent evaluation, customer example or public-sector use is described.
Limitations and uncertainty
Vendor claim; no benchmarks. Interpretation of state vendor bans as applied to hosted models is ours, not the source's. GovCloud availability is not stated in the source.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-11-11; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- A Bedrock catalog now includes a model from a vendor named on at least one state prohibited-technology list, and users may select it for cost.
- Stakeholders
- State or university CISO
- Procurement and legal counsel
- Cloud center of excellence lead
- Discovery questions
- Does your state policy name Moonshot AI or PRC-affiliated AI vendors?
- Does that policy reach models hosted by a US cloud provider?
- Which Bedrock model IDs are allowed today, and how is that enforced?
- Value hypothesis
- A documented allow-list decision reduces compliance risk from new catalog entries; it does not by itself prove compliance.
- Potential engagement
- A short model-governance review that maps state vendor rules to Bedrock model IDs and SCPs.
- Unsupported claims
- That AWS hosting satisfies state vendor bans
- Moonshot's performance claims
- Cost savings from Flex or global routing for any specific workload
Pre-sales engineering
Role takeaway
- Fit
- Relevant to any SLED Bedrock tenant in commercial Regions; GovCloud availability not stated.
- Architecture and integration
- Bedrock runtime with OpenAI-compatible and Converse APIs; global, US and India cross-Region inference profiles.
- Prerequisites
- Organization-level SCPs
- Approved model list
- CloudTrail logging of Bedrock invocations
- Constraints
- Global profile routes to any commercial Region
- GovCloud availability not stated
- No benchmarks published by AWS
- Security
- Restrict model IDs and inference profiles by policy; AWS claims zero retention and no operator access, which should be confirmed in service terms.
- Proof of value
- Verify an SCP denies the model in a test account
- Confirm us. profile keeps processing in US Regions per documentation
- Check CloudTrail records denied calls
Delivery
Role takeaway
- Work
- Add the model to the agency's Bedrock model-intake review and enforce the decision.
- Dependencies
- Legal interpretation of state vendor rules
- SCP management access
- Model inventory
- Ownership
- Cloud governance team; CISO approves exceptions.
- Skills and adoption
- Cloud teams need IAM/SCP skills for model-level controls.
- Governance checkpoints
- Legal review of vendor restriction
- Allow-list update
- Exception approval record
- Acceptance criteria
- Proposed: unapproved model IDs denied in all accounts
- Proposed: each new Bedrock model reviewed within an agreed period
- Proposed: zero unapproved invocations in CloudTrail
- Risks
- Shadow use via cheaper profiles
- Policy ambiguity on hosted models
- Allow-list drift as catalog grows
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Model is reached through bedrock-runtime with OpenAI-compatible Chat Completions/Responses and Bedrock Converse/Invoke APIs. The default recommended global profile can route to any commercial Region; the us. profile keeps processing in US Regions.
Governance
Who approves, reviews and stays accountable for outcomes?
State policies that name Moonshot AI as a prohibited vendor may or may not reach a model hosted on AWS infrastructure; the source does not address this. Agencies should decide explicitly and enforce through model allow-lists.
Security and privacy
What data, permissions and controls need testing?
AWS claims zero data retention and no operator access during inference. Model weights originate from a vendor on at least one state prohibited list; supply-chain and output-integrity questions are not addressed in the post.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
No accessibility or workforce detail.
Procurement
What should contracts, pricing and exit terms secure?
Lower per-token cost via the global profile and Flex tier may attract budget-constrained users, which raises the importance of allow-list controls. GovCloud availability and authorization status not stated.
Operating model
Which teams own the service once it runs?
Cloud governance teams need a standing process for reviewing each newly added Bedrock model against state vendor restrictions.
Publication history
- 2026-11-10AWS · Issue 128 resources
Stable resource ID: aws-bedrock-kimi-k3-1-pro-moonshot-open-weight-20261110