Skip to main content
Back to News
Partnership
2 min read
CN

Moonshot AI's Kimi K3 Goes Live on Amazon Bedrock, Expanding Enterprise Reach

The AMW Read

Kimi K3's Bedrock GA is a hyperscaler-distribution move for an already-tracked CN frontier lab, technically notable (largest open-weight model, first Bedrock prompt-caching) but not debate-resolving.
NoveltySignificance
Foundation Models · Player MapFoundation Models · Recurring Patterns
Moonshot AI
Moonshot AI

Foundation Models / LLMs

View Company Profile

Moonshot AI's Kimi K3 Goes Live on Amazon Bedrock, Expanding Enterprise Reach

Moonshot AI's Kimi K3 — a 2.8-trillion-parameter open-weight model with native vision and a 1-million-token context window — is now generally available on Amazon Bedrock. Moonshot reports roughly 2.5x scaling efficiency over its predecessor Kimi K2, and Bedrock says K3 is the first open-weight model on the platform to support explicit prompt caching, which cuts latency and input cost on repeated long-context calls. It ships in a US geographic profile and a global cross-region profile priced about 10% lower, both running inside Bedrock's standard security boundary: data stays within AWS, with zero retention and zero model-provider access by default.

Putting K3 inside AWS's own security perimeter — the same access, encryption, and audit controls Bedrock applies to proprietary models — lowers the trust barrier that has kept many Western enterprises from calling a China-hosted API directly. The listing lands amid an unusually dense run of Moonshot coverage: AMW's pipeline logged 44 Moonshot AI stories in the trailing 90 days versus 29 in the prior period (per the AI Market Watch index; name-matched, pipeline-ingested coverage only, not a census), spanning a confidential Hong Kong IPO filing targeting a reported $50B valuation and Anthropic's public allegation that Moonshot distilled Claude Opus via routed requests. A hyperscaler listing is a way to monetize an open-weight model in Western markets without asking enterprises to trust a China-hosted endpoint directly — AWS hosts the compliance surface instead.

For builders, the 1M-token context plus prompt caching makes long-repository and multi-document agent work cheaper to run without leaving AWS's compliance boundary. For investors, the Bedrock listing is a distribution proof point ahead of the IPO, at a moment when Moonshot's own $2B annualized-revenue target is still being tested against its valuation ask.

#MoonshotAI #KimiK3 #AmazonBedrock #AWS #OpenWeightModels #ChineseAI

#Moonshot AI#Kimi K3#Amazon Bedrock#AWS#open-weight LLM#related:Amazon Bedrock

How This Connects

Based on Foundation Models · Player Map

  1. 1d agoAnthropic opens Nasdaq IPO roadshow with a reported $70–$78 share price bandAnthropic
  2. 1w agoCohere completes initial close of Series E targeting up to CAD 7 billionCohere
  3. 1w agoCohere secures up to C$1.6 billion from Canada with conditions on jobs and headquartersCohere
  4. 2w agoAnthropic's Public IPO Filing Puts Growth and Compute Costs Under ScrutinyAnthropic
  5. 1mo agoTypeSafe AI Launches Jev System One, a Decision Model That Returns Choices Instead of TextTypeSafe AI
  6. 1mo agoMoonshot AI's Kimi K3 Goes Live on Amazon Bedrock, Expanding Enterprise Reach · THIS ARTICLE

Related News

More news from Moonshot AI

Stay updated with the latest news and announcements from Moonshot AI.

View all Moonshot AI news

Discover AI Startups

Explore 5,000+ AI companies with VC-grade analysis, funding data, and investment insights.

Explore Dashboard