Realtime AI News
Kimi K3 lands on Amazon Bedrock and Alibaba Cloud Bailian as Moonshot's revenue-share model takes effect
Guandian reported on September 21 that Moonshot AI's Kimi K3 is now available on Amazon Bedrock and Alibaba Cloud's Bailian platform, with Amazon publicly confirming the integration. The launch also marks the formal start of Moonshot's revenue-sharing arrangement with cloud providers, changing how the model is distributed and monetised.

Guandian reported on September 21 that Moonshot AI's Kimi K3 has landed on Amazon Bedrock and Alibaba Cloud's Bailian platform, with Amazon publicly announcing the integration, and that Moonshot's revenue-sharing arrangement with cloud providers has formally taken effect. It is a concrete step forward on distribution and business model after the model's capability release.
The two integrations cover one overseas platform and one domestic one. Amazon Bedrock is AWS's managed service for hosting and serving models, while Bailian is Alibaba Cloud's large-model platform. Kimi K3 appearing on both means its call entry point now extends beyond Moonshot's own API.
The more interesting part is the phrase revenue sharing. According to the report, the arrangement between Moonshot and cloud providers is now in place. In practice this usually means the model maker takes an agreed share of the usage revenue generated on the cloud platform, rather than simply licensing the model or selling it only through its own API pricing. For a company whose core asset is model capability, that hands distribution and billing to the cloud provider and leaves Moonshot to focus on iteration.
From an industry perspective, this rearranges the relationship between model makers and cloud platforms. Cloud platforms have largely been compute and hosting providers while model makers sold usage separately. Once revenue sharing is in place, both sides are tied to the same usage curve: the cloud provider has an incentive to recommend the model to its own customers, while the model maker gains distribution without building a sales organisation.
The report does not disclose the revenue split, the billing basis or the geographic scope of availability, so those details still need to be confirmed against official announcements. Whether the model's version and pricing are identical across the two platforms is also not yet public.
The next question is whether the model can be copied elsewhere. If revenue sharing works on Amazon Bedrock and Alibaba Cloud Bailian, whether other cloud providers follow and whether other model makers adopt similar terms will decide if this is a one-off commercial deal or a standard route for Chinese models going overseas.
Why it matters
For Moonshot this is an overseas distribution channel it does not have to build itself; for the cloud platforms it is a third-party model they now have a financial reason to promote. The arrangement is likely to become a template other Chinese model makers reference.
Nearby Updates
All09/21, 14:06
Amazon blocks Meta's Muse AI agent from shopping on its site as agentic commerce fight escalates
Amazon has blocked Meta's new Muse personal AI agent from shopping on Amazon.com, citing security, privacy and transparency concerns, after failing to persuade Meta to exclude the site from the experience. Muse users now see a popup saying continued access by an unauthorized AI agent violates Amazon's Conditions of Use, and Amazon says it never authorized the agent.
09/21, 13:53
MiniMax Open-Sources MiniMax Code, a Terminal Coding Agent
According to TechNode, MiniMax has open-sourced MiniMax Code, a terminal coding agent that developers can pick up and run directly from the command line. The move signals that MiniMax wants to compete for coding-agent users through open source rather than a closed product, though the repository, licence and support scope still need official confirmation.
09/21, 12:50
Qwen open-sources Qwen-Image-2.1, a 7B visual generator that unifies text-to-image and editing
Qwen released Qwen-Image-2.1 on Hugging Face on September 21, a unified text-to-image generation and image-editing model whose visual generation component carries just 7B parameters across 32 single-stream DiT layers. The release adds native RGBA transparency, support for up to 10 reference images and mask-based local edits, and ships a QwenImage21Pipeline integration in diffusers.
09/21, 11:02
China's AI Models Lead Weekly API Calls for a 21st Straight Week as DeepSeek V4.1-Flash Tops the Chart, Up 219%
A weekly tally shows Chinese AI models leading on call volume for a 21st consecutive week, with DeepSeek V4.1-Flash taking the top spot and growing 219% week over week. The figures were compiled by Futu NiuNiu and surfaced through an aggregator feed on September 21.