What Is Moonshot AI?
Moonshot AI is a Chinese artificial intelligence company founded in 2023 by AI researcher Yang Zhilin.
The company is best known for its Kimi family of AI models and has positioned itself around large-scale, open-weight artificial intelligence.
Moonshot has received backing from major Chinese technology companies, including Alibaba, and has raised more than $2 billion in funding, according to Reuters.
What Is Kimi K3?
Kimi K3 is Moonshot AI's latest flagship model.
Moonshot describes Kimi K3 as a 2.8-trillion-parameter open model built for advanced reasoning, coding, and knowledge work.
The company says the model has native vision capabilities and supports a one-million-token context window.
Kimi K3 uses a mixture-of-experts architecture and incorporates Moonshot's Kimi Delta Attention and Attention Residuals technologies.
Why Is Cloud Infrastructure Important?
Running an extremely large AI model requires substantial computing resources.
Although Kimi K3 is an open-weight model, most businesses may not have the infrastructure needed to operate a model of this size independently.
Cloud platforms can provide the computing capacity required to serve AI models to developers and businesses without requiring customers to build their own large-scale infrastructure.
This makes cloud providers an important part of the commercial AI ecosystem.
What Are Microsoft, Amazon and Google Reportedly Discussing?
According to Reuters, Moonshot AI is negotiating potential revenue-sharing arrangements with Microsoft, Amazon and Google.
The reported structure would allow the U.S. cloud companies to host Kimi K3 while sharing revenue generated from related services with Moonshot.
Reuters reported that Moonshot is seeking as much as a 30% share of revenue generated from K3-related services.
The exact terms remain unresolved, and the talks could still change or end without an agreement.
Why Could These Talks Be Significant?
A deal involving major U.S. cloud providers would give Kimi K3 access to some of the world's largest cloud ecosystems.
It could also provide Moonshot with a wider route to enterprise customers outside China.
For Microsoft, Amazon and Google, hosting an advanced open-weight Chinese AI model could create a commercial opportunity while also raising questions around regulation, data access and U.S.-China technology restrictions.
What Issues Still Need to Be Resolved?
Several issues reportedly remain under discussion.
These include how revenue would be divided, how customer and model data would be handled, and how token usage would be measured.
Token usage is particularly important for AI services because many providers use the number of tokens processed to calculate usage and pricing.
A reliable auditing system would therefore be needed to determine how much revenue is generated by Kimi K3 services.
Kimi K3 Is Already Available Through Cloud Infrastructure
Moonshot AI has already partnered with Together AI to make Kimi models available through U.S.-hosted infrastructure.
Together AI announced in July that Kimi K3 was live on its platform and that future open Moonshot models would also be supported.
The partnership shows that cloud distribution is already part of Moonshot's strategy for making Kimi models accessible to developers.
Why Does Kimi K3 Need Large-Scale Computing?
Kimi K3 contains 2.8 trillion parameters, making it one of the largest open-weight AI models.
Moonshot designed the model using a sparse mixture-of-experts approach, which allows only a portion of the model's experts to be activated for a given task.
Even with this efficiency, operating such a large model can require significant computing resources.
Cloud infrastructure can make deployment easier for organizations that do not have access to large AI computing clusters.
Moonshot says Kimi K3 was designed for long-horizon coding, knowledge work, and advanced reasoning.
The model also supports text, image, and video inputs through its native multimodal capabilities.
Third-party evaluations have also attracted attention. Reuters reported that Arena.ai ranked Kimi K3 highly in a benchmark focused on web-interface building, while Artificial Analysis reported competitive performance on complex, multi-step tasks.
These evaluations are independent assessments and should not be treated as guarantees of performance for every use case.
What Are the U.S.-China Concerns?
The potential cloud discussions come at a sensitive time for U.S.-China technology relations.
Washington has imposed restrictions on the export of advanced AI chips to China because of national-security concerns.
Moonshot has also faced criticism from U.S. officials. Reuters reported that U.S. officials have accused the company of intellectual-property violations and illegally obtaining Nvidia chips.
Moonshot has rejected allegations that Kimi K3's performance was achieved through improper model distillation.
What Happens Next?
The next step will depend on whether Moonshot AI and the U.S. cloud companies can resolve the commercial and technical issues involved.
At present, the discussions remain preliminary.
If agreements are eventually reached, Kimi K3 could become more widely available to businesses and developers through major U.S. cloud platforms.
If no agreement is reached, Moonshot can continue using other distribution and infrastructure partners.
TwikUp Insight
The reported Moonshot AI discussions highlight how cloud infrastructure is becoming an important part of the global AI race.
Kimi K3's large scale makes cloud deployment particularly relevant, while its open-weight approach gives developers more flexibility than traditional closed AI systems.
However, the reported talks with Microsoft, Amazon, and Google should not be presented as confirmed deals. The companies have not announced finalized agreements, and the discussions could still change.
For now, the development shows growing international interest in China's AI models despite the wider technology and national-security tensions between Washington and Beijing.
