Moonshot AI
PulseAugur coverage of Moonshot AI — every cluster mentioning Moonshot AI across labs, papers, and developer communities, ranked by signal.
- 2026-11-05 product_launch Moonshot AI launched Kimi Hosted Agents, a managed cloud service for AI agents. source
- 2026-10-20 funding Moonshot AI is reportedly targeting $2 billion in annual revenue. source
- 2026-10-17 product_launch Moonshot AI launched the world's first AI-native credit card in partnership with the Agricultural Bank of China and American Express. source
- 2026-10-06 funding Moonshot AI closed its final private funding round at a valuation of approximately $50 billion. source
- 2026-10-01 controversy OpenAI accused Moonshot AI of using adversarial distillation to extract internal AI reasoning for training its Kimi model. source
- 2026-09-11 funding Moonshot AI is targeting $2 billion in annual revenue by the end of the year. source
- 2026-09-11 funding Moonshot AI is targeting $2 billion in annual revenue by the end of the year. source
- 2026-09-11 funding Moonshot AI is targeting $2 billion in annual revenue by the end of the year. source
- 2026-09-10 funding Moonshot AI is exploring a dual IPO in Hong Kong and Shanghai to raise capital. source
- 2026-09-10 funding Moonshot AI is exploring dual IPOs in Hong Kong and Shanghai to boost capital. source
- 2026-09-03 funding Moonshot AI has confidentially filed for an initial public offering in Hong Kong. source
- 2026-09-03 funding Moonshot AI has confidentially filed for an initial public offering in Hong Kong. source
- 2026-09-01 partnership Moonshot AI is negotiating with Microsoft, Amazon, and Google for a revenue-sharing agreement for its Kimi K3 model. source
- 2026-08-26 partnership Moonshot AI is negotiating potential hosting deals with Microsoft, Amazon, and Google for its Kimi K3 model. source
- 2026-08-26 partnership Moonshot AI is negotiating potential hosting and revenue-sharing deals with Microsoft, Amazon, and Google for its Kimi K3 AI model. source
22 day(s) with sentiment data
Kimi K3 weights remain unavailable despite July 27th target
The July 19th cluster mentions Kimi K3 weights are not yet publicly available, contradicting the July 16th cluster's anticipation of an open-weight release by July 27th. This suggests a potential delay or issue in making the weights accessible to the public.
Moonshot AI to announce infrastructure expansion plans within 14 days
Given the surge in demand causing Kimi AI to pause new subscriptions, it's highly probable that Moonshot AI will soon announce concrete plans for expanding their computing capacity. This is a necessary step to handle the current user load and accommodate future growth.
Kimi K3's high serving costs may limit adoption despite performance
The July 16th cluster highlights Kimi K3's significantly higher pricing for input and output tokens ($3/million and $15/million respectively). Combined with the July 19th cluster noting DeepSeek V4 Pro as more cost-effective, this suggests that Kimi K3's advanced capabilities might be offset by its cost, potentially impacting its widespread adoption.
Moonshot AI to release Kimi K3 weights by July 27th, 2026
Moonshot AI has stated an open-weight release for Kimi K3 is anticipated by July 27, 2026. Given the recent surge in demand and the model's strong benchmark performance, the company is likely to meet this deadline to capitalize on the growing interest in open-weight MoE models.
Kimi K3's high pricing may limit its adoption outside of specialized use cases.
With input tokens at $3/million and output tokens at $15/million, Kimi K3 is significantly more expensive than other models. While its performance is strong, this high cost could restrict its widespread adoption, limiting its use to high-value, specialized applications where the cost is justified by performance gains.
What are Kimi K3.1 Pro's latest capabilities and performance?
Moonshot AI's Kimi K3.1 Pro model significantly advances, featuring a 1 million token context window and strong coding performance.
The new Kimi K3.1 Pro, a 2.8 trillion parameter Mixture-of-Experts model, boasts a 1 million token context window, enhancing its ability to handle complex tasks and large datasets. It excels in frontend coding, often rivaling or surpassing models like Claude Opus 5.5, and integrates native Unity and Unreal Engine tools for 3D design workflows. Moonshot AI also plans to release its weights as open-source.
Why is Moonshot AI restricting Kimi K3.1 Pro's cyber capabilities?
Moonshot AI intentionally restricted Kimi K3.1 Pro's offensive cyber capabilities due to export controls and safety concerns.
The company removed data related to cybersecurity training from the Kimi K3.1 model release, influenced by China's new export restrictions on AI technologies. This decision reflects increasing global concerns about AI-powered cybersecurity exploits and a commitment to responsible AI development, aiming to limit the proliferation of offensive AI capabilities. Vitalik Buterin has endorsed these capability limitations.
What user experience challenges is Kimi K3.1 Pro facing?
Kimi K3.1 Pro users have reported rapid quota depletion and rate limit issues, prompting Moonshot AI to apologize and restore quotas.
Following the launch of Kimi K3.1 Pro, early community reports indicated that the model's quota depleted very quickly, sometimes within ten minutes, likely due to default settings maximizing computational effort and changes in billing for cached inputs. Moonshot AI has since apologized for these excessive rate limits and unnecessary quota deductions, compensating users by restoring 25% of monthly quotas and reducing sub-agent calls.
Are model distillation accusations still impacting Moonshot AI?
Moonshot AI continues to face serious allegations from OpenAI regarding adversarial model distillation and intelligence extraction.
OpenAI has repeatedly accused Moonshot AI of employing adversarial distillation techniques to extract proprietary reasoning data from its models for training Kimi. Recent reports detail thwarted attempts to replicate OpenAI's internal AI capabilities, with some activity linked to Moonshot AI, specifically targeting the Astra model. These persistent claims highlight ongoing intellectual property challenges and competitive tensions within the advanced AI sector.
What is Moonshot AI's current financial standing and IPO timeline?
Moonshot AI has achieved a $50 billion valuation and is preparing for a significant Hong Kong IPO, targeting substantial annual revenue.
The company recently secured its final private funding round, reaching a valuation of approximately $50 billion. Plans are underway for an initial public offering (IPO) in Hong Kong, anticipated for the first quarter of next year. Moonshot AI is also reportedly targeting $2 billion in annual revenue, signaling aggressive growth ambitions despite intense market competition.
How is Moonshot AI expanding its global reach and infrastructure?
Moonshot AI is expanding its global inference capabilities, with AWS adding India-based support for its Kimi K3 model.
Amazon Bedrock now supports India-based inference for Kimi K3, allowing customers with strict data residency requirements to process AI workloads within India. This expansion, particularly in the Mumbai region, is already being leveraged by partners like Ultrahuman for sensitive health data, demonstrating cost reductions and performance improvements. This move signifies Moonshot AI's strategic push into international markets and cloud partnerships.
Recent developments
- — AWS adds India-based inference for Moonshot AI's Kimi K3 model.
- — Moonshot AI retires Kimi K2 models, pushing users to K3.1 Pro.
- — Moonshot AI launches Kimi K3.1 Pro with 3T parameters and 1M context.
- — Kimi.ai apologizes for K3.1 Pro rate limits, restores user quotas.
- — Moonshot AI to release open-weight Kimi K3.1 Pro with restricted cyber capabilities.
- — OpenAI thwarts model-stealing campaign, linking some activity to Moonshot AI.
Why these stories ranked
-
95
OpenAI's direct action against a model-stealing campaign linked to Moonshot AI underscores the severity of ongoing IP challenges and competitive tensions, impacting industry trust.
-
94
The official launch of Kimi K3.1 Pro with its massive context window and parameter count is a major product announcement, directly challenging frontier models and setting new performance benchmarks.
-
95
The $50 billion valuation and impending Hong Kong IPO signal strong investor confidence and a significant strategic financial move for Moonshot AI, highlighting its market position.
-
92
The decision to release Kimi K3.1 Pro with open weights but restricted cyber capabilities reflects a critical balance between open-source contribution and responsible AI development, influencing policy discussions.
-
89
AWS expanding Kimi K3 inference to India demonstrates Moonshot AI's growing global infrastructure and strategic partnerships, crucial for market penetration and data residency compliance.
-
88
Moonshot AI's apology and quota restoration for K3.1 Pro users highlight the practical challenges of scaling new models and the importance of responsive customer service in a competitive market.
Trajectory of Moonshot AI coverage
Trend
Coverage of Moonshot AI is accelerating significantly, primarily driven by the comprehensive launch of Kimi K3.1 Pro, its open-weight strategy, and competitive pricing. The ongoing intellectual property disputes with OpenAI, particularly regarding Astra model distillation, and user experience challenges with quota depletion also contribute to sustained, high-velocity media attention.
Compared to peers
Moonshot AI is aggressively challenging US leaders like OpenAI and Anthropic with Kimi K3.1 Pro's performance, cost-effectiveness, and open-weight approach. Its unique restrictions on cyber capabilities differentiate it, while it continues to face distinct scrutiny regarding alleged IP theft and navigating China's export controls compared to its peers. The AWS partnership also expands its global reach beyond typical Chinese competitors.
Topic mix
This cycle shows a strong shift towards 'product' (Kimi K3.1 Pro features, pricing, open-weight release) and 'policy/safety' (export controls, cyber restrictions). 'User experience' (quota issues), 'IP' concerns (Astra distillation), and 'infrastructure' (AWS India) also feature prominently, indicating a more diversified and mature coverage landscape.
Our take
We see Moonshot AI making a bold statement with the Kimi K3.1 Pro launch, particularly its open-weight strategy and competitive pricing, which intensifies pressure on frontier AI leaders. However, the company is simultaneously navigating complex challenges, from persistent IP allegations to the practicalities of export controls and managing user expectations around model quotas. Our read is that its strategic expansion into global infrastructure, like the AWS India partnership, signals a determined push for international relevance amidst these multifaceted pressures.
Frequently asked
- What are the key features and performance of Moonshot AI's new Kimi K3.1 Pro model?
- Moonshot AI's Kimi K3.1 Pro is a 2.8 trillion parameter model featuring a 1 million token context window, designed for complex tasks and large datasets. It demonstrates strong performance in frontend coding, often rivaling top models like Claude Opus 5.5, and integrates with Unity and Unreal Engine for 3D design. The company also plans to release its weights as open-source, making it a highly competitive offering in the frontier AI space.
- Why did Moonshot AI remove cyberattack data from the Kimi K3.1 Pro model?
- Moonshot AI intentionally removed cybersecurity training data from Kimi K3.1 Pro due to China's new export control regulations and growing global concerns about the misuse of AI for cyber exploits. This strategic decision aligns with efforts to promote responsible AI development and prevent the proliferation of offensive AI capabilities, a stance endorsed by figures like Vitalik Buterin.
- Are the accusations of intellectual property theft against Moonshot AI still ongoing?
- Yes, Moonshot AI continues to face accusations from OpenAI regarding adversarial distillation techniques. OpenAI alleges that Moonshot AI has attempted to extract proprietary reasoning capabilities from its models, specifically the Astra model, to train its own Kimi models. These persistent claims underscore ongoing intellectual property disputes and intense competition in the AI industry, with US agencies also expressing concerns about such activities.
- What user experience issues have been reported with Kimi K3.1 Pro?
- Users of Kimi K3.1 Pro initially reported rapid depletion of their computing quotas and encountered excessive rate limits. This was likely due to default settings that maximized computational effort and changes in billing for cached inputs. Moonshot AI has since addressed these concerns, apologizing for the issues, restoring 25% of monthly quotas for all subscribers, and reducing the number of sub-agent calls to improve user experience. The company is also retiring older Kimi K2 models, pushing users to K3.1 Pro with new cost structures.
Related
-
Alibaba Cloud wins China Patent Gold Award for AI computing power management
Alibaba Cloud has been awarded the prestigious China Patent Gold Award for its innovative technology in managing large-scale intelligent computing power. This patent addresses the critical challenges of dynamic resource…
-
Chinese AI firms target Silicon Valley talent amid fierce competition · 1 source tracked
Chinese AI companies are increasingly attempting to recruit top engineering talent from Silicon Valley, aiming to compete with established U.S. firms like OpenAI and Anthropic. Startups such as MiniMax and Moonshot AI a…
-
Moonshot AI's Kimi K4 reportedly to feature 5.4T parameters by Q2 2027
A Chinese report suggests Moonshot AI's upcoming Kimi K4 model will feature 5.4 trillion parameters and is slated for release in the second quarter of 2027. The report also mentions a Kimi K3.5 model with approximately …
-
Homelabber runs massive 2.8T-parameter AI model on CPU using Optane memory
A Reddit user has successfully run Moonshot AI's 2.8-trillion-parameter Kimi K3 AI model using only a CPU and second-generation Intel Optane Persistent Memory modules. This setup, which eschews GPUs, achieved a speed of…
-
Chinese AI models dominate open-weight spending, Kimi leads market
A report from Menlo Ventures indicates that Chinese open-source AI models have captured 89% of global enterprise spending on open-weight systems. Moonshot AI's Kimi model leads this market, accounting for 61% of the spe…
-
Asia's AI boom offers advantages over U.S., says Antler VC chief · 1 source tracked
Asia is emerging as a significant hub for AI development and investment, offering distinct advantages over traditional tech centers like the U.S. According to Magnus Grimeland, founder of Antler VC, factors such as abun…
-
Open-weight AI models face production hurdles amid hardware shortages and licensing complexities · 2 sources tracked
This week's AI news highlights the growing availability of open-weight models, with Mistral AI and Moonshot AI releasing significant models. However, the practical adoption of these models is increasingly constrained by…
-
Chinese AI startup Moonshot AI releases Kimi K3.1 Pro open weights
Moonshot AI, a Chinese startup, has released the open weights for its Kimi K3.1 Pro model, a 2.8 trillion parameter system. This model demonstrates advanced agentic coding capabilities, nearly matching Anthropic's Claud…
-
Moonshot AI serves Kimi K3.1 Pro model on TPUs and GPUs via Mastodon
Moonshot AI is utilizing its Kimi K3.1 Pro model, powered by llm-d, across a combination of Tensor Processing Units (TPUs) and graphics processing units (GPUs). This infrastructure setup is being served via the Mastodon…
-
Moonshot AI releases Kimi K3.1 Pro model weights for free commercial use
Moonshot AI has released the weights for its Kimi K3.1 Pro model, making them available for free commercial use under general terms. This move allows developers and businesses to integrate the advanced AI capabilities i…
-
Alibaba Cloud integrates Moonshot AI's Kimi K3.1 Pro on new SuperPoD
Alibaba Cloud has announced that its Lingjun Zhenwu M890 SuperPoD now offers day-zero support for Moonshot AI's Kimi K3.1 Pro model. This integration leverages vLLM and SGLang for efficient operation. The system also in…
-
Moonshot AI launches Kimi Hosted Agents cloud service
Moonshot AI has launched Kimi Hosted Agents, a managed cloud service designed to simplify the deployment and operation of AI agents. This service provides a 24/7 managed agent harness, complete with sandboxes, audit tra…
-
Huawei Ascend enables day-zero support for Moonshot AI's Kimi K3.1 Pro
Huawei's Ascend team has achieved day-zero inference support for Moonshot AI's Kimi K3.1 Pro, a 2.8-trillion-parameter model. This integration allows the Kimi K3.1 Pro model to run on Huawei's Atlas A3 and Ascend 950 Su…
-
Together AI and Moonshot AI discuss Kimi K3.1 Pro capabilities
Together AI and Moonshot AI are collaborating to discuss their new Kimi K3.1 Pro model. The conversation aims to provide insights into the capabilities and appropriate use cases for this model. It suggests that the clar…
-
Moonshot AI releases Kimi K3.1 Pro open weights with 1M context
Moonshot AI has released the full weights for its Kimi K3.1 Pro model, a 2.8 trillion parameter model with a 1 million token context window, on Hugging Face. This release allows developers to self-host the model, but re…
-
AWS adds India-based inference for Moonshot AI's Kimi K3 model
AWS has expanded its Amazon Bedrock service to include India-based inference for Moonshot AI's Kimi K3 model. This new capability allows customers with strict data residency requirements to process AI workloads within I…
-
Alibaba launches AI tokens as consumer category on Tmall for Singles' Day
Alibaba has introduced AI tokens as a distinct consumer category on its Tmall platform for the Singles' Day shopping event. Six AI developers, including Alibaba Cloud, Z.ai, Moonshot AI, and DeepSeek, are offering top-u…
-
New malware uses Claude AI to rank victims' stolen assets
A new strain of Windows malware, dubbed VOLTA-Autodoxer, has been discovered that leverages AI models to assess the value of stolen victim data. The malware first exfiltrates a wide range of personal information, includ…
-
Moonshot AI retires Kimi K2 models, pushing users to K3.1 Pro with new cost structures
Moonshot AI is phasing out its Kimi K2 series models, with new accounts no longer able to access them after October 31, 2026, and all K2 models scheduled for retirement by the end of December 2026. The successor model i…
-
Gemini 4 Carbon, GPT-6.1 Terra, and Mistral Large 4 lead AI model updates · 1 source tracked
The AI model landscape is rapidly evolving with new frontier releases and a closing gap in open-weight models. Google DeepMind has launched Gemini 4 Carbon, a new top-tier model excelling in coding and cybersecurity, in…