{"slug": "google-helps-chinese-firm-accused-by-white-house-of-copying-us-ai-models-bring", "title": "Google Helps Chinese Firm Accused by White House of Copying US AI Models Bring Kimi K3 to US Cloud", "summary": "Google Cloud has begun offering 'Day 0' support for Moonshot AI's Kimi K3 model on its US cloud platform, days after White House officials accused the Chinese firm of copying US AI technology through large-scale distillation of Anthropic's models. The White House Office of Science and Technology Policy director Michael Kratsios alleged Moonshot AI used covert distillation to develop Kimi K3 and obtained restricted Nvidia GB300 processors via third countries, while Treasury Secretary Scott Bessent warned of potential sanctions. Google's move puts the cloud provider alongside a company under congressional investigation for national security risks.", "body_md": "# Google Helps Chinese Firm Accused by White House of Copying US AI Models Bring Kimi K3 to US Cloud\n\n## The rollout came days after White House officials accused Moonshot AI of copying US AI technology.\n\nGoogle has begun helping customers run a leading Chinese artificial intelligence system on its US cloud platform, only days after senior White House officials accused the model's developer of copying American AI technology and warned Chinese firms could face sanctions over alleged intellectual property and export‑control breaches.\n\nThe move puts one of the world's biggest cloud providers alongside Moonshot AI, the Beijing‑based company behind the open‑weight [Kimi K3](https://www.ibtimes.co.uk/moonshot-ai-unveils-kimi-k3-revolutionary-open-source-ai-model-1809160) model, at the same time as Washington steps up scrutiny of Chinese AI developers and their access to advanced US hardware.\n\n## Google Announces 'Day 0' Support for [Kimi K3](https://www.ibtimes.co.uk/kimi-k3-ai-discovers-redis-vulnerabilities-1810344) on Its Cloud\n\nGoogle Cloud announced 'Day 0' support for Kimi K3 through its Model Garden catalogue, allowing enterprise customers to deploy the model on Google's infrastructure using Nvidia Blackwell‑based A4 and A4X virtual machines.\n\nThe company also published validated deployment recipes and configuration guidance intended to simplify large‑scale deployment.\n\nGoogle thanked Moonshot AI and infrastructure partner RadixArk for their 'close collaboration and partnership to deliver this great outcome for the community'.\n\nThe announcement means organisations can self‑host Kimi K3 on Google Cloud rather than relying only on Moonshot's own services.\n\nKimi K3 is one of the largest open‑weight AI models released so far. Moonshot says the mixture‑of‑experts model contains 2.8 trillion parameters, activates roughly 104 billion parameters per token and supports a one‑million‑token context window.\n\n## White House Has Publicly Accused Moonshot AI\n\nGoogle's announcement came only days after Michael Kratsios, director of the White House Office of Science and Technology Policy, publicly accused Moonshot AI of using large‑scale distillation to build Kimi K3 from Anthropic's models.\n\n'We have information that Moonshot AI distilled Anthropic's Fable for the development of its K3 model,' Kratsios said.\n\nHe added that 'large-scale, covert industrial distillation aimed at stealing proprietary US technology and undermining American research is unacceptable.'\n\nKratsios also alleged that Moonshot obtained access to restricted Nvidia GB300 processors through third countries despite US export controls.\n\nTreasury Secretary Scott Bessent separately warned that Chinese companies found to have engaged in such conduct could face sanctions or placement on the Entity List, saying, 'Open source is not open season on American IP.'\n\nMoonshot AI has not publicly accepted those allegations.\n\n## Congress Is Already Investigating Chinese AI Models\n\nMoonshot AI was already under scrutiny before Google announced support for Kimi K3.\n\nIn April, the House Select Committee on the Chinese Communist Party, together with the House Homeland Security Committee, launched a joint investigation into potential national security risks posed by several Chinese AI developers, including Moonshot AI, DeepSeek, Alibaba and MiniMax.\n\nLawmakers said they were examining concerns involving data security, censorship aligned with Chinese government policies and the role advanced AI models could play in strengthening Beijing's technological capabilities.\n\nThe investigation remains ongoing.\n\n## Kimi K3 Licence Imposes Conditions on Big Cloud Firms\n\nAlthough Moonshot describes Kimi K3 as an open‑weight model, its licence sets commercial restrictions for large cloud providers.\n\nThe licence allows developers to download, modify and deploy the model without charge in many circumstances.\n\nHowever, companies operating a 'Model as a Service' business with more than $20 million in annual revenue must enter into a separate commercial agreement with Moonshot before offering hosted inference services based on Kimi K3.\n\nIndustry analysts have interpreted the clause as requiring commercial licensing arrangements for qualifying cloud providers that offer managed AI services.\n\nAfter questions were raised following Google's announcement, a Google source clarified that the company was providing infrastructure and deployment guidance for customers wishing to run Kimi K3 on Google Cloud themselves, rather than announcing a fully managed Google‑hosted Kimi service.\n\n## US–China AI Competition Shifts To Cloud and Distribution\n\nThe announcement shows how competition between US and Chinese AI companies is expanding beyond model development into cloud infrastructure, deployment tools and enterprise distribution.\n\nGoogle's support gives enterprise customers access to one of China's most capable open‑weight models on American cloud infrastructure powered by [Nvidia hardware](https://www.ibtimes.co.uk/nvidia-ai-financing-deals-market-concerns-1811093), while US officials continue to question how Moonshot developed Kimi K3 and whether it complied with US intellectual property and export‑control rules.\n\nNeither Google nor Moonshot indicated that the arrangement changes the ongoing investigations or allegations surrounding the Chinese developer.\n\nThe White House accusations have not resulted in formal findings of wrongdoing against Moonshot, and no US agency has announced enforcement action linked to Google's cloud support for Kimi K3.\n\n© Copyright IBTimes 2025. All rights reserved.", "url": "https://wpnews.pro/news/google-helps-chinese-firm-accused-by-white-house-of-copying-us-ai-models-bring", "canonical_source": "https://www.ibtimes.co.uk/google-cloud-supports-chinese-ai-model-kimi-k3-1811403", "published_at": "2026-07-29 21:26:49+00:00", "updated_at": "2026-07-29 21:37:24.940953+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-policy", "ai-products", "ai-infrastructure", "ai-ethics"], "entities": ["Google Cloud", "Moonshot AI", "Kimi K3", "White House", "Michael Kratsios", "Anthropic", "Nvidia", "Scott Bessent"], "alternates": {"html": "https://wpnews.pro/news/google-helps-chinese-firm-accused-by-white-house-of-copying-us-ai-models-bring", "markdown": "https://wpnews.pro/news/google-helps-chinese-firm-accused-by-white-house-of-copying-us-ai-models-bring.md", "text": "https://wpnews.pro/news/google-helps-chinese-firm-accused-by-white-house-of-copying-us-ai-models-bring.txt", "jsonld": "https://wpnews.pro/news/google-helps-chinese-firm-accused-by-white-house-of-copying-us-ai-models-bring.jsonld"}}