The Biden administration has intensified its scrutiny of Chinese technology firms by adding Zhipu, a prominent Chinese start-up specialising in large language models (LLMs) for artificial intelligence, to the US entity list. This move occurred on Wednesday and is part of a broader strategy to reinforce national security measures concerning technology that may bolster Chinese military capabilities.
Zhipu, which hails from Beijing and operates under the direction of Tsinghua University professor Tang Jie, has developed technologies akin to those powering popular AI applications like OpenAI's ChatGPT. The company has engaged with local governmental bodies to improve administrative services, such as deploying chatbots to assist residents with inquiries regarding waste management and parking.
Despite the US government's allegations that Zhipu's activities are contributing to advances in military technology through AI research, the start-up has publicly refuted such claims, asserting that the US decision "lacked factual basis". The blacklisting not only bars Zhipu from acquiring most technology originating from the United States but also casts a shadow on the future of its international collaborations.
Commercial Secretary Gina Raimondo articulated the rationale behind this stringent action, stating, "These rules will further target and strengthen our controls to help ensure that [China] and others who seek to circumvent our laws and undermine US national security fail in their efforts." The implications of this designation extend to the commercial realm, with Zhipu's existing partnerships and operations now subject to the heightened restrictions.
Conversely, insights from an investor in Zhipu suggest that the impact on the company's operational capacity might be minimal, as they possess a robust portfolio of in-house technologies and primarily service the domestic market. This investor contended that the US government’s focus on Zhipu could, paradoxically, reinforce the start-up's standing within China, as state support may increase in light of the US sanctions.
Indeed, past examples indicate that Chinese state entities have stepped in to back firms facing US sanctions, with companies such as SenseTime and Huawei receiving significant funding. Zhipu’s current backers include the National Social Security Fund, Tencent, Alibaba Cloud, and various venture capital groups, having secured $400 million in its last funding round in December.
In addition to Zhipu, the US government has cast its net wider to encompass other Chinese enterprises. The Chinese chip designer Sophgo was also added to the entity list due to concerns regarding its potential collaboration with Huawei and other actions that could further China's indigenous chip production objectives.
These developments signal a shift in the landscape of AI and technology, where the United States continues to implement restrictions on critical components like advanced chips vital for AI training and deployment. The government has imposed new licensing requirements for the export of high-end AI chips, particularly those manufactured by Nvidia, aiming to limit their availability to Chinese firms.
As the situation evolves, the future trajectory of AI start-ups in China, including Zhipu, rests on the interplay between domestic backing and the implications of US technological sanctions.
Source: Noah Wire Services