Chinese AI startup 01.AI to raise $200 million to build the next generation of open-source AI models

Beijing-based 01.AI, a Chinese artificial intelligence startup founded by former Google China chief Lee Kai-fu, is raising up to $200 million in fresh funding, according to an exclusive report from Reuters. The move to raise funds comes amid a surge in global interest in open-source AI models.

According to an insider cited by Reuters, 01.AI, which officially launched in July following a three-month incubation period, achieved a valuation of $1 billion in early November. Another source mentioned that the company is actively seeking additional investments from U.S. dollar-based investors. All the sources contributing to the report requested anonymity because the fundraising details were not yet public, Reuters reported.

Recognized for its Yi-34B model, which became the first Chinese Language Model (LLM) to lead the Huggingface leaderboard in November, 01.AI is part of the wave of Chinese startups opening or planning to open their large language models (LLMs) for public use. This move puts them in competition with larger entities like Meta and Alibaba, all striving to attract users and keep pace with market leader OpenAI.

While the benefits of open-source AI are apparent, concerns persist about potential misuse by malicious actors. The success of 01.AI has the potential to bolster China’s standing in the global AI landscape. However, the company faces tough competition from other Chinese startups like Zhipu AI and Baichuan Technology, both of which have also attracted substantial investments.

In November, 01.AI gained recognition in the open-source LLM community when its Yi-34B model claimed the top spot on the Huggingface leaderboard—a platform where tech firms share and rate LLMs based on performance and popularity.

The rising interest in generative artificial intelligence (AI) has proven beneficial for startups, leading to significant investments from backers eager to support the promise of open AI models. Notably, Zhipu AI, founded in 2019, surpassed a $1 billion valuation, and Baichuan Technology achieved a valuation of $1.2 billion in its last fundraising round in October, according to a source with direct knowledge of the exercise.

Meanwhile, ChatGPT maker OpenAI has voiced criticism against the movement to open-source Large Language Models (LLMs). Unlike some other initiatives, OpenAI has chosen to keep the codes of its models tightly guarded, citing concerns about potential abuse by malicious actors and the resulting societal risks.

In contrast, Alibaba’s cloud arm in China has actively embraced the open-source approach for its LLMs. The latest addition to its lineup, Qwen-72B, recently claimed the top spot on Huggingface’s leaderboard, marking the second Chinese model to achieve this feat.

Qwen-72B stands as the most recent and powerful among the eight AI models Alibaba has open-sourced in the past four months. According to the company, Qwen-72B can surpass OpenAI’s flagship GPT-4 in handling Chinese, as indicated by certain industry benchmarks.

Jingren Zhou, the Chief Technology Officer of Alibaba Cloud, stressed the importance of establishing an open-source ecosystem to foster the development of LLMs and AI applications. In an interview with Reuters on Wednesday, Zhou stated that Alibaba Cloud is committed to being the most open cloud and aims to make generative AI capabilities accessible to a broader audience. He highlighted the increasing demand for AI applications across various industries, developers, and businesses.

When asked about Alibaba’s commitment to open-sourcing its AI models, Jingren Zhou, the chief technology officer of Alibaba Cloud, told Reuters: “Building up an open-source ecosystem is critical to promoting the development of LLM and AI applications building.”