Chinese artificial intelligence developer Zhipu is making headlines in the tech investment world, having secured an impressive 2.5 billion yuan (approximately $342 million) in funding since the beginning of the year. This achievement was officially announced today, with heavyweight tech giants Alibaba Group Holding Ltd. and Tencent Holdings Ltd., along with Xiaomi Corp. and other notable contributors, participating in the funding round.
Zhipu, which reportedly originated from China’s prestigious Tsinghua University in 2019, is led by Professor Tang Jie, an esteemed figure in the university’s computer science department. The company’s primary focus centers on the development of large language models akin to GPT-4, the powerful artificial intelligence behind OpenAI LP’s renowned ChatGPT chatbot.
Notably, Zhipu has introduced its own chatbot, known as Qingyan, into the market. Additionally, the company has actively contributed to the AI community by creating two open-source large language models: GLM-130B and ChatGLM-6B, boasting a remarkable 130 billion and 6 billion parameters, respectively.
In the realm of artificial intelligence, parameters serve as the essential configuration settings that dictate how AI processes data. Developers manually define certain settings, such as the number of artificial neurons integrated into a language model, during the coding phase of AI projects. Other parameters are autonomously determined by the neural network itself during the training phase.
AI models with an extensive number of parameters typically exhibit a greater capacity for tackling advanced tasks compared to their smaller counterparts. However, neural networks with fewer parameters offer the distinct advantage of being significantly more hardware-efficient. Notably, Zhipu’s ChatGLM-6B, the smaller of the two open-source language models, is designed to operate efficiently on consumer-grade graphics cards.
The timing of Zhipu’s fundraising announcement coincides with another significant AI development in the Chinese tech industry. Recently, Baichuan, another emerging AI startup, unveiled a remarkable $300 million investment. This substantial funding included contributions from several companies that also supported Zhipu, such as Alibaba, Tencent, and Xiaomi. Baichuan’s valuation now reportedly exceeds $1 billion.
Founded in April of the same year, Baichuan operates in direct competition with Zhipu within the burgeoning large language model market. The company has already developed four open-source language models and two proprietary models. Baichuan distinguishes itself by providing neural networks through application programming interfaces (APIs), enabling developers to construct tools like customer support chatbots.
The substantial investments in both Zhipu and Baichuan reflect the burgeoning interest in AI startups, particularly those specializing in large language models, during the early stages of this year. Amazon.com Inc., for instance, recently declared its intention to invest up to $4 billion in Anthropic, a rival to OpenAI, renowned for its AI chatbot Claude.
Furthermore, AI enthusiasts are closely following OpenAI’s journey, with reports suggesting that the company is considering selling shares. In recent news, sources revealed that OpenAI may conduct a secondary sale, potentially valuing the AI developer at around $86 billion. An earlier report in The Wall Street Journal even proposed a valuation as high as $90 billion, signaling the tremendous financial interest in the AI sector.
This year has already been marked by significant investments and growth in the AI industry, and Zhipu’s accomplishment is emblematic of the sector’s dynamism and potential. The competition and collaborations among emerging AI startups promise to further accelerate technological advancements and innovation in the field. With such substantial funding and the backing of tech giants, Zhipu appears poised to make a lasting impact in the world of artificial intelligence.