Generative AI, which translates text into prose, graphics, and videos, is developing in a fragmented tech landscape between the U.S. and China due to the decoupling.
China is developing huge language models like OpenAI’s GPT models to minimize its reliance on the U.S. Unlike the U.S., Baidu is one of its most sophisticated AI projects.
Google launched its ChatGPT competitor in March. The 23-year-old corporation wants to invest in other AI firms. It wants to own AI startups. Baidu co-founder and CEO Robin Li unveiled a billion yuan ($145 million) fund for generative AI businesses at a JPMorgan symposium in China this week.
As my colleague Connie highlighted, the OpenAI Startup Fund started at $100 million and grew to $175 million. The fund will invest 10 million yuan, or $1.4 million, in a project. Given the check size, the fund is clearly targeting early-stage AI applications, which isn’t surprising since Chinese generative AI businesses haven’t seen widespread adoption and most investments are in the seed and early stages.
Baidu plans to utilize the investment to promote its huge language model Ernie Bot. American developers are creating ChatGPT-based apps. Li predicted more Chinese developers will use Ernie to build AI apps.
Thus, the fund appears to favor AI applications above core layer developers. Pitches will flood the fund. Chinese startups have innovated in live streaming, live commerce, and short films. Li projected that Chinese enterprises will again pioneer commercial AI applications in the generative AI age.
I’m optimistic about China’s AI. China has welcomed new innovations for decades. We created breakthrough apps like WeChat, Douyin, and Didi while not inventing Android, iOS, or Windows. Many are popular and useful. AI follows the same pattern. We’re skilled at capturing technology’s many possibilities to build applications.”
However, the basic level—China’s domestic huge language models—may not be powerful enough to accommodate the diversity of real-life scenarios anticipated of them. China needs its own LLMs to avoid U.S. sanctions that shut off technology supplies, like in the semiconductor industry. Alibaba and Tencent, like Baidu, are building massive language models.