China’s obsession with creative artificial intelligence has sparked a wave of product announcements from startups and tech giants on an almost daily basis, but investors are warning that as cost and profit pressures mount, There is a possibility of a stir.
The buzz in China, sparked by the success of the first OpenAI’s Chat GPT Nearly a year ago, it has sparked what a senior Tencent executive described this month as “a war of a.” Hundred modelsAs it and competitors from Baidu to Alibaba to Huawei promote their offerings.
China now has at least 130 major language models (LLMs), accounting for 40% of the global total and just behind the United States’ 50% share, according to brokerage CLSA. Additionally, companies have also announced dozens of “industry-specific LLMs” that are tied to their core model.
However, investors and analysts say most had yet to find viable business models, were too similar and were now struggling with rising costs.
Tensions between Beijing and Washington have also affected the sector as US dollar funds. Invest lessIn early stage projects and Difficulties Getting AI chips made by the likes of Nvidia Start cutting.
“Only those with the strongest capabilities will survive,” said Esme Pau, head of China internet and digital asset research at Macquarie Group, who expects consolidation and price wars as they compete for customers.
Several leading companies have indicated that they will compete on price to gain market share, he added. Cloud Services As have those from Alibaba and Tencent.
“Over the next six to 12 months, low-cap LLMs will be phased out due to chip restrictions, high costs and intensifying competition,” Pau said.
Founder and Responsible
Opinions about which firms vary widely.
Yuan Hongwei, chair of Shenzhen-based venture capital Z&Y Capital, said he believes only two to three general-purpose LLMs will dominate the market.
That’s why his firm looks for experienced founders when deciding which startups to invest in.
Z&Y, whose past investments include drone maker DJI and autonomous driving startup Pony.ai, has finally decided to back Baichuan Intelligence to take on Meta Platform’s Llama 2. A five-month-old firm is building an open-source AI model.
Baichuan was started by Wang Xiaochuan, the founder of China’s No. 2 Internet search engine Sugo Inc., and became one of the first five companies to receive Beijing’s approval to release a public chatbot in late August. The company is on track to close a second round that will be valued at $1 billion, Wang said.
“We see an opportunity here,” Yuan said. “Wang himself is leading the project. His understanding of digital business, his success with Sogo and how he focuses on the entire industry, we think is our best bet.”
Behind the new Chinese are several other big-name business and tech executives. AI startupsSuch as former Google China chief Kai Fu Li and former vice president of SenseTime Yan Joji.
Others said China’s biggest tech giants Alibaba, Tencent and Baidu ultimately had the biggest headstart and deep pockets to succeed, thanks to their large user bases and extensive services. For example, they can easily offer creative AI services to their cloud customers as an additional plugin.
“Current tech giants have inherited an unfair advantage of most low-hanging fruit business scenarios from their established ecosystems,” said Tony Tung, managing director of Gobi Partners GBA.
Tung added that some investors regret investing prematurely in LLM firms at the height of the hype earlier this year and that many such startups are struggling to build strong business cases. And now looking to partner with tech giants to explore use cases or possibly sell them.
“Many pursue similar LLMs, looking for similar problems to solve with micro-innovations in data processing techniques or model architecture,” he said.
“At this particular moment, investors are definitely a little bit higher than earlier this year.”