TL;DR:
- China’s AI sector is experiencing a surge in generative artificial intelligence, driven by OpenAI’s ChatGPT success.
- The market sees a proliferation of large language models (LLMs), with over 130 models now accounting for 40% of the global total.
- Challenges emerge as LLMs struggle to find viable business models and grapple with rising costs.
- Geopolitical tensions, particularly between Beijing and Washington, add complexity, impacting funding and chip supply.
- Experts anticipate consolidation and a fierce price war among LLM providers, with lower-capacity models likely to exit the market.
- A few general-purpose LLMs are expected to dominate, and experienced founders become key investment criteria.
- Notable entrepreneurs and tech executives were backing new Chinese AI startups, while established tech giants like Alibaba, Tencent, and Baidu leveraged their user bases for AI services.
- Investors are cautious after a hype-driven rush into LLMs, with many startups seeking partnerships or acquisitions by tech giants.
Main AI News:
In the wake of OpenAI’s groundbreaking success with ChatGPT a year ago, China has been caught in the throes of a generative artificial intelligence frenzy. This exhilarating wave has brought forth a relentless stream of product unveilings from both nimble startups and tech titans. Nevertheless, astute investors are sounding a cautionary note, predicting an impending shakeout as cost and profitability pressures mount.
This vibrant AI landscape in China, initially sparked by OpenAI’s ChatGPT, has culminated in what a high-ranking Tencent executive recently referred to as the “war of a hundred models.” The likes of Baidu, Alibaba, and Huawei have been fervently endorsing their respective offerings, flooding the market with a multitude of large language models (LLMs). China now boasts at least 130 such LLMs, constituting 40% of the global total and second only to the United States, which commands a 50% share, according to CLSA, a brokerage firm. Furthermore, numerous companies have unveiled a plethora of “industry-specific LLMs” tailored to complement their core models.
Despite this fervor, investors and industry analysts have observed that most of these AI ventures are yet to devise sustainable business models. Many appear strikingly similar to one another and are grappling with escalating operational costs. The ongoing tensions between Beijing and Washington have further complicated matters, with U.S. dollar funds showing less enthusiasm for early-stage AI projects and the difficulties in securing AI chips from manufacturers like Nvidia starting to take a toll.
In the midst of these challenges, Esme Pau, the head of China’s internet and digital asset research at Macquarie Group, asserts, “Only those with the strongest capabilities will survive.” Pau foresees an impending consolidation phase and an intense price war as players vie for market dominance. She notes that several leading companies have signaled their intentions to compete aggressively on pricing, akin to the strategies employed by cloud service giants Alibaba and Tencent. “In the next six to twelve months, LLMs with lower capacities will gradually be eliminated due to chip restrictions, high costs, and intensifying competition,” Pau predicts.
Yuan Hongwei, the chair of Shenzhen-based venture capital firm Z&Y Capital, opines that only two to three general-purpose LLMs will ultimately emerge as market dominators. Accordingly, Z&Y Capital places a premium on backing startups with experienced founders. Notably, they chose to support Baichuan Intelligence, a fledgling company founded by Wang Xiaochuan, the creator of China’s second-largest internet search engine, Sogou Inc. Baichuan has embarked on the ambitious mission of building an open-source AI model to rival Meta Platform’s Llama 2. In late August, Baichuan secured Beijing’s approval to launch a public chatbot, making it one of the first five companies to achieve this milestone. Wang anticipates closing a second funding round valuing the company at $1 billion.
“We see an opportunity here,” Yuan explains. “Wang himself is leading this project. Given his understanding of the digital business, his success with Sogou, and his industry-wide influence, we believe this is our best bet.”
Several other prominent entrepreneurs and tech executives have also thrown their weight behind new Chinese AI startups. This includes Kai-Fu Lee, Google China’s former chief, and Yan Juejie, a former vice-president of SenseTime. Meanwhile, some observers argue that China’s tech giants—Alibaba, Tencent, and Baidu—hold the most substantial advantages, thanks to their vast user bases and extensive service portfolios. These companies could seamlessly offer generative AI services as an added feature for their cloud customers.
Tony Tung, managing director at Gobi Partners GBA, comments on the competitive landscape, stating, “The incumbent tech giants have inherited an unfair advantage of a majority of low-hanging fruit business scenarios from their established ecosystems.” He further notes that some investors may regret rushing to invest in LLM firms during the peak of the hype earlier this year. Many of these startups are currently grappling with the challenge of establishing robust business cases and are now exploring partnerships with tech giants or the possibility of acquisition.
“At this particular moment, investors have certainly sobered up quite a bit compared to early this year,” Tung concludes.
Conclusion:
China’s AI market is reaching a critical juncture. The surge in LLM development has led to saturation and challenges, prompting a phase of consolidation and intense competition. Only a select few LLMs are expected to thrive, while experienced founders and strategic partnerships become pivotal. Established tech giants with extensive user bases have a distinct advantage, and investors are now taking a more cautious approach, emphasizing sustainable business models and differentiation in a crowded landscape.
