China is lagging behind the US in the development of its own OpenAI technology.

• China is striving to surpass OpenAI in the U.S. AI market, which is dominated by tech giants Microsoft, Google, and Amazon, as well as well-funded startups such as Anthropic.
• While OpenAI and Google's Gemini dominate foundational large language models, China is narrowing the gap by utilizing open source LLMs such as Meta's Llama 1.
• The restrictions on U.S. export of advanced chips have hindered its AI acceleration, but Chinese market leaders like Huawei are intensifying their chip efforts, and China boasts nearly triple the number of AI undergraduate degrees compared to the U.S.

The success of Nvidia in the stock market highlights the importance of chip quality and availability in determining the victors in the generative AI era. However, measuring early leads in the space is not limited to just China, which is striving to produce its own chips or obtain more from Nvidia. Despite the presence of numerous Chinese tech giants and startups, no dominant gen AI contender to OpenAI has emerged yet in China.

China is attempting to overtake OpenAI's lead in the U.S. AI market, which is dominated by tech giants Microsoft, Google, and Amazon, as well as well-funded startups such as Anthropic, which received a $2.7 billion investment from Amazon this week.

The technology gap between the U.S. and China is perceived as significant, with Chinese companies looking to benchmark against ChatGPT, highlighting their lagging position.

Few companies can afford to develop their own large language model due to the significant capital required. Jenny Xiao, a partner at AI VC firm Leonis Capital in San Francisco, stated that Silicon Valley is leading the way in this area.

The U.S. continues to be the largest investment market for artificial intelligence, with VCs and corporate investors driving $31 billion in AI investment across 1,151 deals in 2022, led by large outlays in OpenAI, Anthropic, and Inflection. This compares with $2 billion in 68 deals in China, which marked a significant drop from 2022's $5.5 billion in 377 deals. The decline is partly due to restrictions on U.S. venture investment into China.

Rui Ma, an AI investor and co-founder of investment syndicate and podcast TechBuzz China, stated that China faces a significant disadvantage in developing the foundation models for Gen AI.

While China is behind in foundational models, which are mainly dominated by OpenAI and Google's Gemini, it is closing the gap by utilizing Meta's open-source large language model Llama 1. According to Triolo, the Chinese contenders, although behind, are improving on the U.S. model.

According to Ma, many China models are essentially copies of Llama, and it is widely agreed that these copies are about one to two years behind the leading U.S companies, OpenAI and its video-to-text model Sora.

In the years ahead, China has the potential to make a significant impact in the AI competition due to its tech talent.

A study by Marco Polo, a think tank run by the Paulson Institute, reveals that the U.S. is home to 60% of top AI institutions and remains the leading destination for elite AI talent at 57%, compared to China with 12%. However, China outperforms the U.S. in producing top-tier AI researchers based on undergraduate degrees, with China at 47% and the U.S. lagging with 18%. Furthermore, among top-tier AI researchers working at U.S. institutions, 38% have China as their country of origin, compared with 37% from the U.S.

New Chinese gen AI market entries can also achieve rapid mass adoption. Baidu's ChatGPT competitor, Ernie Bot, reached 100 million users in just one year after its August 2023 release. Samsung is integrating Ernie AI into its new Galaxy S smartphones, while Apple is in talks with Baidu to supply the iPhone 16 with the Chinese company's gen AI technology.

According to Leong, Baidu's Ernie Bot models are considered among the most advanced within the current slate of AI contenders.

Major players in the Chinese technology market, including cloud companies like Baidu and Alibaba, social media players ByteDance and Tencent, and tech companies SenseTime, iFlyTek, Megvii, and Horizon Robotics, as well as research institutes, are funding the efforts of several other Chinese companies.

Alibaba and Hongshan, previously Sequoia China, are funding Moonshot AI, which is constructing large language models capable of processing extensive content inputs. Meanwhile, Kai-Fu Lee, the former Google China president, has developed an open source gen AI model, 01.AI, with funding from Alibaba and his firm Sinovation Ventures.

The development of China's AI industry has been hindered by U.S. restrictions on exporting high-end AI chips, which are primarily controlled by Nvidia, as part of a new tech rivalry between the two countries.

Although Chinese AI developers strive to create native solutions, they mostly depend on foreign hardware, mainly from U.S. companies, which poses a risk in the current political context, according to Bernard Leong, founder and CEO of Analyse Asia in Singapore.

The ongoing tensions between the U.S. and China over technology innovation and national security issues are leading to a split in gen AI development, with two parallel ecosystems for the technology, one in the U.S. and one in China. This is a pattern of other impactful technologies caught up in superpower tech arms races. Regulations and bans over sensitive, cutting-edge technologies are likely to result in this outcome. For example, ChatGPT is blocked in China while Baidu's Ernie Bot can only be accessed in the U.S. with a mainland Chinese cell phone number. According to Xiao, "U.S. companies can't go into China and Chinese companies can't go into the U.S."

Gina Raimondo, the U.S. Secretary of Commerce, has stated that the goal of U.S. curbs on AI chip exports is to prevent China from acquiring or producing advanced chips. While mainland China focuses on homegrown capabilities, Chinese companies SMIC or Huawei could be an alternative to Nvidia. However, the future for these alternates is uncertain if export controls cut off these companies from the most advanced designs for manufacturing. Triolo noted that Huawei recently developed a series of AI chips as a rival to Nvidia.

"The chip shortage is crucial for training foundational models that require specific chips, but it's not as significant for applications that don't need those chips," Ma stated.

The "real killer app" for gen AI, according to Triolo, will be in companies that are willing to invest in the technology as part of their business operations. Alibaba is focusing on integrating AI into its e-commerce ecosystem. Huawei, while competing more successfully against Apple's iPhone in the consumer market in the past year, also has broader ambitions, developing AI for specific industries including mining, using its in-house hardware, Leong said.

According to research by Boston Consulting Group, it may take some time for the wider gen AI market to expand beyond the tech industry. Out of 1,400 executives surveyed, 60 percent are waiting for gen AI regulations to develop, while only 6 percent of companies have trained their employees on gen AI tools.

The leadership of China has placed AI and tech concerns at the forefront, with the release of guidelines on AI in 2023 following ChatGPT's success, and subsequent adjustments to certain measures.

While the use of open source gen AI technology by many Chinese developers can foster collaboration and shared insights globally as AI progresses, Leong warned that it also presents challenges related to maintaining the quality and security of the models, managing bias, and preventing misuse of AI.

Triolo stated that China aims to control the content being released and wants their companies to take the lead, even if it means implementing strict measures.

The battle for control over OpenAI's mission highlights how ethical and social concerns hinder gen AI advances in both China and the U.S. Additionally, in China, maintaining control over gen AI applications, particularly in sensitive areas, could slow AI acceleration, as stated by Leong.