People learn about Baidu's artificial intelligence chatbot Ernie Bot during the 2nd World Digital Trade Expo at Hangzhou International Exhibition Center on November 23, 2023 in Hangzhou, Zhejiang Province, China.
China News Service | China News Service | Getty Images
Nvidia The stock market's rocket ship ride highlights how the quality and availability of chips will influence the winners in the era of generative artificial intelligence. But there's another side to measuring early leads in this space. In China, which is seeking to produce its own chips or acquire more from Nvidia, no dominant AI competitor to OpenAI has yet emerged among dozens of Chinese tech giants and startups.
Late in the game, China is seeking to take the OpenAI lead in the broader US AI market shaped by the tech giants Microsoft, alphabet Google and AmazonAnd well-funded startups including Anthropic, which this week received a $2.7 billion cash infusion from Amazon.
In this fast-moving field, the gap between the United States and its technological rival China is wide. “Leading Chinese companies are benchmarking against ChatGPT, which indicates how far behind they are,” said Paul Triolo, senior vice president for China and director of technology policy at Dentons Global Advisors in Washington, D.C.
“A lot of companies can't support their own big language model,” said Jenny Xiao, a partner at AI VC company Leonis Capital in San Francisco. “It takes a lot of capital. Silicon Valley is certainly way ahead of the game.”
The United States remains the largest investment market. Last year, AI startup funding accounted for nearly half of the $42.5 billion invested globally in AI companies, according to CB Insights. In the US, VC and corporate investors drove AI investment to $31 billion across 1,151 deals, led by significant expenditures at OpenAI, Anthropic, and Inflection. This compares to $2 billion in 68 deals in China, which is a significant decline from $5.5 billion in 2022 in 377 deals. This decline is partly due to restrictions on investment in US projects in China.
“China is at a severe disadvantage in building the basic models for the AI generation,” said Roy Ma, an AI investor and co-founder of investment syndicate and podcaster TechBuzz China.
But while China lags behind in core models, dominated by Google's OpenAI and Gemini, it is closing the gap with open source Meta and the large language model Llama 1, and Triolo said Chinese contenders, if falling behind, are improving on the American model.
“Many Chinese models are actually llama forks, and the consensus is that these forks are a year or two behind leading US companies OpenAI and their video-to-text model Sora,” Ma said.
China has the technical talent to make a difference in the AI competition in the coming years.
A new study by the Marco Polo Research Center, run by the Paulson Institute, shows that the United States is home to 60% of the best AI organizations, and that the United States remains by far the leading destination for elite AI talent at 57% of the total, compared China by 12%. But the research finds that China outperforms the United States by some other measures, including outperforming the United States in producing top-tier AI researchers, based on university degrees, with China at 47% and the United States behind at 18%. . Additionally, among senior AI researchers working at US institutions, 38% are from China as their country of origin, compared to 37% from the United States.
New entries to the Chinese AI market could also reach widespread adoption quickly. Baidu's ChatGPT competitor, Ernie Bot, released in August 2023, reached 100 million users by the end of the year. Samsung is planning to integrate Baidu's Ernie AI into its new Galaxy S smartphones, while in another high-profile development that speaks to US-China relations, Apple is in talks with Baidu about equipping the iPhone 16 with the Chinese company's new-generation AI technology.
Among the current list of AI contenders, Baidu's Ernie Bot models are among the most advanced, according to Leung.
Several other Chinese companies are moving forward, with funding from major players in their technology market. Helping in this effort are large cloud companies such as Baidu and Alibaba, social media players ByteDance and Tencent, technology companies SenseTime, iFlyTech, Migvi, and Horizon Robotics, as well as research institutes.
Moonshot AI, funded by Chinese e-commerce giant Alibaba and VC firm Hongshan (formerly Sequoia China), is building large language models that can handle long content inputs. Meanwhile, former Google China head Kai-Fu Lee has developed an open source model for artificial general intelligence, 01.AI, with funding from Alibaba and his company Sinovation Ventures.
While China has accelerated the development of its domestic chip industry and advanced AI, its AI development has been limited in part by US restrictions on the export of advanced AI chips, a market that Nvidia has cornered, as part of a new battleground for US-U.S. technological superiority. United States and China.
“Despite efforts to develop home-grown solutions, Chinese AI developers still largely rely on foreign hardware, especially from US companies, which represents a vulnerability in the climate,” said Bernard Leung, founder and CEO of Analyze Asia, a consultancy in Singapore. current geopolitics.
Ongoing tensions between the United States and China over technological innovation and national security issues are leading to a rift in the development of artificial general intelligence, similar to the pattern of other influential technologies caught in technological arms races between great powers. Given the regulations and bans on sensitive and cutting-edge technologies, the likely outcome is two parallel AI generation ecosystems, one in the United States and the other in China. ChatGPT is blocked in China while Baidu's Ernie Bot can only be accessed in the US using a mainland Chinese mobile phone number. “American companies cannot enter China, and Chinese companies cannot enter the United States,” Xiao said.
US Commerce Secretary Gina Raimondo said the goal of US restrictions on AI chip exports is to prevent China from acquiring or producing advanced chips. As mainland China focuses on local capabilities, Chinese companies SMIC or Huawei could be an alternative to Nvidia. But the future of alternatives is likely to be uncertain if export controls prevent these companies from obtaining the most advanced designs in manufacturing. Triolo noted that Huawei recently developed a series of artificial intelligence chips as a competitor to Nvidia.
China is making progress in applying artificial intelligence to certain categories, such as computer vision. “Chip shortages are very important for training basic models where you need certain chips, but for applications, you don't,” Ma said.
The “real killer application” of the AI generation, according to Triolo, will be in companies willing to pay money to harness the technology as part of their business operations. Alibaba is focusing on integrating artificial intelligence into its e-commerce ecosystem. Leung said Huawei, although competing more successfully against Apple's iPhone in the consumer market last year, also has broader ambitions, working to develop artificial intelligence for specific industries including mining, using its own in-house hardware.
Boston Consulting Group research suggests that it may be some time before this broader AI market expands beyond technology. 60% of the 1,400 executives surveyed are waiting to see how AGI regulations evolve, while only 6% of companies have trained their employees on AGI tools.
AI and technology issues are at the forefront of the Chinese leadership's concerns, with the country launching guardrails on AI in 2023 following the ChatGPT hack, and then tweaking some measures.
The open source AI technology used by many Chinese developers could encourage global collaboration and lead to shared insights as AI advances, but Leung said open source also leads to issues around ensuring the quality and security of models, as well as managing bias and potential. Misuse of artificial intelligence.
“China wants to make sure the content doesn't spread,” Triolo said. “They also want their companies to take the lead, and they are willing to rein in the crackdown.”
Ethical and social concerns are hindering the progress of AI in China and elsewhere, including the United States, as we see in the battle for control of the OpenAI mission. Within China, there is another factor that could slow the acceleration of AI, according to Leung: maintaining control over public AI applications, especially in areas sensitive to state interests.
