Baidu shares jumped almost 4.5% during Wednesday trading in Hong Kong following expectations-beating revenue from the Chinese tech giant. Baidu is trying to solidify an early lead in the race to win China's AI market, starting with the launch of its ChatGPT-like ERNIE Bot earlier this year.
Baidu generated revenue of $4.7 billion for the three months ending Sept. 30, a 6% year-on-year increase. The company also earned $916 million in net income, compared to a $20.6 million loss for the same quarter last year.
“Our AI-centric business and product strategy should set the stage for sustained multi-year revenue and profit expansion within our ERNIE and ERNIE Bot ecosystem,” CEO Robin Li said in a statement on Tuesday.
Baidu and Li, also the company's founder, hope AI will revive the tech company's fortunes, after the company lost ground to competitors like Tencent and Alibaba. The company is primarily known for its search engine, but is now shifting to new sectors like automated driving and generative AI.
Baidu launched ERNIE earlier this year, though observers were underwhelmed by the presentation compared to its non-Chinese peers like Google and Microsoft. Yet the Chinese company has continued to update the model and its chatbot, releasing ERNIE 4.0 in October.
The tech company also shared details on its robotaxi service, named Apollo Go, which operates in major cities like Wuhan, Shenzhen and Beijing. The autonomous ride-hailing service carried 821,000 passengers last quarter, a 73% increase from a year ago.
China's race for AI
Baidu is part of a growing rush in China's tech sector to launch generative AI products, and arguably leading the way: The company is the only Chinese firm featured in Fortune's inaugural "AI Innovators" list, released on Tuesday, which highlights 50 companies at the forefront of AI.
The company's ERNIE bot is perhaps China's closest equivalent to OpenAI's ChatGPT, currently banned in China. The bot outperforms ChatGPT in several Chinese-language tasks, Baidu says.
Yet Baidu's big tech peers are also barreling into the space. Alibaba, Tencent and <a href="https://www.scmp.com/tech/big-tech/article/3227583/chinese-e-commerce-giant-jdcom-unveils-own-large-language-model-chatrhino-drive-increased-adoption">JD.com</a> have all announced their own large language models. (JD.com CEO Sandy Xu Ran is joining Baidu's board as an independent director, the company announced Tuesday). Several smaller AI companies and startups are also developing their own models: There are now over 130 large language models being developed in China today, according to one estimate.
Yet China's AI companies need to work within the limits of what Beijing allows. According to rules approved in July, Chinese developers must ensure that their AI services align with "core socialist values" and national security. Yet they also highlight the importance of innovation, and revisions weakened provisions on how to penalize companies that break the rules.
Developers in China face another threat: U.S. rules limiting the sale of advanced AI chips from firms like Nvidia to Chinese companies. Last week, Alibaba shelved its plan to spin-off its cloud computing division as an independent company, blaming uncertainty from U.S. export controls.
On Tuesday, Li warned that these restrictions could force the consolidation of large language models in China. Baidu has enough AI chips stockpiled in the near term, he said.
Regulation and access to chips are the primary risks faced by China's AI sector, seemingly more real-world concerns compared to the recent worries in the U.S. about safety and more existential threats from the new technology. OpenAI, the developer behind ChatGPT, fired its CEO Sam Altman on Friday, reportedly due to concerns that he was moving too quickly on releasing the organization's products. (Altman returned as OpenAI's CEO on Wednesday morning, ending days of negotiations to bring him back to the organization.)
“The debate around the existential risks around [artificial general intelligence] has not been as much of a priority within the Chinese AI community, which has focused more on developing solid use cases for enterprise deployments of generative AI,” Paul Triolo, an associate partner for China and technology policy lead at the advisory firm Albright Stonebridge, told Fortune on Monday.
Conversations on AI risk "will be very much a government-driven thing in China. No CEO is going to be forced out because of disputes over the lack of guardrails to tackle the existential risks of AGI," he added.
Baidu established an ethics committee in October to "guide the practices of technology professionals," the company said in its earnings statement.
Additional reporting by Nicholas Gordon
This story was originally featured on Fortune.com