Written by 
Entering the second half of 2024, becoming OpenAI is no longer the common goal of large model companies. Slowed technological progress and hesitant investors have led to the gradual differentiation of strategies among six well-known Chinese large model companies.
The adjustment of 01.AI has become relatively clear: focusing on consumer products, promoting AI search applications; while also opening up a business front, continuing to provide standardized services through APIs, and collaborating with ecosystem partners to jointly serve key scenarios such as e-commerce live streaming and marketing.
We have learned that 01.AI is currently planning to spin off an AI application company, which will be based on 01.AI’s large models to develop applications in areas such as gaming. The company is currently seeking external financing at a valuation of tens of millions of dollars.
After the split, the new company is named ‘Oasis,’ with Ma Jie, co-founder and vice president of 01.AI, serving as chairman. The split is expected to be completed by the end of this year. After the split, 01.AI will remain Oasis’s largest shareholder.
Ma Jie, who will be in charge of ‘Oasis,’ previously served as a vice president at Baidu and was responsible for Baidu‘s metaverse product ‘XIRANG.’
Prior to that, he developed personal antivirus software and enterprise-level security hardware products at Rising Antivirus Software Company from 2000 to 2010.
Afterwards, Ma Jie joined Sinovation Ventures founded by Kai-Fu Lee and became one of the first Entrepreneurs-in-Residence (EIR). Later incubated by Sinovation Ventures, he founded China’s first SaaS-based cloud security service company ‘Anquanbao.’ The company was acquired by Baidu in 2015 and Ma Jie joined Baidu.
Before the split, Ma Jie was in charge of cloud gaming, cloud architecture, and other directions that were originally one of the revenue sources for 01.AI, now transferred to Oasis; security and compliance responsibilities under Ma Jie will continue at 01.AI.
An insider close to the matter said: “01.AI now wants to focus more on model and computing power output. Allowing some applications to explore product and commercial development paths independently is beneficial for both product projects and 01.AI.”
Previously, 01.AI had tried several AI products targeting consumers (to C), except for a productivity product with cash flow, other products have ceased operations or are only maintained.
It is understood that AI search will become the key product direction for 01.AI next. Since September, 01.AI has tested and launched two AI search products.
The business-to-business (B2B) strategy of 01.AI is also clearer in the second half of this year: providing an integrated service of ‘infra + large models + applications,’ targeting three key scenarios: first is collaborating with governments and enterprises to build large model computing power and service platforms while providing model capabilities; second is offering ‘Ruyi’ digital human solutions for e-commerce live streaming, office meetings, etc.; third is providing ‘Wanshi’ short video solutions for marketing scenarios.
A primary market practitioner told us that splitting up benefits both parties: For 01.AI, it means immediate income recognition without significant burden as they don’t need to invest money into new businesses; For Oasis, it allows them to access large models at lower costs while being able to operate independently with separate financing.
It’s as if the ‘fast forward’ button has been pressed, and the large-scale industry is accelerating through the stages from hot financing to intense competition, and then to a period of being questioned in the technology entrepreneurship cycle.
The first business split in the large-scale industry also came earlier – it happened when the company was less than two years old. Among the previous batch of AI ‘four little dragons,’ YITU was the earliest to split its business. In 2016, YITU, founded in 2012, spun off its medical business.
“The remaining large-scale companies probably do not have such (subsidiary splitting) development plans for now,” analyzed a primary market practitioner. “For other companies, one reason is that their division of labor strategy is different; another reason is that their founders are younger and have different views on business and capital compared to Kai-Fu Lee.”
Kai-Fu Lee worked as an executive at Microsoft and Google early in his career before founding Sinovation Ventures investment firm, witnessing many ups and downs in businesses. In May this year, when enthusiasm for investment in large models had not yet cooled down, Li Kaifu told us that going public is very important for large model companies from a commercial perspective.
He said at that time: “Going public isn’t about cashing out or making money.” Instead, going public will assess whether a company has high-quality revenue streams and can sustain growth. Going public can also bring other financing opportunities like private placements.
“Nowadays with Chinese big model startups…forget about investments from America or Europe; Middle East countries are still observing…the VC environment in China has changed too; many companies are already expensive now—do you expect them to be valued at $4-5 billion USD next round?”
Amongst these six big model companies, 01.AI places great emphasis on product ROI (Return on Investment) and financial indicators. “Large model companies sooner or later will face scrutiny from investment banks; they need to prepare ahead of time because otherwise these house-of-cards will eventually collapse,” Kai-Fu Lee once said.
SEE ALSO: 01.AI Reveals Its ToB Solution Based on the Yi Model