Kai-Fu Lee: Such startups may face "liquidation" from investors
Kai-Fu Lee, chairman of Sinovation Ventures, hopes that his start-up Zero One Everything can learn from Huawei's strategy of becoming a market leader. IMAGE COURTESY OF LUCAS SCHIFRES—FORTUNE
Even as researchers burn cash and little yields results, venture capitalists are still willing to spend billions of dollars to invest in the latest AI startups, and the golden age may be nearing its end. On March 27, at the Fortune Innovation Forum in Hong Kong, China, Kai-Fu Lee, chairman and CEO of Sinovation Ventures, said that as AI technology matures, unprofitable AI companies will soon face a "liquidation."
Too many large language model (LLM) startups are too focused on pursuing breakthroughs and too little on commercializing their research, Lee said. In a conversation with Alyson Shontell, editor-in-chief of Fortune magazine, Kai-Fu Lee said: "Many of the big language model companies are run by researchers, and they are just concerned with building great models. This stage of the science contest should be over. ”
If America's three tech giants have anything in common, it's that they've all successfully commercialized emerging technologies such as Microsoft's personal computers and Apple's and Google's smartphones.
Kai-Fu Lee was the president of Google China and a researcher in the field of artificial intelligence. He founded an AI startup in March 2023. The company is called Zero One Ten Thousand Things (01. AI), which was valued at more than $1 billion in less than eight months after its founding.
Mr. Lee said his former employer, Google, was a lesson in the past. Although Google has by far the densest network of AI talent in the world, it has been robbed of the lead in the AI space by OpenAI, which Lee believes is because Google always wastes time and resources on competing internal initiatives.
"As a startup, if you have too many researchers and you promote a culture where everyone can try their own ideas, you're going to run out of money very quickly," he said. ”
Huawei's focus vs Google's "Hundred Flowers"
Kai-Fu Lee believes that if his company wants to one day become a global leader in artificial intelligence, it must make the most of every spending extremely efficiently.
On March 27, the AI expert used Huawei as an example to illustrate the practical effects of this focus. Huawei, China's leading telecommunications equipment manufacturer, has leveraged an inconspicuous technological advancement by Turkish IT researcher Edar Alikan to bring almost the entire company together to commercialize his polar code technology breakthrough. This allowed Huawei to eventually overtake larger Western competitors such as Ericsson and seize most of the 5G mobile network market.
"This strategy has had an important impact," Lee said. We're taking the same approach, working very hard to save on GPU costs. ”
He believes that by focusing on efficient execution, the gap between 010000 and American companies like OpenAI has narrowed from eight years to less than 12 months in just one year. All studies of Zero One are published on open websites such as Hugging Face.
By contrast, if competitors in the AI space adopt what Lee calls Google's "let a hundred flowers bloom" strategy, it will be difficult for them to become profitable.
Kai-Fu Lee pointed out: "At this time, the company will usher in 'liquidation'. Investors will ask: What are your results, what is your P&L, what is your revenue, what is your growth rate, when are you going to break even?"
If AI startups can't give a convincing answer, the era of it concentrating on technology as if it were participating in a "science race" will be over. (Fortune Chinese Network)
作者:Christiaan Hetzner
Translator: Liu Jinlong
Reviewer: Wang Hao