The halcyon days when venture capitalists were content to spend billions on the latest artificial intelligence startup while researchers spent money without producing any results may be over. A “reckoning” is coming soon for artificial intelligence companies that struggle to turn a profit as new technology develops, Kai-Fu Lee, chairman and chief executive of Sinovation Ventures, said at the Fortune Innovation Forum in Hong Kong on Wednesday.
Lee said too many large language model (LLM) startups are focused on pursuing breakthroughs and too little on commercializing their work. “Many LLM companies are run by researchers who only care about creating a great model,” he said in a conversation with Luck Editor-in-Chief Alison Shontell. “This science fair must end.”
If there’s one thing that America’s top three mega-cap tech companies have in common, it’s that they have successfully monetized new technology—Microsoft with personal computers, Apple and Google with smartphones.
A former president of Google China and a researcher in the field himself, Li founded his own artificial intelligence startup in March 2023. The firm, called 01.AI, was valued at more than US$1 billion. less than eight months.
Lee said his own former employer, Google, could serve as a cautionary tale. Even though it has the densest network of AI talent in the world today, he argues that Google lost the lead to OpenAI because it wasted time and resources pandering to the competing agendas of all its employees.
“If you have too many researchers and a culture where everyone can try out their ideas, the startup will quickly run out of money,” he said.
Huawei vs Google Focus “Let a Hundred Flowers Bloom”
Lee argued that for his company to one day become one of the world’s leaders in this field, it would need to make the most of every dollar it spends.
On Wednesday, an artificial intelligence expert pointed to Huawei as an example of how such a trick could work in practice. A leading Chinese telecommunications equipment manufacturer has seized on the little-known achievement of Turkish IT researcher Erdal Arıkan, investing its efforts almost exclusively in commercializing its breakthrough in polar code. This allowed them to eventually surpass larger Western competitors such as Ericsson and continue to control a large portion of the 5G mobile network market.
“It made all the difference,” Lee said. “We’re taking the same approach of trying very, very hard to preserve GPUs. [costs]”
He believes that due to the focus on effective execution of 01.AI, which publishes all of its research into open source sites like Hugging Face has closed the gap with US companies like OpenAI from eight years to less than twelve months in just a year.
AI competitors that instead embrace Google’s “let a hundred flowers bloom” strategy, as Lee puts it, will struggle to achieve profitability by comparison.
“There comes a point of reckoning where investors will say: What can you show yourself?” – said Lee. “What is your profit and loss? What is your income? What is your height? When will you break even?”
If an AI startup doesn’t have a compelling answer, its “science fair” days are over.