Enterprise

Kai-Fu Lee wanted to teach the US about Chinese AI. Instead he provoked a rivalry.

World-renowned researcher and investor Kai-Fu Lee wanted to spread knowledge about AI in China with the U.S. so both countries could succeed. Now he may be forced to live by a different philosophy.

Kai-Fu Lee, founder of Sinovation Ventures, at a Bloomberg panel at the World Economic Forum in Davos, Switzerland, in 2018.

After years of trying to bind AI businesses in the U.S and China, Kai-Fu Lee — seen here at the World Economic Forum in Davos in 2018 — is struggling to keep a grip on the two tech power spheres as geopolitical pressure rips them apart.

Photo: Jason Alden/Bloomberg via Getty Images

Click here for more coverage

Contributing to a $6 million investment in a fledgling AI startup may not be a headline-grabbing move for Sinovation Ventures, the Chinese venture capital firm led by legendary AI technologist Kai-Fu Lee.

But those 2021 and 2022 investments in U.S.-based HPC-AI Tech could serve a greater purpose for the celebrity AI researcher and longtime investor in AI developed in China. They could help Lee maintain a decades-long cross-border bond with the U.S. that is slowly eroding.

Lee had years of firsthand experience learning how the U.S. and China worked together to advance AI before launching Sinovation Ventures in 2009. Since 2013, the venture capital firm has participated in investment rounds totalling at least $1.06 billion in AI companies around the world, according to data provided by CB Insights and analyzed by Protocol.

Before becoming an investor, Lee led AI research in China for U.S. tech giants Microsoft and Google, watching his groundbreaking speech recognition and natural language processing technologies come to fruition in search, mobile, and conversational AI products that touched people’s lives in both countries and across the globe. Along the way, he seemed to harvest gems of wisdom from both tech cultures.

Lee’s cross-border philosophy of absorbing knowledge from one country and applying it in the other is evident in his investment approach. Sinovation, which has a presence in Beijing, Shanghai, and other cities in mainland China, focuses its investments on Chinese companies.

However, over the years the VC firm has also funded U.S. companies to gather knowledge, as is the case with HPC-AI Tech. The company, whose chief technology officer is based in China, aims to sell its AI optimizing software to the U.S. market, its founder Yang You told Protocol in September.

Lee has admonished Americans, urging them to pay attention to China’s AI accomplishments. “They’re under a disadvantage because Chinese companies are studying Chinese companies and American companies, but American companies don’t study Chinese companies, so American companies have only half the knowledge that Chinese companies have, and that’s a big problem,” Lee said at a 2016 TechCrunch conference in Beijing.

It was a concept he tried to get across with his bestselling 2018 book, “AI Superpowers: China, Silicon Valley, and the New World Order.”

The book depicted the U.S. and China traveling on parallel tracks, helping advance an important category of tech that could benefit everyone. Lee believed the U.S. could benefit from knowing more about the rapid advancements China had made in areas like machine learning for ecommerce and social media, and deep learning and computer vision for autonomous vehicles.

Instead of grasping the book’s nuanced takeaway, many misconstrued its narrative, calling it a shallow story portraying a winner-take-all AI fight.

Ironically, what may have seemed to Lee like a win-win way to share knowledge — while conveniently attracting investment for the Chinese AI startups Sinovation Ventures funded — had the opposite effect, serving up red meat for national security hawks and ripe fodder for hyperbolic headlines amplifying the so-called AI race between the U.S. and China.

The media mischaracterization has helped fuel extreme reactions from U.S. policymakers, stymying cross-border tech business partnerships, blocking China from AI supplies built in the U.S. like powerful chips used to train machine learning models, and potentially clogging the flow of U.S. dollars toward the very Chinese AI companies Lee bets on.

Now after years of playing an increasingly influential role as a technologist and businessman trying to bind AI businesses in the U.S and China, Lee is struggling to keep a grip on the two tech power spheres as geopolitical pressure rips them apart. He’s still holding on, investing in AI companies straddling both worlds — including HPC-AI Tech; AI drug discovery company Insilico Medicine, which has operations in the U.S., China, the U.K., and elsewhere; and China’s WeRide, which has a license to test its autonomous vehicles in San Jose.

Years after his book came out, the distorted, oversimplified version of Lee’s message was still repeated. In November 2021, a Bloomberg reporter asked him, “In the AI race between the U.S. and China, who will win in decades to come?”

When a weary-looking Lee looked into his laptop camera to answer the reporter’s question, he patiently explained that there was no head-on AI fight happening between the two countries. Rather, he said each had its own strengths in AI — the U.S. in areas such as enterprise AI software, and China in things like robotics for manufacturing. Both did well in AI for autonomous vehicles and consumer internet platforms, he told the reporter.

“I hope AI will end up making companies in both countries winners,” Lee said.

But hope may not be enough for Lee’s knowledge-sharing philosophy to function much longer. He could be forced to live by a new credo.

Sinovation Ventures did not respond to multiple requests to interview Lee for this story.

A tech celeb emerges

Thirty years ago, Lee stood under bright lights on the set of one of the top morning shows in the U.S.

Joan Lunden, longtime host of “Good Morning America,” looked at Lee quizzically. She needed an explanation for the outlandish number he had just uttered.

“A trillion different sentences?” marveled Lunden.

Lee’s eyes widened and he allowed a wry yet sweet smirk to appear across his face.

“Yes,” he responded, explaining that the futuristic technology he helped build, now at work in the Macintosh desktop computer at their side, could recognize human speech because it had been trained using a trillion sentences.

It was 1992. Moms dishing out instant oatmeal for the kids or getting ready for work may have wondered why Lee — then age 30 but looking all of 18 — was peering out from their TV screens. One day this youthful computer nerd with the twinkle in his eye would amass huge global followings on yet-to-be invented social media platforms favored both in the U.S. and China, and even spark a turf war between rivals Microsoft and Google.

“There are two breakthroughs here,” Lee told Lunden, explaining the significance of voice recognition research he’d conducted for Apple that would later form a foundation for work he’d do while running AI labs for Microsoft and Google in China.

The voice recognition technology Lee demoed that day was baked into an Apple product called Casper. Lunden and Apple CEO John Sculley took turns giving Casper pre-scripted voice commands to direct the computer to open a calendar application to schedule an appointment and automatically write a check to pay a phone bill.

As Lee stood there, the glitz of the “GMA” set may have seemed fitting. He clearly had exhibited enough confidence for Apple to make the atypical decision to have its young research scientist play sidekick to the company’s more seasoned CEO in front of a TV audience of millions. Already Lee had made a name for himself through highly regarded Ph.D. research he conducted while in the computer science department at Carnegie Mellon University.

Lee’s CMU work used mathematical learning to enable continuous speech recognition. While it did not involve entirely new concepts, Rashid said it was a risky approach that earned Lee a reputation for making bold yet smart moves that would help actualize AI.

“That’s something people weren’t necessarily expecting was going to work. A lot of the orthodoxy during that period of time was that the way to solve these problems was to do more rule-based systems — systems where you were basically taking the expertise of an individual and trying to codify that in some fashion,” Rashid said.

There was even a glimmer of catchy branding in Lee’s wonky research. Rather than calling it something academic and complicated, it had a snappy, memorable name: Sphinx.

The creature’s lion body represented a large database, its human head the system’s knowledge, and the bird wings its speed, Lee would later explain in his 2011 book, “Making a World of Difference.” Sphinx got attention and was even covered in a 1988 New York Times article.

“It was one of the things that he became well known for. It was a surprise. In some sense it helped launch his career,” Rashid said.

As Lee moved up the professional ladder, his self-assurance would prove to be immensely valuable. While launching the new Microsoft Research, or MSR, lab in Beijing in 1998, Lee would have to foster relationships with universities to recruit talent and even schmooze with local government officials in the hopes that they’d support the lab.

He also had the skills to get the attention of decision-makers at Microsoft. Having moved to the U.S. in the early 1960s at age 11 from his birthplace of Taiwan, Lee was immersed in American culture and the English language at a young age.

“He could articulate what his plans were about his software,” said Nathan Myhrvold, founder of tech startup investment business Intellectual Ventures, who, along with Rashid, hired Lee for the pivotal role guiding the creation of Microsoft’s China lab. “An articulate researcher can sell you on his or her plan better than an inarticulate researcher,” Myhrvold said.

Lee helped the lab garner a good reputation inside China. “Within just a few years MSR China was famous within China as a super cool place to work, where you could do really, really interesting stuff,” Myhrvold said.

Google wanted in on the action. The company, which had ambitious plans to upend Microsoft’s dominance through search and email in the early 2000s, managed to poach Lee in 2005 to help it establish its own China operation.

An articulate researcher can sell you on his or her plan better than an inarticulate researcher.”

Microsoft was not happy. That July, it sued Google arguing Google had violated Microsoft’s noncompetition agreement with Lee. The companies settled the suit a few months later.

Lee’s networking and leadership acumen and his strengths in groundbreaking speech recognition were no match for the geopolitical and business headwinds Google faced in China. Lee quit the company in 2009; by the following year, Google began shuttering or moving products and services away from the country that Lee helped nurture for the China market.

Lee took what he’d learned and parlayed it into his next endeavor.

“[W]e could see the progress Android was making. And we knew that would be the answer in China,” he told Wired in 2018. “So when I left Google … I started an investment company specifically for mobile internet, mostly Android-based. This was Sinovation Ventures. We invested in social networks, education, entertainment. We got very good in these areas before AI,” he said.

‘Learn what’s so great in America, and then move to China’

At an “AI Superpowers” book tour event held in 2018 at the UC Berkeley Haas School of Business, a business student from China studying in California had a question for Lee.

“You mentioned that the AI development in China and the U.S. are two parallel universes. So what kind of advice would you give for future entrepreneurs who want to work between these two countries?” the student asked.

Lee was blunt. “I think cross-border will be very difficult,” he said. “I wouldn’t advise you to work for an American company trying to do business in China, or a Chinese company trying to do business in America — tech company anyway.”

What Lee said next prompted nervous laughter from the crowd of wannabe Silicon Valley startup execs: “Learn what’s so great in America, and then move to China and use that knowledge,” Lee said. As the crowd gasped, he paused for a beat. Then he added, “Or the reverse.”

Startling, but revealing, there in a nutshell was Lee’s guiding philosophy of cross-border learning, adjusted for a new geopolitical reality.

Learn what’s so great in America, and then move to China and use that knowledge”

During the time Lee gave his talk at Berkeley, HPC-AI Tech’s You was a researcher studying ways to streamline training of deep neural networks at UC Berkeley’s Electrical Engineering and Computer Sciences, a half-mile down Gayley Road from the business school.

Like so many aspiring AI entrepreneurs, You looks to Lee for advice. In fact, like some of Lee’s 48.6 million fans on China’s Twitter-like platform Weibo, You uses a term of respect and endearment for him.

“I call Kai-Fu ‘Professor Lee.’ Kai-Fu Lee is not a professor, but [it is] to respect him,” You told Protocol during an interview in September, a day before he had plans to meet with Lee for dinner in Singapore.

“I hope I can get some suggestions from him for my company on how to build the product, and also how can [I] make the company attract more VCs in the future,” You said.

‘Professor Lee’ rides the decoupling wave

Whether a startup like HPC-AI Tech will adjust its sales strategy as U.S.-China trade tensions grow is anyone’s guess. But as the Biden administration signals restrictions on cross-border investments that could affect national security, Lee seems determined to find windows of opportunity.

“Decoupling will create challenges for many companies. They’ll lose sales, but they’ll also create opportunities for ones that can straddle and take advantage of the decouple,” Lee said in April during a Bloomberg event.

“I don’t worry about a worst case because any extreme measure in decoupling will end up in a very severe lose-lose,” Lee said.

Until recently, Lee’s philosophy of applying cross-border experience has served him well. Sinovation actively coaxed U.S. investors to help fund the Chinese AI startups in its portfolio as AI became more viable in China — in part as a result of the massive data influx generated by mobile device use he recognized as so important while at Google. By 2016, the company had raised the equivalent of $675 million for its Chinese and American funds.

Lee’s AI investment activity fluctuated over the years, but CB Insights data shows that 2021 was an especially big one. He participated in eight funding rounds in AI companies in 2021 worth at least $380.6 million. In contrast, he joined seven funding rounds in the first three quarters of 2022 worth around a third of that amount — at least $130.7 million. The full amounts of Sinovation’s AI-related funding rounds evaluated for this story were not disclosed publicly.

Sinovation has found it more difficult to attract U.S. investors as U.S. tariffs and export controls darken prospects for AI business collaboration between the two countries. The company reportedly closed its Silicon Valley office in 2019. Then, this August, The Wall Street Journal reported that the investment firm first closed its latest China-based fund after raising $200 million, approaching the midway of the $500 million it had hoped to attract as funds raised by China-focused general partners had dropped to multiyear lows, according to the report.

Still, ongoing investments in two of Lee’s largest AI bets — China’s autonomous vehicle company WeRide and Insilico Medicine, which both operate in the U.S. and China — indicate that he has not entirely given up on transcontinental AI deals. Sinovation joined investment rounds in WeRide totaling $167 million in 2017 and as late as 2021. In 2019 and 2021, CB Insights reported that Sinovation contributed to funding rounds totaling $292 million. And between the first quarter of 2017 and Q2 2022, CB Insights data shows WeRide raised $767 million in investment rounds involving U.S.-based venture capital firms.

Despite what seems like eternal optimism for cross-border AI collaboration between the U.S. and China, Lee has felt the pang of disappointment from lost opportunity for years. In “Making a World of Difference,” he cited a Chinese proverb about finding common ground and accepting differences: “yi zhong qiu tong.”

“Over the years,” he wrote, “I have learned that if each country could understand the other’s history, culture, and viewpoint, and accept that there are some issues that the two countries will ‘agree to disagree,’ there would be tremendous progress.”

Protocol data researcher AJ Caughey assisted with data analysis on this story. This story was updated to clarify the nature and timing of Sinnovation Ventures' U.S. investment funds.

Fintech

Judge Zia Faruqui is trying to teach you crypto, one ‘SNL’ reference at a time

His decisions on major cryptocurrency cases have quoted "The Big Lebowski," "SNL," and "Dr. Strangelove." That’s because he wants you — yes, you — to read them.

The ways Zia Faruqui (right) has weighed on cases that have come before him can give lawyers clues as to what legal frameworks will pass muster.

Photo: Carolyn Van Houten/The Washington Post via Getty Images

“Cryptocurrency and related software analytics tools are ‘The wave of the future, Dude. One hundred percent electronic.’”

That’s not a quote from "The Big Lebowski" — at least, not directly. It’s a quote from a Washington, D.C., district court memorandum opinion on the role cryptocurrency analytics tools can play in government investigations. The author is Magistrate Judge Zia Faruqui.

Keep Reading Show less
Veronica Irwin

Veronica Irwin (@vronirwin) is a San Francisco-based reporter at Protocol covering fintech. Previously she was at the San Francisco Examiner, covering tech from a hyper-local angle. Before that, her byline was featured in SF Weekly, The Nation, Techworker, Ms. Magazine and The Frisc.

The financial technology transformation is driving competition, creating consumer choice, and shaping the future of finance. Hear from seven fintech leaders who are reshaping the future of finance, and join the inaugural Financial Technology Association Fintech Summit to learn more.

Keep Reading Show less
FTA
The Financial Technology Association (FTA) represents industry leaders shaping the future of finance. We champion the power of technology-centered financial services and advocate for the modernization of financial regulation to support inclusion and responsible innovation.
Enterprise

AWS CEO: The cloud isn’t just about technology

As AWS preps for its annual re:Invent conference, Adam Selipsky talks product strategy, support for hybrid environments, and the value of the cloud in uncertain economic times.

Photo: Noah Berger/Getty Images for Amazon Web Services

AWS is gearing up for re:Invent, its annual cloud computing conference where announcements this year are expected to focus on its end-to-end data strategy and delivering new industry-specific services.

It will be the second re:Invent with CEO Adam Selipsky as leader of the industry’s largest cloud provider after his return last year to AWS from data visualization company Tableau Software.

Keep Reading Show less
Donna Goodison

Donna Goodison (@dgoodison) is Protocol's senior reporter focusing on enterprise infrastructure technology, from the 'Big 3' cloud computing providers to data centers. She previously covered the public cloud at CRN after 15 years as a business reporter for the Boston Herald. Based in Massachusetts, she also has worked as a Boston Globe freelancer, business reporter at the Boston Business Journal and real estate reporter at Banker & Tradesman after toiling at weekly newspapers.

Image: Protocol

We launched Protocol in February 2020 to cover the evolving power center of tech. It is with deep sadness that just under three years later, we are winding down the publication.

As of today, we will not publish any more stories. All of our newsletters, apart from our flagship, Source Code, will no longer be sent. Source Code will be published and sent for the next few weeks, but it will also close down in December.

Keep Reading Show less
Bennett Richardson

Bennett Richardson ( @bennettrich) is the president of Protocol. Prior to joining Protocol in 2019, Bennett was executive director of global strategic partnerships at POLITICO, where he led strategic growth efforts including POLITICO's European expansion in Brussels and POLITICO's creative agency POLITICO Focus during his six years with the company. Prior to POLITICO, Bennett was co-founder and CMO of Hinge, the mobile dating company recently acquired by Match Group. Bennett began his career in digital and social brand marketing working with major brands across tech, energy, and health care at leading marketing and communications agencies including Edelman and GMMB. Bennett is originally from Portland, Maine, and received his bachelor's degree from Colgate University.

Enterprise

Why large enterprises struggle to find suitable platforms for MLops

As companies expand their use of AI beyond running just a few machine learning models, and as larger enterprises go from deploying hundreds of models to thousands and even millions of models, ML practitioners say that they have yet to find what they need from prepackaged MLops systems.

As companies expand their use of AI beyond running just a few machine learning models, ML practitioners say that they have yet to find what they need from prepackaged MLops systems.

Photo: artpartner-images via Getty Images

On any given day, Lily AI runs hundreds of machine learning models using computer vision and natural language processing that are customized for its retail and ecommerce clients to make website product recommendations, forecast demand, and plan merchandising. But this spring when the company was in the market for a machine learning operations platform to manage its expanding model roster, it wasn’t easy to find a suitable off-the-shelf system that could handle such a large number of models in deployment while also meeting other criteria.

Some MLops platforms are not well-suited for maintaining even more than 10 machine learning models when it comes to keeping track of data, navigating their user interfaces, or reporting capabilities, Matthew Nokleby, machine learning manager for Lily AI’s product intelligence team, told Protocol earlier this year. “The duct tape starts to show,” he said.

Keep Reading Show less
Kate Kaye

Kate Kaye is an award-winning multimedia reporter digging deep and telling print, digital and audio stories. She covers AI and data for Protocol. Her reporting on AI and tech ethics issues has been published in OneZero, Fast Company, MIT Technology Review, CityLab, Ad Age and Digiday and heard on NPR. Kate is the creator of RedTailMedia.org and is the author of "Campaign '08: A Turning Point for Digital Media," a book about how the 2008 presidential campaigns used digital media and data.

Latest Stories
Bulletins