The Global Contest for Breakthrough Technologies

Guest post by Mehran Gul on his new book: The New Geography of Innovation

I first met Mehran Gul in 2017 when we were both finalists for the Bracken Bower Prize—McKinsey & The Financial Times’ little sibling to their Business Book of the Year Award. Mehran went on to win that year, and we’ve stayed in touch as fellow travelers exploring the big questions shaping global innovation. His book The New Geography of Innovation just came out (see FT feature here) offering a fresh lens on how technology, policy, and entrepreneurship are transforming economies far beyond the traditional hubs. The topic is all the more relevant given the global contest emerging for dominance in AI. That’s why I asked Mehran to do a guest post for this edition. It is already available in Europe, and I’m excited to read his book when it comes out in the U.S. (available for pre-order).

---

Is the center of gravity for where the most consequential developments in technology are happening moving from the US to other places? This question has lately been the subject of much commentary globally and perhaps found its most scholarly expression in Alan Greenspan and Adrian Wooldridge’s book Capitalism in America in which they argued that the economic dynamism that set the US apart from other nations is now beginning to fade: “Look at any measure of creative destruction, from geographic mobility to company creation to tolerance of disruption, and you see that it is headed downward. The United States is becoming indistinguishable from other mature slow-growth economies such as Europe and Japan in its handling of creative destruction.”

The parallel rise of China has only brought this sentiment into sharper relief. I’ll skip the usual examples of the country’s growing relevance to frontier tech and reference one that doesn’t get nearly the amount of attention that it deserves: ResNet, short for “Deep Residual Learning for Image Recognition”, is a paper published by four researchers at the MSRA lab in Beijing in 2015 which outlines a method to train neural networks hundreds of layers deep; a seminal advancement in machine learning that paved the way for more powerful AI systems. The ResNet paper has in the span of a decade racked up over quarter of a million citations on Google Scholar. It is the most cited artificial intelligence paper ever. It is the most cited computer science paper ever. It is in fact the most cited paper in any academic field published in the 21^st century.

ResNet was authored by Kaiming He, Xiangyu Zhang, Shaoqing Ren and Jian Sun. All four got their undergraduate, graduate, and doctoral degrees at Chinese universities. None of them had worked outside China prior to publishing their landmark paper. They have all won awards at top global AI conferences and computer vision contests. In 2015, they won ImageNet, a major competition in artificial intelligence at Stanford which invites researchers to submit AI algorithms to identify objects in images. The MSRA submission was the first time that an AI system surpassed human-level performance at image-recognition on their dataset. ResNet was not a one-off. The last three editions of the ImageNet challenge were all won by teams that came from China.

The ResNet example ought to dispel any lingering doubts about China’s technical capabilities. The country doesn’t just imitate, it also innovates. Its competence is not limited merely to developing applications of technologies but extends to the underlying research breakthroughs that enable them. “China produces a lot of good quality research and some absolutely world-class research in artificial intelligence,” Geoffrey Hinton, the acclaimed British computer scientist and professor at the University of Toronto, told me in an interview. Hinton is the author of AlexNet, the paper that kicked off the modern revolution in AI. Ilya Sutskever, Hinton’s coauthor on that paper, would go on to found OpenAI. Hinton has won every major accolade in computer science, including the Turing Award, often called the Nobel Prize in computing, and the actual Nobel Prize, this time in physics. When I asked him which country was best positioned to take the lead in AI his answer was unequivocal: China.

That would appear to be a ringing endorsement of Greenspan and Wooldridge’s thesis from an unassailable source. Except the question of who’s winning the global tech competition is interesting precisely because it defies easy answers and divides opinion among the field’s loftiest authorities. It is ironic that Capitalism in America came out in 2018, the year when Apple became the first company ever to cross a trillion-dollar valuation. Since then seven other US tech companies have joined the club. The extent of America’s primacy in tech has grown to a scale that can be hard to grasp. The combined value of just the ten largest US tech companies is greater than the entire GDP of every single country in the world except the US itself. Just five American tech companies are worth more than every company listed on every stock exchange in all of mainland China combined. If that is what decline looks like, then what is ascendance?

John Hennessy, the Chairman of Alphabet, who previously served as the President of Stanford for 16 years, was nonplussed when I asked him whether he thought the Valley’s best years can now best be seen in the rear-view mirror. “It’s the killer app that has driven the Valley time and again,” he said. Small and medium-sized advances have always come from different places. But the Valley’s been the place that consistently throws up that big, hairy, epoch-defining technology that changes everything: semiconductors, personal computers, software, smartphones. And it doesn’t look like that’s about to change anytime soon. “It is the center of the next generation of AI companies,” he said. “There will be ones elsewhere, but certainly the Valley probably outnumbers anybody else by a factor of, I don’t know, two, three, four or five probably?”

As the world moves from the dot com era to the dot ai era, the US does appear to have an unassailable lead. But that doesn’t mean that China is not a plausible competitor. The commentariat has a poor track record of predicting the country’s technological trajectory. Just four years ago, in 2021, an article in Foreign Affairs argued that “China has worked for decades to develop a car brand that can rank among the world’s top automakers – but its car companies have had difficulty producing anything that consumers in developed countries actually want to buy.” With BYD now outselling Tesla and the entire German car industry in EVs, one does wonder whether there is another industry assessment out there that has aged quite so badly so quickly.

China winning the next platform shift might just be what it takes for it to tip the balance definitively in its favor. Seven out of ten of the highest profile researchers poached by Meta for its new Superintelligence Lab got their undergraduate degrees at just three institutions on mainland China. Is that an early sign of the shape of things to come? The status quo seems immutable until it doesn’t. Hennessy has been on the scene long enough to have experienced the last major shift. When he first moved to Palo Alto in 1977 all the major technology companies – IBM, Raytheon, DEC – were out in the east. “When you wanted to talk to the movers and shakers in the computer industry you got on a plane and you flew back to either New York or Boston, and that’s just the way it was,” he recalled. But change came fast and it didn’t take longer than a decade for the momentum to switch. “Quite frankly, I’m surprised at how once things began to swing how quickly it happened,” he said.

While it would be prudent to acknowledge the complexity of the US-China tech competition – not simply the story of the rise of one and the fall of the other – it is also I think necessary to look beyond it. The global tech race is often framed as a two-way contest with everyone else a mere also ran. And while it is true that there is unlikely to be a third competitor who can become relevant across the entire spectrum of new technologies, we do have hard to ignore countries and centers of excellence in specific domains coming up in various parts of the world. Asia’s dominance of the battery industry would be a case in point. All ten of the largest battery manufacturers are now Asian: six Chinese, three Korean, and one Japanese. In the time it took me to write my new book Northvolt went from promising upstart to Europe’s best funded startup ever to a collapsed has been. When I interviewed its founder Paolo Cerruti in 2020 he was upbeat that the company could own 25% of the continent’s market for batteries which would ensure that Europe “takes a strategic part of the value of the future car in its own hands.” The ambition struck me as realistic, even modest, given it was regional and not global in scope – which in my mind set it apart from companies in the Valley which sell far more fantastical visions. But the fact that it couldn’t deliver on what seemed to me to be circumscribed ambition shows the stranglehold that Asian players have on this corner of the tech industry.

Another reason to widen our peripheral vision when it comes to tech is that just because two countries dominate the landscape doesn’t mean that valuable lessons and life-changing opportunities can’t be found elsewhere. Let’s take an example from the venture industry. Earlybird was one of 107 new funds that came up in Germany during the first dot com boom in 90s. A quarter century later only three are still standing and Earlybird is the largest of them all. In its early years it opened an office in Silicon Valley to take on marquee names in the industry. The experiment failed. “We totally misjudged and underestimated the challenge of getting access to the best deals in the US in a very well-funded market,” Hendrik Brandis, the firm’s founder, told me. “You have to be pretty lucky to find extremely good opportunities because so many people have looked at it before you and you must see something that nobody has seen yet.”

The firm’s failed expansion to the US gave Hendrik the seed of an idea that would pave the way for future triumphs. It shifted its focus from the West to the East and opened an office in Istanbul to prospect opportunities in Eastern Europe. In 2015 the eastern fund invested in a small startup called DeskOver run by two entrepreneurs from a shabby building in Bucharest, Romania. The company would later change its name to UiPath and when it went public at the NYSE in 2021 it debuted at a valuation of $36 billion, the largest IPO for a European company ever at the time. Earlybird was the first outside investor in UiPath. Its modest million-dollar investment would net the fund upwards of $2 billion, perhaps the best early stage bet in European history.

Another interesting case study is Altos Ventures. It was founded in Palo Alto in 1996 by two Korean immigrants, Han Kim and Ho Nam, who met at Stanford GSB. The fund found it impossible to compete with the more established names and almost went out of business. Their struggles are the subject of two HBS case studies. Han told me that the company eventually found its footing by expanding out to regions overlooked by traditional funds where they worked with immigrant entrepreneurs who got little love from the bigger names in the industry. They also eventually entered the Korean market. Altos is today one of few funds that have managed to succeed on both sides of the Pacific. It was the largest shareholder in Roblox, a stake that was worth $8 billion at IPO, and an early investor in Coupang, Korea’s largest e-commerce platform, whose market cap tops $50 billion.

But perhaps the most important reason for tuning into other geographies is that at some point our collective beliefs become self-fulfilling prophecies. If everyone thinks that tech is a US-China duopoly then that’s what it’s going to be. Alex has previously written about how we’ve gone from vibe coding to vibe investing – AI companies raising vast fortunes without any product or customers or financials to speak of. Perception drives reality. While we have strong opinions when picking sides between the US and China; one of the few things that probably everyone can agree on is that it can only be a good thing for more places join the mix. The more we move the spotlight to other regions the more likely it is that they will attract the resources they need to at least show up to the starting line. Talk is cheap and that’s an excellent reason to have more of it.