The subatomic age: Asia's quantum computing arms race

SINGAPORE/TOKYO -- Dzmitry Matsukevich's laboratory is cramped and windowless, dominated by two large tables laid out with arrays of lasers, lenses, cables and black boxes arranged like some impenetrably complex board game. It is an experimentalist's space -- practical, prioritizing form over function. Plastic sheeting hangs in curtains from the ceiling; paperwork, boxes of components and rolls of electrical tape are piled on every spare surface.

Matsukevich, a principal investigator at Singapore's Center for Quantum Technologies, is a physicist out of central casting: mop-haired, earnest, and infectiously enthusiastic about two slightly blurred gray squares on a computer monitor. The squares are trapped ions, held in near total stillness by lasers: the foundational element of a quantum computer.

"If you haven't seen a single atom before," Matsukevich said, "now's your chance."

Matsukevich has spent decades working on quantum information in academic institutions around the world. As recently as 10 years ago, the science had become moribund. Early breakthroughs had not been followed with any practical advances, and the reality of a functional quantum computer seemed a long way off.

"If you wrote on your research proposal that you are going to do quantum computing, most of the universities would just throw your application away," Matsukevich said. "They thought it was a dead end."

  © Illustration by Hiroko Oshima

For all of its potential, quantum computing has always felt just out of reach. Although it has attracted various waves of investment and interest over the last 30 years, scientists struggled to scale the technology beyond academically interesting, but largely nonfunctional devices, incapable of matching mainstream classical computers.

In October last year, that changed. U.S. technology giant Google announced that it had achieved "quantum supremacy" -- it had built a quantum computer that performed, in just under four minutes, a series of calculations that it claims would have taken a conventional supercomputer 10,000 years.

Technology giants from Hitachi to Huawei are buying up talent from university departments. Venture capital is hunting promising startups, alive to the commercial possibilities of a technology that could rewire the IT sector. At the same time, governments are throwing billions of dollars into quantum computing initiatives, starting a high-tech arms race that has drawn parallels with the mad rush into artificial intelligence over the last decade.

The prize is enormous. A stable, functioning quantum computer could make power grids super-efficient, trivialize complex financial risk management, shave years off the process of pharmaceutical development and power a revolution in machine learning. It could crack unbreakable encryption, give militaries unbelievable capabilities in the field and leave vast amounts of personal data open to exploitation and manipulation. In theory, it could change the world.

"I think right now we are still underestimating its potential," said Norishige Morimoto, vice president of research and development at IBM Japan. "An analogy is to compare it to the abacus. ... I think it's almost that level of a jump. Abacus vs. classical computers. That's what today's computers vs. quantum computers could be."

Quantum supremacy

The idea of the quantum computer is popularly -- but probably erroneously -- ascribed to Richard Feynman, an infamous lockpicking, womanizing, bongo-playing physicist who helped develop the atomic bomb.

Feynman was a leading figure in the development of quantum mechanics, the mathematics that describe the interactions between matter and energy at a subatomic scale. Those interactions are intuitively complex -- the distinction between matter and energy is blurred, objects can exist simultaneously in different states, and the act of observation can change those states. To simulate these systems, Feynman said in a 1982 lecture, would need a computer built on the same lines.

By the early 1990s, physicists had figured out how that computer might work. Classical computers store and transmit information in bits -- binary states of 0 or 1. Anything that can have those two states can, in theory, be used for computation; 20th century computers used vacuum tubes, magnetic cores and transistors to encode those bits. Thanks to the laws of quantum mechanics, a quantum bit, or "qubit," can occupy both states, 1 and 0, at the same time.

Computation using these qubits would be much more complex, but potentially much more powerful than a digital device. A computer with 25 bits can encode 25 pieces of information. A quantum computer with 25 qubits could encode 2^25 bits of information -- more than 33.5 million. A quantum computer with 500 qubits would be able to encode more bits of information than there are atoms in the observable universe.

"Our devices are constrained in the way that the mechanisms inside them move. If we could build a device that's unconstrained, perhaps we could do different kinds of computations," said Chris Ferrie, associate professor at the Center for Quantum Software and Information at the University of Technology Sydney and author of a series of children's books, including Quantum Entanglement for Babies. "The question is, is there something in the world in which you can encode these computations?"

That meant complex algorithms, and physical computers on which to run them. The search for hardware has thrown up some exotic solutions, from Matsukevich's approach of trapping ions with laser beams, to electrons frozen in synthetic diamonds, to loops of superconducting wire cooled to below -270 C. The computers are often physically enormous to accommodate the electromagnetic shielding and liquid helium needed to create qubits that flash into use for a few fragile moments.

As those approaches yielded results the technology advanced, qubit by qubit, although some of the most significant challenges remain unsolved -- in particular that of correcting the errors and noise that are inevitable with such a complex and sensitive type of computer. In the last few years, however, progress has accelerated, lubricated by a sudden rush of money into corporate programs, which have dwarfed the resources of small, grant-funded laboratories that made up most of the field.

"I think the real catalyst was money that was put towards a single effort, rather than competing efforts," Ferrie said. "You're not going to land on the moon with five people."

Today's quantum computers are still not game-changers. IBM's Morimoto compares the state of the technology to where classical computers were at in the 1940s -- unwieldy, inaccurate machines based on banks of vacuum tubes. "The system engineer at that time would have no idea that this machine would play movies or music, or even connect you to different countries," he said, in an interview in IBM's Tokyo headquarters.

  © Illustration by Hiroko Oshima

IBM has doubled the computational power of its quantum computers annually for the past four years, Morimoto said, and is confident that it can keep up the pace. Although he did not want to be drawn on specific timelines for an error-free, universal quantum computer, he insisted that real change is imminent.

"In five to ten years ... I'm pretty confident that we will reach the point where the usable applications, not just mathematical puzzles, are going to exceed things that a classical computer could ever do," he said.

Because of the way that quantum computation works, even relatively low-powered quantum computers should be able to solve certain types of problems far more rapidly than classical machines. These include multidimensional analyses, such as machine learning on enormous and complex data sets; differential equations, such as those used in derivative pricing and financial risk analysis; factoring of prime numbers, used in cracking cryptography; and the analysis and design of complicated molecules, which would be valuable to the materials science and pharmaceutical fields. The COVID-19 pandemic has demonstrated the value of computational approaches in the pharmaceutical industry, with classical supercomputers around the world diverted to the search for drugs that might act against the virus.

This could result in profound disruptions for some industries. In the banking sector, a working quantum computer could perform Monte Carlo simulations -- commonly used in risk analysis -- to greater depth and at much greater speed than a classical computer. They could crunch massive data sets to allow for real time portfolio optimization.

"People are competing in milliseconds on machine trading. That is based on your knowledge of being able to see things two minutes out. How about if you can predict two days out?" Morimoto said.

"People are competing in milliseconds on machine trading. That is based on your knowledge of being able to see things two minutes out. How about if you can predict two days out?" said Norishige Morimoto, vice president of research and development at IBM Japan. (Photo by Peter Guest)

In the materials sector, it could change the balance of power in an industry that Japan has long dominated.

"In materials, long-term accumulation of knowledge and experience is the key, that's why it's so difficult for other countries to catch up," Morimoto said. "But if you can invent a new material overnight with this new machine, without all that base knowledge and accumulated experience, data, experiments. Just put it in the simulator. That will dramatically reduce the advantage that today's materials companies have. ... That's kind of a horror story."

Tellingly, the first industry partners for a quantum computing partnership between IBM and Tokyo's Keio University were two banks and two chemicals companies -- Mizuho Bank, MUFG Bank, Mitsubishi Chemical and JSR.

This year, Morimoto will oversee the delivery of two quantum computers to Japan, the first time that the company has moved its hardware outside of its New York State headquarters. One will be placed at Tokyo University, and the other at a secure IBM site. The company is trying to recruit industry partners, inviting them to familiarize themselves with the technology before it accelerates away from them.

The initiative has been boosted by the Japanese government, which has thrown its weight behind a national quantum strategy, including 30 billion yen in the 2020 budget to support research and development in the field, and a 100 billion yen in funding for a "moonshot" initiative. The government said it wants to develop a 100-qubit quantum computer -- around double the power of IBM and Google's current machines -- within a decade.

  © Illustration by Hiroko Oshima

It is not alone. In 2018 the European Union launched a three-year, 132 million euro program to support the development of strategic quantum technologies.

Probably the largest and most ambitious program is in China, which elevated quantum computing to the status of a strategic technology around 2017, according to people with knowledge of the initiative. China has had advanced programs in other forms of quantum technology -- particularly communication and cryptography -- since the 1980s, but is now devoting resources to catching up in computation. Specific numbers are hard to come by, but several people familiar with the government's approach told Nikkei that the total investment -- in the form of grants and other incentives -- totals billions of dollars. A new quantum research center is under construction, and leading universities have started or expanded their programs with government funding.

That, in turn, has prompted a reaction in the U.S. The American government and defense establishment have supported quantum computing for years, but in February 2020, the White House's proposed budget included a doubling of national support for quantum technologies to more than $200 million a year, with millions more going into specific initiatives by the departments of defense and energy. The sudden surge of interest in the defense establishment was, according to Elsa Kania, adjunct senior fellow with the technology and national security program at think tank Center for a New American Security, came from a realization that, while the U.S. lead in many technologies such as semiconductors may seem insurmountable, that may not be the case in quantum computing.

"These new frontier technologies like ... quantum information and computing present unique, perhaps historic opportunities for China to be more of a first mover, or at least to be neck-and-neck," Kania said.

Strategic tech

Quantum physicist Yung Man-Hong spent nearly 20 years in academia in the U.S. and China, focusing on the science of quantum information. Like others in the field who spoke to Nikkei, he had started to worry that the pace of development was just too slow.

"Five years ago I told my students, if you don't work hard, you're going to be unemployed," he said. "Sometimes people say it may still take another 20 years to have a quantum computer, but then I say, I cannot wait another 20 years. That would be 40 years -- my whole career -- without any practical outcome."

Just over two years ago, while Yung was a professor at the Southern University of Science and Technology in Shenzhen, Huawei came calling. He jumped at the chance to get access to the company's vast resources. "This was a real moment I've been expecting for 20 years," he said.

Other Chinese tech giants Baidu and Tencent began to invest in the science in earnest at around the same time. Alibaba was much earlier, according to people with knowledge of the company, starting around 2015.

Yung's role, as chief quantum computing software and algorithm scientist at Huawei, is to develop quantum software, including the programming languages and libraries that will allow others to build applications. In late 2018, he oversaw the release of Hi-Q, a publicly-accessible quantum computing simulator and programming framework. Other colleagues are working on hardware, including quantum chips, he said.

Huawei's interest in the technology is not based on any short-term expectations, according to Yung. "I'm not being asked to make revenue, or [any] kind of business plan. My role is to make this thing happen, or to help be part of the development," he said. He declined to put a time scale on the company's expected results, but said that the company was "patient."

  © Illustration by Hiroko Oshima

For now, much of Yung's effort is directed into hiring and training staff. China's universities now pump out more than twice as many graduates per year than U.S. institutions, with a significant proportion of those in practical disciplines, such as engineering and computer science. That production line of technical specialists has fed the country's swollen technology sector, but quantum computing requires a fundamentally different type of skills and intuition.

"Companies like IBM, Google, they [are] able to get a lot of very experienced people to join. In China, we still need to train more people to become experts in quantum computing. This is the reality," he said. "We cannot do much about it. We just need to keep training people."

This skills shortage is likely to hamper Chinese companies' attempts to catch up with their U.S. competitors. The U.S., Japan and European countries have invested for decades into fundamental academic research in universities, building a depth of knowledge and experience that cannot be bought off the shelf.

"They could put a lot of money in -- and once the government decides that they are going to do this, they would probably put a lot more money in than any other country," said Mark Greeven, professor of innovation and strategy at IMD Business School, and a research associate at China's National Institute for Innovation Management. "But that's probably not enough."

Greeven, who has worked extensively within China's technology sector, said that while the government has clearly acknowledged the strategic importance of quantum computing -- just as it did with artificial intelligence last decade -- it is not clear how it will actually support the development of the technology. "They are not going to support something that they do not feel comfortable and confident about that they can actually win in it," he said.

When the Chinese government decided that artificial intelligence would be a strategic focus, it was able to seed the industry by becoming an anchor client for startups, and by providing vast quantities of data. "Data [is] something in which China naturally has an advantage, and the U.S. does not," Greeven said. "You cannot say the same about quantum computing."

Quantum computing is a complex field, and one with high barriers to entry. It is highly unlikely that thousands of startups will emerge, and development is likely to remain concentrated at a few large companies.

It is also likely that Chinese companies will be denied access to research and technology from the U.S., as part of the White House's broad spectrum crackdown on China's tech companies.

Speaking anonymously to avoid jeopardizing their future funding opportunities, several academics at international institutions in Asia and Europe told Nikkei that there is already a suspicion around researchers with links back to China. One said that it was "inevitable" that individual scientists, research groups and even universities would eventually be forced to choose between working exclusively with Chinese companies, or avoiding them entirely.

Quantum winter

Some degree of Balkanization in the field was inevitable. As the technology edges towards commercial viability, companies were always going to hold their research closer to their chests. And as the center of gravity shifts towards those big companies, venture capitalists are sniffing at the fringes of the field, looking for the startups that will eventually be acquired by the industry's big guns.

"In quantum, we always think: Who could be the potential buyer?" Fuyuki Yamaguchi, managing partner at Tokyo-based Abies Ventures, which has invested in several quantum-related startups, and is looking for more. U.S. companies are likely to vacuum up startups in order to acquire talent and software, but Japan's tech giants are also likely to join the hunt.

"In Japan, NTT or NEC, Fujitsu, or Toshiba, they focus on leading technology," Yamaguchi said. "They might be good buyers."

University researchers are hurrying to create spinoffs, eager to profit from their research.

After an academic career at top universities in Ireland, the U.K. and Singapore, Joe Fitzsimons spun off Horizon Quantum Computing, a quantum software company, from CQT in 2018. The tipping point was on the horizon, he said, and he wanted to be established by the time one of the big tech companies announced that they had achieved quantum supremacy.

"Everyone in the field has known this was coming," he said. "We thought: Once this result is announced, there's going to be this rush of people looking to start companies. ... I think the day following the Google announcement, I got contacted by three VCs."

Fitzsimons called it right. Over the past 12 months there has been a proliferation of startups in the quantum space, many headed by prominent researchers. It is a strange time for academics, who have been used to collaborating freely and internationally, and scraping together grant money to fund their experiments.

The promise of money is seductive. As Fitzsimons said: "There are a number of different things that can give you the advantage. For sure, access to capital is one of them. Give me $100 billion in the morning, I reckon I could have a fair stab at making a quantum computer."

But at the same time, the sudden rush of interest has made many people in the field nervous. Artificial intelligence went through several waves of excitement through the 80s, 90s and 2000s. Each time capital flooded in, then washed back out as it became clear the technology could not deliver on the hype.

The sudden interest has made some nervous. "There is a lot of hype in the field. ... We have to be careful about what we promise," said Dimitrios Angelakis, principal investigator at the Center for Quantum Technologies in Singapore. (Photo by Peter Guest)

For the dozen startups that Nikkei spoke to, the fear of a "quantum winter" was a common thread.

"There is a lot of hype in the field. Being 25 years in the field, I don't want it to die," said Dimitrios Angelakis, another principal investigator at CQT, who has recently established a consulting business on the side. "We have to be careful about what we promise. The supremacy is proof we can fly, in the sense that we took off for 200 seconds, then we crashed."

There might be real implications for society if the current quantum computing wave does collapse under the weight of expectation.

At IBM, Norishige Morimoto said that the world is already coming up against the practical limits of current technology. The vast amount of data being produced in the 21st century requires a doubling of computing power every year, something that is simply too expensive and too energy-intensive to sustain.

Morimoto also, unbidden, spoke lyrically about the potential of a technology that would allow scientists to see the world in a fundamentally different way.

"Like tiny little birds, they fly over dark nights for thousands of miles, from one country to others. How do they keep track of the directions in the dark, or flying over the ocean? Where is the GPS that these little guys have?" he said. "There are some [theories] that some quantum physics happen somewhere in those little creatures' brains. But we cannot see it or prove it, so we think it's some kind of miracle. ... There are many things that today we cannot explain, and we think are magical."