Quantum Computing Will Change the World. How to Play the Stocks.
Investors finally have ways to play the first radical shift in computing since the 1950s, but you're better off waiting before jumping in.
Last month, the Royal Swedish Academy of Sciences awarded the Nobel Prize in physics to three scientists for their research on quantum information science, one of the most surprising and—as it turns out—most commercially alluring scientific discoveries of the past 100 years.
Yes, the quantum computing revolution is coming. But it will take time, probably longer than is comfortable for investors in the handful of quantum start-ups that have managed to reach the public market. This is a textbook case of Amara’s Law, coined by the futurist and engineer Roy Amara: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
The new crop of small, public companies devoted to quantum—such as Rigetti Computing, D-Wave Quantum, and IonQ—will be challenged to generate significant revenue for years to come.
“We want to solve problems that are intractable for today’s—or even next century’s—supercomputers,” says Krysta Svore, vice president of quantum software at Microsoft (ticker: MSFT). With conventional supercomputers, she says, there are unsolvable problems—like sorting through potential drug candidates—that would require compute times longer than the current lifespan of the universe. “We want to bring this technology forward and see how we can use it in conjunction with classical technology.”
The nature of quantum computing makes it useful for solving computationally intensive problems with huge numbers of variables. Quantum computing will potentially speed up drug development, improve financial modeling, and boost the efficiency of electric batteries.
The Nobelists’ work demonstrated a mind-bending concept crucial to quantum computing called “quantum entanglement.” The three prizewinning physicists—Alain Aspect, John Clauser, and Anton Zeilinger—all contributed to the discovery that particles in an entangled state can affect other particles, even when vast distances apart.
The new research undermines the thinking of none other than Albert Einstein. Einstein was skeptical of the ability of quantum mechanics to describe the universe in full; in particular, he was uncomfortable with quantum theory’s reliance on what he called “spukhafte fernwirkungen”—“spooky action at a distance.” It turns out that Einstein’s doubts were misplaced.
Though the exact mechanism remains uncertain—“nobody understands quantum mechanics,” the physicist Richard Feynman once said—experiments from the new Nobelists prove that quantum theory really does describe the natural world and that entanglement exists. That discovery has set the stage for an entirely new branch of computing, and there’s a race under way to develop the first commercial quantum computers, with potentially vast riches at stake.
The combatants include some of the biggest players in “classical” computing: Microsoft, Intel (INTC), Alphabet (GOOGL), Amazon.com (AMZN), and IBM (IBM) are all building quantum hardware, along with Japan’s Toshiba, NEC, and NTT, and China’s Baidu (BIDU), Huawei Technologies, Tencent, and Alibaba. At the other end of the scale are a handful of small firms that rode quantum hype into the public markets, mostly through special purpose acquisition company, or SPAC, mergers, including Rigetti Computing (RGTI), D-Wave Quantum (QBTS), and IonQ (IONQ). And that’s just the tip of the iceberg: According to PitchBook, 251 quantum start-ups have together raised more than $5.4 billion in venture capital since the beginning of 2017.
It’s easy to see the allure.
IDC last year estimated that the market for quantum computing services, mostly delivered by the cloud, could grow to $8.6 billion in 2027, up from $412 million in 2020, a compound annual growth rate of more than 50%.
More tantalizing is a 2021 report from Boston Consulting Group that put the potential value creation from quantum computing at $450 billion to $850 billion—with $90 billion to $170 billion of that flowing to the quantum industry players. But investors will have to be patient—Boston Consulting Group doesn’t expect the industry to reach that scale until 2040 or later.
William Zeng, head of quantum research at Goldman Sachs, is fascinated by quantum’s game-changing potential. “We are very much in research mode,” Zeng says. “We do not have systems in production yet. We’re figuring out how to get there. It starts with looking at business problems—what things now are too slow, too expensive, or can’t be solved at all. And then you try to pair those with places where quantum has a theoretical advantage.”
One early area of focus for Goldman is the potential to speed up Monte Carlo calculations, complex algorithms used to assess the value and risks of derivatives and other securities. He sees other potential applications in portfolio optimization and machine learning for anti-money-laundering, among other things. But not yet, and not particularly soon.
Tony Uttley, president and COO of Quantinuum, left. He says the technology at the heart of Quantinuum’s “trapped ion” quantum computers, right, could help the company reach the market fastest. PHOTOGRAPHS BY THEO STROOMER
But you’ll need a quantum computer to predict the long-run winners.
“We’re not at the stage where a quantum computer is improving the bottom line of any company not in the field of quantum computing,” says Ryan Babbush, head of quantum algorithm and applications at Alphabet‘s Google unit.
That said, there are real-world examples of quantum computers being used today. IBM arguably has the early lead. Big Blue has built more than 30 quantum computers since 2016, and more than 20 of them are online right now, accessible via the web. IBM’s director of research, Dario Gil, says there are more than 500,000 users for those IBM quantum systems at over 180 institutions, primarily for research. “We have more quantum computers online than the rest of the world combined,” Gil says.
“It is still super-early,” agrees Simone Severini, the director of quantum computing at Amazon Web Services. “There is still substantial scientific and engineering work to do before we get quantum computing at scale.” Like other quantum industry leaders, Severini says there is growing interest from customers that want to explore the technology, but he says that it’s too early to know which technological approaches will succeed.
IBM CEO Arvind Krishna thinks we’re still probably five years away from anyone generating material revenue from quantum computing, but he adds that “there are a lot of very smart people with a lot of capital chasing the space.” He thinks that at some point—maybe in two years, or three, or five—“you will solve problems that will just astonish people.”
His back-of-the-envelope math suggests there could be a $100 billion market for quantum computing by the end of the decade, which he thinks will be split by a handful of players.
To be clear, quantum computers are never going to replace conventional computing. You’ll never use one to check your email, play games, or run Excel, and there will be no quantum smartphones or laptops. Instead, quantum systems will work in tandem with conventional computing to solve problems that can’t be addressed with current technology.
“This is the first time that computing is branching,” says IBM’s Gil, describing the radical shift away from the computing architecture pioneered by Intel, whose co-founder Gordon Moore accurately forecast in 1965 that the power of microchips would double roughly every two years.
Advertisement: Scroll to Continue
“Some people say this is another step in Moore’s Law,” Gil says, “but it is more fundamental than that. It is not that often that we get to redefine the nature of information.”
It has now been 77 years since ENIAC, the first programmable digital computer, launched at the University of Pennsylvania. Originally intended to be used for artillery targeting, ENIAC played a role in the early development of nuclear weapons. Two years later, in 1947, Bell Labs unveiled the first transistor, the basic building block for all modern electronics.
Over the decades, every computer ever built—PCs, mainframes, supercomputers, game consoles, and mobile phones—have relied on transistor-based binary computation with “bits,” the smallest possible pieces of data, which can only exist in two states. On or off. Zero or one. True or false. Yes or no.
In quantum computing, the most basic piece of information is the qubit—a quantum bit. Like a classical computing bit, the qubit also has two potential states—on or off, zero or one. But quantum computing allows for qubits to be in a state known as “superposition,” in which they are zero and one at the same time—or more precisely some statistical probability of being either. The physicist Erwin Schrödinger whimsically illustrated this paradoxical idea by describing a hypothetical situation in which a cat could be simultaneously both alive and dead.
More important than understanding the deep weirdness of quantum mechanics is the fact that entanglement and superposition give quantum computers phenomenal computational power.
In conventional computing, computing power grows linearly with an increase in bits. But with entangled qubits, computing power grows exponentially as you add more qubits. With three entangled qubits, you can get eight simultaneous calculations—add a fourth qubit, and the system can do 16 calculations in parallel. As the number of qubits increases, eventually you get systems that can’t be matched by conventional computing. To match the computing power of a system with 100 qubits, you’d need the equivalent of 10 trillion years of classical computing time. Classical computing solves problems by considering each potential solution sequentially; quantum computing evaluates all possible solutions simultaneously.
Consulting firm McKinsey estimates that quantum computing has the potential to “revolutionize” research and development on molecular structures in biopharmaceuticals, speeding up drug discovery and development. In the chemicals industry, McKinsey says quantum should drive speedier development of new catalysts, with implications in areas like carbon capture and energy efficiency. Auto makers such as BMW and Volkswagen have begun research on the application of quantum computing to supply-chain management, traffic routing, and electric-vehicle battery design. There are implications for financial services, particularly in portfolio management, risk analysis, machine learning, and options pricing.
Advertisement: Scroll to Continue
Yet, what could prove miraculous in finance, chemistry, and drug discovery also threatens to undermine all current versions of cybersecurity. The Global Risk Institute wrote in a 2021 report that “the threat posed by quantum computers could lead to a catastrophic failure of cybersystems, both through direct attacks and by disrupting trust.”
The threat is still viewed as far out—a survey of 47 quantum experts found that a little more than half saw a better than 50% chance that quantum computers will be able to break an encryption key of 2,048 numbers in under 24 hours within 15 years.
A growing worry is that a U.S. adversary—China being the most likely—could beat the U.S. to the punch, effectively putting any information protected with traditional forms of cryptography at risk of discovery—think industrial secrets, financial information, personal data, everything. The cybersecurity industry is racing to create “quantum safe” alternatives to traditional approaches to data security.
The process of creating quantum computing is taking years longer than some people might have expected, and for good reason. It turns out that building stable and useful quantum systems is fiendishly difficult. Qubits are fragile, sensitive to changes in temperature, materials impurities, radiation, vibration, and other environmental conditions. That makes them prone to high error rates.
In one room at IBM’s research center in Yorktown Heights, N.Y., there’s a quantum system sitting alone, connected to the web for the benefit of researchers experimenting with coding quantum systems. Down the hall, more systems are under construction using chips with higher qubit counts.
Earlier this month, IBM announced a 433-qubit processor, more than triple the number of qubits in the company’s last-generation quantum chip; the company is targeting the 4,000 qubit level by 2025. A few other players have made similar promises.
On the outside, IBM’s quantum systems are just imposing boxes. Inside, they look far different than a traditional computing system, mostly because quantum computers require supercooling to keep the systems running. The cooling systems to reach the required ultralow temperatures result in the distinctive candelabra-shaped designs for systems created by IBM, Rigetti, and others.
By some estimates, it could take thousands or even millions of physical qubits to create a viable quantum computer. “You need enormous redundancy,” says Amazon’s Severini. Tim Costa, director of high-performance and quantum computing at Nvidia (NVDA), says that while there are a variety of quantum computers accessible today via the large cloud vendors, none have more than a few hundred qubits; to do useful work, he says, will require systems with millions of qubits.
Most of the larger aspirants—including IBM, Alphabet, Amazon, and Alibaba—are pursuing a similar approach, using superconducting qubits, their chips cooled to extremely low temperatures, a few microkelvin, colder than the vacuum of outer space and controlled with microwaves.
Quantinuum, a spinout from Honeywell International (HON), and newly public IonQ, rely on “trapped ion” systems, using naturally occurring atoms as qubits; trapped ion systems are considered to be more reliable, but slower, than superconducting quantum computers. Quantinuum President Tony Uttley says Honeywell decided early on that the trapped ion approach would give the company a chance to reach the market sooner. “Our thesis was that if you make really high-quality qubits, you can do more with those systems in early stages of quantum computing,” he says. “We believe in a future with multiple kinds of quantum processors.”
IonQ’s CEO, Pete Chapman, says the company has systems running on all three of the leading cloud platforms—AWS, Azure, and Google Cloud—and contends that the company will have commercial applications running for clients by the end of 2023. “We have a shot at being the first one there,” he says. “We should have the market to ourselves for the next few years.”
Other companies are taking different approaches. Start-ups PsiQuantum and Xanadu use photons controlled by mirrors and other devices. ColdQuanta and Atom Computing rely on neutral, or cold, atoms—rather than the charged atoms, or ions, used by other technologies. Intel is working on “quantum dot” technology and says that its expertise in chip production can be applied to building quantum processing units, or QPUs. Microsoft is betting on a technology called “topological qubits,” which, in theory, have fewer errors than other approaches but which remain in the early research stage, with no functioning systems to date.
Advertisement: Scroll to Continue
D-Wave, which recently went public via a merger with a SPAC, offers an approach called “quantum annealing” that is targeted specifically at solving optimization problems. D-Wave CEO Alan Baratz, a longtime tech executive who in the late 1990s built the Java software business at Sun Microsystems, contends that his company is “the only commercial quantum computing company, working with real companies on real applications.”
For all the promise, today’s quantum revenue is virtually nonexistent. D-Wave’s June-quarter revenue, for instance, was just $1.4 million. For the full year, the company sees revenue of $7 million to $9 million.
The clean room at a Rigetti Computing lab. PHOTOGRAPH BY WINNI WINTERMEYER
Despite the lack of current commercial applications, companies have begun to put timelines on major breakthroughs. IBM intends to build a 1,000 qubit system by the end of 2023—and 4,000 by 2025—and sees commercial workloads evolving before the end of 2025. PsiQuantum has vowed to develop a one-million qubit computer as soon as 2025. And it may take systems at that scale to reach quantum’s potential. James Clarke, director of quantum hardware at Intel, says it will take at least one million qubits “to do something earth shattering.”
Rigetti founder Chad Rigetti, who announced his resignation as the company’s CEO earlier this month, said in a recent interview that his company has the fastest quantum computers in the world, and next year will launch a system with 336 qubits, “opening the window where quantum advantage really becomes possible.”
The real question is when quantum computing will generate meaningful revenue. “If I were to draw a graph of probability of quantum computations really assisting businesses in making decisions—it would peak in three to four years—and then again in 10 years or so,” says Alphabet’s Babbush. He sees some potential for the earliest systems to find niche applications, including in financial services, in the next few years.
Richard Moulds, who runs Amazon Braket, an AWS-hosted quantum computing research service, says the cloud-based computing giant’s customers “expect us to be a guiding hand and to be ready with commercial quantum infrastructure.” But, he adds, so far the primary use of the online quantum services is to build better quantum computers.
“No one is using this in a production sense,” Moulds says. “We’re still discovering which applications are likely to be the most useful. We’re getting ready for quantum.”
IBM’s Gil views quantum computing as a high-stakes game that few players will survive. To build quantum hardware, he says, will take “stamina, capital, and know-how.” He suspects that there will be more pretenders than successes. “I don’t think there will be many players at the end; you’ll be able to count them on one hand. In the U.S., maybe two or three. It’s not for the faint of heart. On a scale of one to 10 for technical difficulty, it is a 10.”
After a short period of hype, investors have begun to recognize the long lead times. Small-cap quantum stocks have been terrible performers this year, particularly as investors flee high-risk assets. Rigetti is down 88%, IonQ is off 71%, and D-Wave has lost 64%. The Defiance Quantum exchange-traded fund (QTUM) is down a more modest 27%, but that’s because the portfolio includes not only pure plays like IonQ but also stocks such as Taiwan Semiconductor Manufacturing, Baidu, Microsoft, and Texas Instruments.
Meanwhile, the big tech players in quantum—Alphabet, Microsoft, and Amazon.com—are all dealing with broader business issues that have weighed on their stocks.
IBM is perhaps the only quantum player currently operating from a position of strength. Its stock is up 10% this year.
IBM’s Gil thinks that quantum will be worth the work, and the wait. “The future of computing is bits plus neurons and qubits”—a combination of conventional computing, quantum, and artificial intelligence. “That is how computing is going to run.