News in focus virologists and researchers with expertise in samples from March 2019 and found SARS- The investigation should also prioritize public health, animal health and food safety CoV-2 fragments, says Raina MacIntyre, an carnivorous mammals farmed for fur, such will lead the WHO’s COVID-19 investigation. epidemiologist at the University of New South as raccoon dogs and civets, which had a role The agency has not released their names. Wales in Sydney, Australia. “If this study was in the SARS outbreak, says Martin Beer, a The team held its first virtual meeting, correct, we have to ask how the virus was in virologist at the Federal Research Institute for including researchers in China, on 30 October, Spain in March last year,” she says. Animal Health in Riems, Germany. “It is surpris- and is reviewing the preliminary evidence and Plans to look beyond China are sensible, ing that there is no mention of these animals in developing study protocols, says the WHO. given that extensive surveillance in bats in the report, and we have no information from The initial phase of investigations in Wuhan China since the 2002 SARS outbreak has iden- China about whether these animals have been will probably be conducted by researchers tified only a distant relative of SARS-CoV-2, tested,” says Beer. who are already in China, and international says Wang. A growing number of experts think A spokesperson for the WHO says the researchers will travel to the country after that the immediate or close ancestors of SARS- mission will be guided by science, and “will reviewing those results, the agency says. CoV-2 are more likely to exist in bats outside be open-minded, iterative, not excluding any In Wuhan, researchers will take a closer China, says Wang. He says the WHO team hypothesis that could contribute to gener- look at the Huanan meat and animal market, should survey bats and other wildlife across ating evidence and narrowing the focus of which many of the earliest people diagnosed southeast Asia for SARS-CoV-2 antibodies. research”. with COVID-19 had visited. What part the market played in the virus’s spread remains a mystery. Early investigations sampled frozen animal carcasses at the market, but none found evidence of SARS-CoV-2, accord- ing to a 5 November report on the WHO UNDERDOG TECH MAKES mission’s terms of reference (see go.nature. com/2uiz8ik). However, environmental sam- GAINS IN QUANTUM ples, taken mostly from drains and sewage, did test positive for the virus. “Preliminary studies COMPUTER RACE have not generated credible leads to narrow the area of research,” the report states. Trapped-ion technologies are gaining momentum in The WHO mission will investigate the wild and farmed animals sold at the market, includ- the quest to make a commercial quantum computer. ing foxes, raccoons (Procyon lotor) and sika deer (Cervus nippon). They will also investigate By Elizabeth Gibney month, University of Maryland spin-off firm other markets in Wuhan, and trace the animals’ IonQ in College Park announced a trapped- journeys through China and across borders. technology for building quantum ion machine that could prove to be compet- The investigators will prioritize animals that computers that has long been side- itive with those of IBM or Google, although are known to be susceptible to the virus, such lined by commercial developers is the company has yet to publish details of its as cats and mink. gaining momentum. As quantum performance. Smaller spin-off firms — such The team will also look at Wuhan’s hos- computing has transformed from as Universal Quantum in Brighton, UK, and pital records, to find out whether the virus Aacademic exercise to big business over the Alpine Quantum Technology in Innsbruck, was spreading before December 2019. The past decade, the spotlight has mostly been Austria — are also attracting investment for researchers will interview the first people iden- on one approach — the tiny superconducting trapped-ion projects. tified to have had COVID-19, to find out where loops embraced by technology giants such as Trapped-ion quantum computers, which they might have been exposed, and will test IBM and Intel. Superconductors last year ena- store information in the energy levels of indi- blood samples collected from medical staff, bled Google to claim it had achieved ‘quantum vidual charged atoms held in an electric field, laboratory technicians and farm workers in the advantage’ with a machine that for the first are far from new: they were the basis of the weeks and months before December, looking time performed a particular calculation that qubits in the first basic quantum circuit in 1995, for antibodies against SARS-CoV-2. The report is beyond the practical capabilities of the best long before anyone used superconducting acknowledges that some of this work might classical computer. But a separate approach, loops (C. Monroe et al. Phys. Rev. Lett. 75, 4714; already be under way in China. using ions trapped in electric fields, is gain- 1995). But efforts to put all the building blocks ing traction in the quest to make a commercial together to build viable commercial systems Longer-term plans quantum computer. are “sort of bursting on the scene now”, says The initial investigation in Wuhan will inform Earlier this year, technology and manufac- Daniel Slichter, a quantum physicist at the US longer-term studies into the pandemic’s turing company Honeywell launched its first National Institute of Standards and Technology origins, which could take investigators quantum computer that uses trapped ions in Boulder, Colorado. outside China. “Where an epidemic is first as the basis of its quantum bits, or ‘qubits’, detected does not necessarily reflect where which it had been working on quietly for more Rising challenger it started,” the WHO report states, noting than a decade. Honeywell, headquartered in “I think nowadays people say ‘superconduc- preliminary reports of viral RNA detected in Charlotte, North Carolina, is the first estab- tors’ and ‘trapped ions’ in the same breath, sewage samples before the first cases had been lished company to take this route, and it has and they weren’t saying that even five years identified. a 130-strong team working on the project. In ago,” says Chris Monroe, a physicist at the This statement could refer to a study, October, seven months after the launch, the University of Maryland who worked on the posted on the preprint server medRxiv with- firm unveiled an upgraded machine; it already 1995 experiment and is a co-founder of IonQ. out peer review (G. Chavarria-Miró et al. Pre- has plans to scale this up. is still in its infancy, and print at medRxiv https://doi.org/fhw5; 2020), And Honeywell is not the only company plan- although various companies are jockeying to which retrospectively tested Spanish sewage ning to make trapped-ion systems at scale. Last claim that their quantum computer is the most

342 | Nature | Vol 587 | 19 November 2020 ©2020 Spri nger Nature Li mited. All ri ghts reserved. ©2020 Spri nger Nature Li mited. All ri ghts reserved.

advanced, it is too early to say which types of useful in the short term. some complex calculations, she says. hardware — if any — will prevail. As companies In the past few years, rapid progress in But trapped ions have drawbacks: they are embrace a range of technologies, the field is superconducting loops risked leaving trapped slower at interacting than superconducting wider than ever. ions in the dust. Google, IBM and others have qubits, which will be important when it comes Classical computers store their information developed machines with around 50 or more to accounting for real-time errors coming out as 1s and 0s, but qubits exist in a delicate super- high-quality qubits. IBM aims to have a 1,000- of the system, says Michele Reilly, founder position of 1 and 0. Through the quantum phe- qubit machine by 2023. John Martinis, a quan- of quantum software company Turing in nomenon of entanglement, qubits’ states can tum physicist at the University of California, New York City. And there are limits to how become intertwined, and interference of their Santa Barbara — and, until April, head of quan- many ions can fit in a single trap and be made wavelike quantum states should allow a quan- tum hardware at Google in Mountain View, Cal- to interact. IonQ’s latest model contains tum computer to carry out certain massive ifornia — thinks that Google will use the same 32 trapped ions sitting in a chain; plucking calculations exponentially faster than the best basic architecture it used to achieve quantum any 2 using lasers causes them to interact. To classical machines can. This includes finding advantage to achieve error correction, the next scale up to hundreds of qubits, the company is the factors of prime numbers. big milestone. working on ways to link up multiple chains of Superconducting qubits have so far qubits using photons. The firm aims to double Pros and cons benefited from feeling familiar to many its number of qubits each year. Any system with two possible quantum companies, as their basic components are Meanwhile, Honeywell plans to intercon- mechanical states — such as the oscillations compatible with classical chip technology. nect all the ions by physically shuttling them in a superconducting loop or energy levels around a giant chip (J. M. Pino et al. Preprint at of an ion — could form a qubit, but all hard- “I want to help someone https://arxiv.org/abs/2003.01293; 2020). The ware types have pros and cons, and each faces latest system by the firm’s Honeywell Quantum substantial hurdles to forming a full-blown to build the first quantum Solutions (HQS) division, called H1, consists of quantum computer. A machine capable of computer. It doesn’t have to just ten qubits, but its chief scientist, Patty Lee, living up to the original promise of quantum be me.” says that the firm is already working on its next computing by, for example, cracking conven- iteration. In the next 5 years, the team plans to tional encryption, would require millions of connect around 20 qubits, which should allow individually controllable qubits. But size is not But trapped-ion qubits have many inherent the machine to carry out problems that are the only issue: the quality of the qubits and advantages, says Sabrina Maniscalco, a quan- otherwise impractical on classical machines, how well they connect to each other are just tum physicist at the University of Helsinki. says Tony Uttley, president of HQS. as important. Their operations are much less prone to The challenge is to keep the quality and pre- The frequency of errors in qubits and errors and the delicate quantum states of cision of qubits, while controlling dozens, their operations, caused by noise, tends to individual ions last longer than those in or even hundreds, at once — which neither increase as more are connected. To operate superconducting qubits, which, although Honeywell nor IonQ has yet shown it can do. at scale, each qubit needs to work with error small, are still made of a very large number Although many of the necessary components rates that are low enough to allow mistakes of atoms. Moreover, superconducting qubits have been mastered individually, “what is to be detected and fixed in a process known tend to interact only with their nearest neigh- needed is a system-level integrative approach as error correction, although physicists also bours, whereas trapped ions can interact with putting it all together, testing it, solving its hope that smaller, noisier systems will prove many others, which makes it easier to run problems”, says Barbara Terhal, a theoretical physicist at Delft University of Technology in the .

No clear victor Trapped-ion hardware isn’t the only tech- nology attracting substantial investment. The success of superconducting qubits has opened doors for various technologies, says Slichter, including silicon-based spin qubits, which store in the nuclear spin states of an atom embedded in a silicon crystal. In a coup for this technol- ogy, Martinis joined Silicon Quantum Com- puting in Sydney, Australia, on a six-month sabbatical in September — his first move away from superconducting systems in almost two decades. Martinis doesn’t mind which design ends up winning. “I want to help someone to build the first quantum computer. It doesn’t have to be me [or] whatever I’m working with,” he says. The race is also far from being called, says Maniscalco, and a winner might never emerge. “It may be that there isn’t one

HONEYWELL QUANTUM SOLUTIONS QUANTUM HONEYWELL winning platform, but we have a hybrid or An ion trap from Honeywell’s quantum computer inside a vacuum chamber. different platforms being useful for different tasks.”

Nature | Vol 587 | 19 November 2020 | 343 ©2020 Spri nger Nature Li mited. All ri ghts reserved. ©2020 Spri nger Nature Li mited. All ri ghts reserved.