
The research, published Tuesday as a preprint on arXiv and not yet peer-reviewed, was conducted at Caltech in collaboration with Oratomic, a Pasadena-based startup that launched the same day with a mission to build the machines. If the theory holds, practical quantum computers — devices powerful enough to solve problems that today’s computers cannot and to break the encryption that protects bank accounts and private messages — could be operational by the end of the decade, according to the researchers, who cautioned that significant engineering challenges remain.
In simple terms, quantum computers use tiny units called qubits instead of the ones and zeros that power ordinary computers. But qubits are fragile. They make errors constantly, and fixing those errors has required piling on enormous numbers of extra qubits — roughly 1,000 backup qubits for every one that does useful work, according to the Caltech press release describing the research. That meant a working quantum computer would need about 1 million qubits in total, a number far beyond current engineering ability.
The Caltech and Oratomic team found a way, in theory, to get that ratio down to about five backup qubits per working qubit. The result: a functioning quantum computer might need only 10,000 to 20,000 qubits instead of 1 million, according to the Caltech press release.
“It’s actually very surprising how well this works,” said Manuel Endres, a professor of physics at Caltech. “It’s what we call ultra-efficient error correction.”
The breakthrough relies on a type of quantum computer built from individual atoms arranged by laser beams — a technology known as neutral-atom quantum computing. Unlike other designs, where qubits are essentially locked in place and can only communicate with their nearest neighbors, neutral-atom qubits can be physically moved across the system and connected over long distances, according to the Caltech press release.
Endres and his Caltech colleagues last year set a record by assembling the largest-ever array of these atomic qubits — 6,100 of them, trapped by lasers — a result published in the journal Nature in September.
“Unlike other quantum computing platforms, neutral atom qubits can be directly connected over large distances,” Endres said, according to the Caltech press release. “Optical tweezers can shuttle one atom to the other end of the array and directly entangle it with another atom.”
That ability to move atoms is central to the new error-correction method. The study, titled “Shor’s algorithm is possible with as few as 10,000 reconfigurable atomic qubits,” was co-first-authored by Madelyn Cain, lead theoretical scientist at Oratomic, and Qian Xu, the Sherman Fairchild Postdoctoral Fellow at Caltech who is now also a research scientist at Oratomic.
Senior authors include Endres; John Preskill, the Richard P. Feynman Professor of Theoretical Physics and the Allen V. C. Davis and Lenabelle Davis Leadership Chair of the Institute for Quantum Information and Matter at Caltech; Hsin-Yuan (Robert) Huang, an assistant professor of theoretical physics and William H. Hurt Scholar at Caltech who is currently on leave while serving as Oratomic’s chief technology officer; and Dolev Bluvstein, a visiting associate in physics at Caltech who is Oratomic’s chief executive officer. Other authors include Oratomic’s Robbie King and Lewis Picard, and Harry Levine of Oratomic and UC Berkeley.
“I’ve been working on fault-tolerant quantum computing longer than some of my coauthors have been alive,” Preskill said, according to the Caltech press release. “Now at last we’re getting close.”
The title of the paper refers to an algorithm developed in 1994 by Peter Shor (BS ’81), a Caltech alumnus who is now a professor of applied mathematics at the Massachusetts Institute of Technology. Shor’s algorithm demonstrated that a quantum computer could break the mathematical codes underpinning modern encryption — a finding that was largely theoretical at the time because no machine was powerful enough to run it.
The new research suggests that a neutral-atom quantum computer with as few as 10,000 qubits could be sufficient to execute that algorithm, according to the Caltech press release. The researchers stressed that the accelerated timeline underscores the urgency of migrating to new encryption standards designed to resist quantum attacks. Global guidelines call for that transition by 2035, according to the Oratomic press release.
“It is plausible, although not guaranteed, that we will have a fault-tolerant quantum computer by the end of the decade,” Bluvstein said, according to the Oratomic press release. “Although exciting and opening the door to a broad range of applications, such advances would also put modern cryptography at-risk. Our results emphasize the importance of transitioning vulnerable cryptosystems to post-quantum encryption.”
Oratomic launched Tuesday from Pasadena with a founding team that includes members from Caltech, Harvard, UC Berkeley, Amazon, Google, and other institutions, according to the company’s press release. The company will work in close collaboration with Caltech’s Advanced Quantum Computing Mission, an on-campus interdisciplinary effort, according to the Caltech press release. The Caltech team has said it plans to eventually have quantum “supercomputers” on campus for scientific research.
The paper’s findings are theoretical, and the researchers acknowledged in the Caltech press release that significant engineering challenges remain before these capabilities can be combined into a working system. The paper was posted to the arXiv preprint server and has not been published in a peer-reviewed journal. CoinDesk reported that all nine authors of the paper are shareholders in Oratomic, with six employed by the company.
“Now it’s time to build the machines,” Bluvstein said, according to the Caltech press release.











