Oratomic Launches With Caltech Research Showing Shor's Algorithm Is Feasible With Just 10,000 Neutral-Atom Qubits
A Caltech-backed startup claims utility-scale quantum computers need far fewer qubits than previously estimated, potentially accelerating the timeline for cryptographically relevant machines to the end of the decade.
Overview
A new quantum computing startup backed by researchers from Caltech, Harvard, and Berkeley has emerged with a claim that could reshape the field’s timeline: useful, cryptographically relevant quantum computers may need only 10,000 to 20,000 qubits, not the millions that have been the standard assumption for over a decade.
Oratomic launched on March 31 with a founding team of more than 15 quantum computing researchers drawn from Caltech, Harvard, UC Berkeley, Amazon, and Google, according to a company announcement via PR Newswire. The company is led by CEO Dolev Bluvstein, a pioneer in fault-tolerant quantum computing, and is built around breakthrough research conducted in collaboration with Caltech scientists.
The timing is notable. Neutral-atom quantum computing has been gaining momentum throughout 2026, with companies like QuEra, Pasqal, and Atom Computing deploying commercial error-corrected systems, as previously reported by The Machine Herald. Oratomic’s research suggests the hardware threshold for practical quantum computation may be far lower than the field assumed.
What We Know
The core claim rests on a paper titled “Shor’s algorithm is possible with as few as 10,000 reconfigurable atomic qubits,” co-authored by Oratomic founders and Caltech researchers. According to Caltech’s announcement, the research demonstrates that the number of physical qubits per logical qubit can be reduced from roughly 1,000 to as few as five, enabled by the unique properties of neutral-atom platforms.
The key technical innovation is the use of dynamically reconfigurable qubit arrays. Unlike superconducting qubits, which are fixed on chips, neutral atoms are suspended in vacuum and positioned by optical tweezers. As Caltech physics professor Manuel Endres explained, according to Caltech, “Unlike other platforms, neutral atom qubits can be directly connected over large distances.” This reconfigurability enables “high-rate codes” where each physical qubit participates in multiple logical qubits, dramatically improving error-correction efficiency.
The practical implications are stark. The paper estimates that under plausible assumptions, a system with approximately 26,000 physical qubits could compute discrete logarithms on the P-256 elliptic curve in a matter of days, according to Caltech. Factoring RSA-2048 integers would require one to two orders of magnitude more time but remains within reach of a machine far smaller than previously thought necessary.
As Science News reported, the research indicates that quantum computers based on atoms could decrypt internet encryption schemes “much sooner than scientists thought,” a conclusion that has implications for the global transition to post-quantum cryptography.
Oratomic co-founder Endres has already demonstrated arrays of 6,000 atomic qubits, according to The Quantum Insider, placing the company’s existing hardware roughly halfway to the 10,000-qubit threshold identified in the research. Co-first author Madelyn Cain said, as reported by Caltech, that the results “reduce qubit counts by up to two orders of magnitude.”
What We Don’t Know
Several critical questions remain unanswered. Oratomic has not disclosed any funding details, leaving the company’s financial runway and investor backing unknown. The research paper’s estimates rely on “plausible assumptions” about qubit performance, but whether those assumptions will hold in engineered hardware at scale is unproven.
The gap between trapping 6,000 atoms and running fault-tolerant algorithms on 10,000 is not trivial. High-fidelity multi-qubit gates, real-time error decoding, and sustained coherence across a reconfigurable array present engineering challenges that the paper’s theoretical framework does not resolve. Bluvstein himself has acknowledged uncertainty, stating that achieving a fault-tolerant quantum computer by the end of the decade is “plausible, although not guaranteed,” according to PR Newswire.
It also remains to be seen how Oratomic will position itself against established neutral-atom competitors like QuEra, which raised over $230 million in 2025, and Atom Computing, now part of Quantum Circuits through its merger. The competitive landscape in neutral-atom quantum computing is crowding quickly.
Analysis
The significance of the Oratomic research is not that it proves quantum computers will break encryption soon. It is that it lowers the bar. For years, the quantum computing field has operated under the assumption that millions of physical qubits would be needed for practical computation, a target so distant that it placed useful quantum computers at least a decade away. A two-orders-of-magnitude reduction in that estimate changes the calculus for governments, financial institutions, and anyone relying on current encryption standards.
The research also underscores the growing divergence between quantum computing platforms. Neutral-atom systems are emerging as the frontrunner for fault-tolerant computing specifically because of their reconfigurability, an advantage that superconducting and trapped-ion platforms cannot easily replicate. That Oratomic’s founding team includes John Preskill, one of the originators of the concept of quantum computational supremacy, lends the effort additional credibility.
The urgency of post-quantum cryptographic migration, already a priority for agencies like NIST, may intensify in light of these findings. If 10,000 qubits is indeed sufficient, the window for transitioning to quantum-resistant encryption standards is shorter than most organizations have planned for.