Ivan Deutsch, a foundational theorist whose work underpins the entire neutral-atom quantum computing industry including companies like QuEra and Pasqal, is questioning a core assumption of the field: the exclusive use of two energy levels of an atom for qubits. This challenge to established practice comes from a key figure who helped build the current theoretical foundations, giving his doubts considerable weight. Deutsch suggests the field “may not have gotten this right,” potentially leaving significant computational advantages untapped by prematurely settling on this two-level approach and overlooking the potential of utilizing an atom’s multiple energy levels, a concept known as a “qudit.” He isn’t fully convinced, and the reason matters, revealing a nuanced perspective on a debate that could reshape quantum computing’s trajectory.
Neutral Atom Quantum Computing: Foundations Laid by Ivan Deutsch
His journey to this point began with a 1992 PhD focused on quantum optics, sparked by Alain Aspect’s experiments, and a subsequent postdoc at France Telecom. A 1994 conference hosted by NIST, featuring Peter Shor, fundamentally altered his trajectory. “I was blown away,” Deutsch recalls, describing his reaction to Shor’s algorithm and the potential of quantum information processing. This early exposure led to decades of foundational work, culminating in his current position at the University of New Mexico’s Center for Quantum Information and Control. He points to a connection between bosonic cat qubits in superconducting cavities and multi-level atomic encodings, suggesting a pathway for error correction that bypasses the overhead of surface codes. The reframing that multiple hardware groups are making, connecting “leakage” previously considered noise as a resource, is also shifting perspectives and potentially unlocking new fault-tolerance strategies.
QuDits and Multi-Level Systems: Beyond Two-Level Qubits
The prevailing architecture of neutral-atom quantum computing, used by companies like QuEra, Pasqal, Atom Computing, and Infleqtion, currently relies on manipulating only two energy levels within an atom to represent a qubit. However, Ivan Deutsch, a foundational theorist for this field, is now questioning this fundamental assumption, suggesting the field may have prematurely settled on a limited approach. Recent research demonstrates the potential of these multi-level systems; a LANL/CQuIC paper showcased arbitrary SU(10) maps with 0.9992 average fidelity, the core technical result behind the qudecimal computing approach. This connection suggests a pathway to encode error correction within a single atom, potentially reducing the overhead associated with traditional qubit-based encoding. This shift in perspective challenges established practices and invites a deeper exploration of the full potential of atomic systems.
I don’t think it’s a slam dunk. I don’t feel I don’t want to be an evangelist because I don’t really haven’t studied this well enough to be able to say, we really should do quantum computation base 10 rather than base two or base three or what have you.
9992 Fidelity: Nuclear Spin Qudits in Strontium-87
QuEra Computing, Pasqal, Atom Computing, and Infleqtion, companies driving the neutral-atom quantum race, are built upon theoretical foundations largely established by Ivan Deutsch, yet the physicist is now prompting a reevaluation of a fundamental assumption underpinning their technology. Recent research demonstrates the viability of this qudit approach, with Deutsch and colleagues achieving 99.92% fidelity using ten-level nuclear spin qudits in strontium-87. This benchmark, detailed in a paper published by LANL and CQuIC, is the core technical result behind the qudecimal computing approach. The implications extend to error correction, as Deutsch’s team explores encoding qubits within a single qudit, drawing parallels to bosonic cat qubits used in superconducting systems. This reframing of “leakage”, previously considered detrimental noise, as a potentially useful resource is altering how hardware groups approach fault tolerance.
He observes, “Ions are great because they’re charged. You can hold onto them very tightly… Ions are terrible because they’re charged. You can’t push many ions together in a single trap,” highlighting the trade-offs inherent in different quantum modalities. The exploration of qudits represents a willingness to challenge core assumptions, even from a figure instrumental in defining the current landscape, and suggests a nuanced debate is unfolding within the quantum community.
Ions are great because they’re charged. You can hold onto them very tightly… Ions are terrible because they’re charged. You can’t push many ions together in a single trap.
Spin-Cat Codes: Fault Tolerance via Atomic Encoding
The pursuit of stable quantum computation increasingly focuses on how to encode information within the very fabric of atomic structure, moving beyond the conventional reliance on two-level qubits. This re-evaluation stems from a novel approach to error correction, detailed in recent publications, that explores encoding qubits within the spin states of atoms. Deutsch’s work connects the principles of bosonic cat qubits, traditionally used in superconducting circuits, to a distinct proposal for neutral atoms. This allows for the creation of “spin-cat codes,” where a qubit is encoded in a large-spin qudit, potentially surpassing the fault-tolerance thresholds of standard qubit-based encodings. The team demonstrated arbitrary SU(10) maps with average fidelity 0.9992, the core technical result behind the qudecimal computing approach. He isn’t fully convinced, and the reason matters. This intellectual honesty, from a figure so central to the field’s development, highlights a willingness to re-examine foundational principles and explore unconventional pathways toward robust quantum computation.
Conventional wisdom in quantum computing centers on minimizing error, treating any deviation from a defined quantum state as detrimental noise. Notably, Deutsch isn’t yet prepared to become an advocate for the qudit approach, and the reason matters. The shift in thinking extends to how errors themselves are perceived.
New Mexico Quantum Ecosystem: UNM, Sandia, and Los Alamos Collaboration
The collaborative spirit driving quantum innovation in New Mexico is rapidly solidifying, with the University of New Mexico (UNM), Sandia National Laboratories, and Los Alamos National Laboratory forging an increasingly integrated ecosystem. This tri-institutional alliance isn’t merely co-location; it represents a deliberate strategy to build a comprehensive quantum industry within the state, anchored by the federally designated Elevate Quantum Tech Hub. The foundation of this ecosystem rests heavily on the work originating at UNM’s Center for Quantum Information and Control (CQuIC), led by Ivan Deutsch, whose research has underpinned much of the current neutral-atom landscape. Deutsch’s recent questioning of fundamental assumptions, specifically the reliance on two-level qubits, highlights a willingness to re-examine established practices. Sandia National Laboratories contributes significant expertise in Rydberg gate experiments, collaborating closely with UNM researchers.
Los Alamos is demonstrating qudit-based approaches with a LANL/CQuIC paper achieving 0.9992 average fidelity using ten-level nuclear spin qudits; the core technical result behind the qudecimal computing approach. Leakage, long treated as pure noise in qubit-based systems, is being reframed as a resource, potentially unlocking new avenues for fault tolerance. The combined strengths of these institutions are positioning New Mexico as a key player in the evolving quantum landscape, with a stated goal of moving beyond theoretical exploration toward practical, scalable quantum technologies.
NISQ Era Assessment: Deutsch’s Perspective on Early Quantum Progress
This questioning from a foundational theorist carries considerable weight, given his decades of contributions to neutral-atom control and quantum information theory. Deutsch’s concern centers on the possibility that the field prematurely converged on two-level qubits, potentially overlooking computational advantages offered by utilizing the full spectrum of an atom’s multiple energy levels, an approach known as “qudits.” He highlights a historical pattern where early consensus can inadvertently lock in suboptimal choices, hindering future progress. This isn’t merely about incremental improvements to existing qubits, but a potential paradigm shift in how quantum information is encoded and processed. His work is currently exploring how concepts like “leakage”, previously considered noise, can be reframed as a resource, potentially altering fault-tolerance strategies across multiple quantum platforms, and ultimately, unlocking more powerful quantum computation.
I’m interested in foundations. I work on questions about quantum complexity and where does quantum advantage lie and the role of noise in limiting complexity.
This reframing connects directly to error correction strategies, specifically the intriguing parallel between bosonic cat qubits in superconducting circuits and a novel proposal for neutral atoms. The reframing is quietly changing how multiple hardware groups think about errors. “Ions are great because they’re charged.”
