A Chinese technology company called SpinQ has begun selling desktop quantum computers that run at room temperature, marking the first time buyers can purchase such hardware without the extreme cooling systems that define most quantum platforms. The machines, called Gemini and Triangulum, use nuclear magnetic resonance (NMR) technology and fit on a standard lab bench. They are aimed squarely at universities and research groups looking for hands-on quantum instruction tools rather than large-scale computational power.
Why bench-top NMR quantum machines change the buying equation
Most quantum computers require cooling to temperatures near absolute zero, which means multimillion-dollar dilution refrigerators and specialized facilities. That cost barrier has kept quantum hardware out of reach for the vast majority of university physics and computer science departments. SpinQ’s approach sidesteps the cooling problem entirely by encoding qubits in the nuclear spins of molecules in liquid samples, a well-understood NMR technique that operates in ordinary laboratory conditions.
The practical result is that a department chair can, in principle, order a unit, place it on a countertop, and have students running basic quantum algorithms within weeks. If universities begin purchasing these machines in meaningful numbers, a plausible outcome is a measurable rise in published research on small-scale NMR quantum algorithms. That growth could outpace comparable gains in cryogenic qubit research simply because access is cheaper and faster. Within two years, the split in academic output between NMR-based and cryogenic-based quantum work could become visible in citation databases, creating a new category of experimental quantum computing literature driven by affordability rather than raw qubit counts.
That scenario depends on whether the machines deliver enough capability to produce publishable results. NMR qubits are inherently limited in scalability. The Gemini platform offers two qubits, and the Triangulum extends that to three. Those numbers are far below the dozens or hundreds of qubits available on cloud-accessible superconducting platforms from major providers. The question is whether the educational and small-experiment value justifies the purchase when cloud alternatives exist at no hardware cost.
There is also a pedagogical trade-off. A physical device that students can see, hear, and occasionally troubleshoot may offer a different learning experience than abstracted access through a web portal. For instructors building laboratory courses in quantum information, the ability to design experiments around the quirks and constraints of a specific machine can be valuable. At the same time, the tiny qubit counts and lack of scalability could limit how much of the modern quantum computing curriculum can be explored on these systems alone.
SpinQ Gemini and Triangulum: what the technical record shows
The strongest available documentation for both machines comes from technical preprints authored by SpinQ-affiliated researchers and posted on arXiv. The Gemini hardware is described in detail in a paper on a desktop quantum computer designed for education and research, with explicit claims of room-temperature operation. The device uses standard NMR pulse sequences to manipulate nuclear spins, and the preprint outlines how users can implement basic quantum gates and simple algorithms on the platform.
A separate preprint describes the Triangulum as a three-qubit system that extends the same NMR architecture. That paper positions the machine as a step up from the Gemini, adding a third qubit while retaining room-temperature operation and a compact footprint suitable for classroom or small-lab environments. Both documents present pulse-level control, gate sets, and example algorithms as core features, suggesting that users can move beyond canned demonstrations to custom circuit design.
The preprints emphasize that these are not one-off laboratory prototypes but commercial products. They describe enclosures, control software, and user interfaces meant for non-specialist operators, including undergraduate students. The authors highlight use cases such as teaching quantum superposition and entanglement, verifying textbook algorithms like Deutsch–Jozsa on real hardware, and giving students exposure to experimental uncertainties that do not appear in idealized simulations.
Importantly, both preprints have entered the formal academic record through DOI-linked entries. The Gemini identifier and a corresponding listing for the Triangulum allow other researchers to cite the hardware unambiguously. That citation infrastructure matters because it gives the machines a stable identity in the literature, which is a prerequisite for any downstream research community to form around them. Over time, if multiple groups report experiments on “Gemini-class” or “Triangulum-class” devices, those references can be traced back to the same technical baselines.
Gaps in independent testing and shipment data for SpinQ hardware
The technical preprints establish that the Gemini and Triangulum exist and describe their intended capabilities. What they do not provide is independent verification. No third-party benchmarks of gate fidelity, coherence times, or error rates appear in the available primary documents. The performance measurements reported so far come exclusively from SpinQ-affiliated authors, and no university or national laboratory has published an assessment of either machine based on the sources reviewed here.
This absence of outside testing leaves important questions unanswered. For example, it is unclear how closely real-world behavior matches the idealized gate operations presented in the educational materials. Instructors designing experiments around interference patterns or entangled states would need to know how reproducible those states are over many runs and how quickly decoherence degrades results. Without independent benchmarks, prospective buyers must rely on vendor claims or conduct their own evaluations after purchase.
Sales volume is equally opaque. The preprints describe the machines as commercial products targeting education and research buyers, but no invoices, shipment records, or customer counts appear in the technical documentation. Whether a handful of early adopters or a broader set of institutions have acquired units is not answerable from the available record. Pricing, maintenance contracts, and warranty terms are also absent, leaving department heads without public reference points for budgeting or comparing SpinQ’s offering to other instructional technologies.
Long-term reliability data is another gap. NMR systems require periodic calibration, careful magnetic shielding, and, in some cases, replacement or reconditioning of the sample materials. The operational burden over months or years of classroom use is not documented in either paper. Questions such as how often the device must be recalibrated, how sensitive it is to environmental noise in typical teaching labs, and what downtime to expect during maintenance are all left open. A university considering a purchase would need to request that information directly from SpinQ, because the published literature does not address it.
The limited qubit counts raise a separate question about staying power. Two and three qubits allow demonstrations of basic quantum phenomena, including superposition, entanglement, and simple gate operations. They do not allow meaningful exploration of quantum error correction, variational algorithms on nontrivial problem sizes, or any application that requires more than a handful of qubits. In practice, the machines occupy a narrow band between classical simulation of quantum circuits, which any laptop can do efficiently for small qubit numbers, and the larger-scale devices that underpin current research into quantum advantage.
How universities might integrate SpinQ systems
Given these constraints, the most plausible role for Gemini and Triangulum is as specialized teaching tools rather than general-purpose research platforms. In an introductory quantum computing course, instructors could use the machines to anchor laboratory modules on single-qubit rotations, Bell-state preparation, and simple two- or three-qubit algorithms. Students could compare ideal circuit simulations with experimental data, gaining intuition about noise and measurement statistics.
In more advanced courses, the hardware might support small projects on pulse sequence design, calibration strategies, or optimization of specific gate decompositions within the NMR framework. Even without scaling to larger qubit counts, these topics intersect with real challenges in broader quantum engineering. For institutions that lack access to large cryogenic systems, SpinQ’s devices could serve as an accessible bridge between purely theoretical coursework and the complexities of experimental practice.
Whether that bridge justifies the investment depends on factors that the current literature does not quantify: acquisition cost, ongoing support, and the comparative value of alternative options such as cloud-based access to larger quantum processors. For now, the public record shows that room-temperature, bench-top quantum computers are technically feasible and commercially offered, but leaves open how widely they have been adopted and how well they perform outside the vendor’s own lab.
For researchers and department heads watching this space, the next development to track is whether any independent group publishes benchmarking results or longitudinal case studies of classroom use. Such reports would clarify where SpinQ’s Gemini and Triangulum fit in the evolving ecosystem of quantum technologies-and whether room-temperature NMR machines become a standard feature of quantum education or remain a niche experiment in making quantum hardware more tangible.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
