Picks and shovels for quantum computing: the companies that supply the industry. What are the picks and shovels for quantum computing? A checkable supply chain map: dilution refrigerators, control electronics, lasers, and the compute layer connecting QPUs to classical hardware. This is not investment advice, it's a map of who builds the infrastructure. FreeQuantumComputing "What are the picks and shovels for quantum computing" is a question that comes up regularly in investing circles, most recently on r/ValueInvesting. It's a fair question and a real supply chain sits behind it. This is not stock-picking advice, and nothing here is a recommendation to buy anything. It's a map of who builds the infrastructure every quantum hardware company depends on, checked against real market data and real deployments rather than a forum thread's comments. Superconducting qubits run at temperatures colder than deep space, and every superconducting quantum computer needs a dilution refrigerator to get there. Two companies dominate that market: Bluefors, a private Finnish company with roughly 34% global share, and Oxford Instruments (LSE: OXIG), whose NanoScience division holds close to 21% share with its Proteox and TritonXL systems. Together they cover more than 70% of the dilution refrigerator market, and quantum computing now drives roughly 64% of demand for that equipment. Bluefors isn't publicly traded, so Oxford Instruments is the direct public route into this specific bottleneck. It's also a diversified scientific instruments business, so a dilution refrigerator is one product line among several, not the whole business. A quantum processor is useless without hardware that generates precisely timed microwave pulses to drive qubits and reads the result back. Keysight Technologies (NASDAQ: KEYS) built its Quantum Control System around exactly that problem, and it isn't a paper product. Keysight's QCS is embedded in Fujitsu and RIKEN's 256-qubit superconducting computer, and in mid-2025 the company installed what it describes as the world's largest commercial quantum control system, capable of driving more than 1,000 superconducting qubits, at Japan's AIST G-QuAT center. Zurich Instruments, a smaller private player, supplies similar control hardware and partnered with IQM and NVIDIA in March 2026 on a real-time error-correction demonstration, worth knowing about even though it isn't separately investable. Trapped-ion and neutral-atom qubits run on precisely tuned lasers, and every cryogenic system needs vacuum equipment to hold its insulating vacuum. MKS Instruments (NASDAQ: MKSI) sells both: tunable lasers used in quantum research, and vacuum- and gas-based process equipment used across the cryogenic stack. Coherent Corp (NYSE: COHR) is a direct photonics competitor to MKS. The two nearly merged in 2025, when MKS made a competing bid for Coherent during an acquisition battle Coherent ultimately settled with a different suitor, so they remain separate public companies today rather than one combined photonics supplier. NVIDIA (NASDAQ: NVDA) builds no qubits at all, and that's the point of its quantum strategy. Its NVQLink architecture, made generally available through the cudaq-realtime API at GTC 2026, is aimed at being the standard low-latency connection between GPUs and QPUs regardless of which hardware modality wins. The partner list is broad rather than narrow: IQM and Zurich Instruments' joint error-correction demo, Pacific Northwest National Laboratory's open-source GPU-QPU framework, Dell server validation for sub-4-microsecond real-time hosting, and integrations from Quantum Machines, Qblox, SDT, Infleqtion, and Quantinuum. Its own coverage of the real-time decoding bottleneck covers why that low-latency link matters as much as the qubits themselves. NVIDIA's quantum business is a rounding error against its AI chip revenue today, which cuts both ways: real diversification if quantum stalls, real dilution if it takes off. Every ion trap chip, photonic circuit, and control ASIC still needs a foundry, and the government contracts increasingly require a domestic one. GlobalFoundries (NASDAQ: GFS), which Freequantumcomputing covered in detail, took a $300 million CHIPS Act award to scale US silicon photonics manufacturing, relevant to photonic quantum hardware even though the award itself targets AI infrastructure broadly. SkyWater Technology used to be the cleanest public solution to "who fabricates quantum hardware's supporting electronics," until IonQ acquired it outright in July 2026. Freequantumcomputing examined what that deal changes here: SkyWater is now a wholly owned IonQ subsidiary, not a separate, diversified supplier competitors also buy from. Honeywell (NASDAQ: HON) isn't a supplier in the same sense as the companies above, but it's the closest thing to a backdoor into Quantinuum's technology without buying Quantinuum stock directly. Honeywell retained roughly 48-49% of Quantinuum after its June 2026 IPO, which Freequantumcomputing covered when it happened, giving HON shareholders real economic exposure to Quantinuum's trapped-ion roadmap folded into a much larger, diversified industrial business. A dilution refrigerator, a control system, a laser, and a low-latency compute link are useful whether the winning architecture turns out to be superconducting, trapped-ion, neutral-atom, or photonic. That's a genuinely different risk profile from betting on a single hardware vendor's roadmap, the kind of bet its ranking of the top quantum computing companies covers directly. In an actual gold rush, the shovel seller's whole business was shovels. Here, quantum computing is a minor, early slice of Oxford Instruments', MKS's, Keysight's, and NVIDIA's total revenue, not the main event. That cuts two ways: these companies aren't exposed to quantum computing's risk the way a pure-play hardware vendor is, and quantum computing's success or failure won't move their results by much either, at least not yet. Treating this list as a shortcut to quantum-computing-sized returns from established, diversified industrial companies is a different bet than "picks and shovels" investing usually means. This is not investment advice. It's a supply chain map, verified against real market share data and real deployed systems. Its ranking of the hardware vendors themselves covers the firm-level comparison instead.

IQM and Zurich Instruments develop real-time QEC via NVIDIA NVQLink. IQM Quantum Computers and Zurich Instruments have launched a joint project to build a real-time Quantum Error Correction (QEC) demonstrator integrated with the NVIDIA NVQLink platform. The system utilizes IQM's 20-qubit superconducting quantum processor and the newly developed Zurich Instruments ZQCS Quantum Control System to establish a high-performance control loop. This architecture is designed to move beyond passive error monitoring by enabling active, low-latency feedback necessary for maintaining stable logical qubits. The demonstrator serves as a technical blueprint for the industrial-grade integration of superconducting hardware into standard enterprise datacenter environments. The technical core of the demonstrator is the NVIDIA NVQLink interconnect, which provides a direct, low-latency link between the quantum control electronics and GPU-accelerated classical compute nodes. NVQLink facilitates RDMA over Ethernet to achieve an end-to-end latency of less than 4 microseconds (μs), enabling the high-throughput data transfer required for real-time error decoding. By offloading complex syndrome decoding tasks to NVIDIA GPUs, the system can process error detections and issue corrective feedback within the coherence time of the physical superconducting qubits. This integration utilizes the NVIDIA CUDA-Q platform to unify quantum kernels and classical control tasks into a single, coherent programming environment. By aligning the quantum processing unit (QPU), control electronics, and classical acceleration within a singular operational architecture, the project provides a scalable path toward fault-tolerant computing. The ZQCS control system is specifically engineered to handle the synchronized timing requirements of the 20-qubit array while interfacing with the NVQLink platform for continuous calibration and active error mitigation. This project transitions quantum infrastructure from a peripheral, API-based service to a native, high-performance peer within the supercomputing stack, allowing for the deployment of logical qubits with error rates in the targeted range of $10[-5] to $10[-6] for future industrial applications. For further details on the NVQLink architecture and the 20-qubit superconducting hardware, consult the official IQM announcement here. March 16, 2026