IQM Quantum Computers is marketing a fundamentally different product with Spark: not a research system or cloud QPU, but a 5-qubit superconducting processor packaged specifically for university computer science and physics departments to install on-campus.
Hardware Reality
Spark ships in a BlueFors LD-250 dilution refrigerator, pre-configured and factory-calibrated. The package is designed to be operational within 72 hours of delivery — a stark contrast to the weeks-to-months commissioning time of research-grade QPUs. IQM handles remote calibration updates via their cloud control plane, so users don't need on-site quantum hardware expertise.
Qubit count is intentionally minimal at 5 qubits with ring connectivity. This is a pedagogical choice: 5 qubits is large enough to demonstrate quantum entanglement, Bell states, and basic variational algorithms, while small enough that students can track full 32-dimensional state vectors classically for comparison.
Software Stack
Spark uses IQM's Resonance cloud platform for circuit submission, supporting Qiskit and Cirq front-ends. A purpose-built educational curriculum from IQM Academy accompanies every Spark deployment — 12 lab modules covering superposition through simple VQE.
The real-time access model is novel: students reserve 20-minute execution slots via a web scheduler, submit circuits, and receive results within their window. For classrooms of 20–30 students, the bottleneck is the scheduling system, not the hardware.
Does On-Premises Actually Matter?
The honest pedagogical question: does running circuits on real hardware in the building deliver meaningfully different learning outcomes than using cloud QPUs? Two physics department chairs we spoke to argued yes — the physical presence of the cryostat demystifies the technology and enables lab tours that cloud access cannot replicate. One noted that grants for campus infrastructure were more accessible than cloud compute budgets.
The counter-argument: 5 physical qubits offers limited quantum advantage exploration; students working on real algorithms will quickly outgrow Spark and move to cloud systems anyway. IQM's pitch is that Spark is the gateway, not the destination.
At approximately €500,000 per system (estimated, IQM doesn't publish list prices), Spark is competitive with other major research equipment purchases and within reach for well-funded physics departments.