PsiQuantum's photonics bet on million qubits

PsiQuantum is making a manufacturing bet, not a demo bet. The point of targeting millions of qubits is that useful quantum computing likely starts only after heavy error correction, which means a machine needs far more physical qubits than the small systems most rivals showcase today. That pushes the company to solve foundry scale chip production and optical networking early, because a lab scale device does not prove the path to a practical machine.

• PsiQuantum’s system is built around standard 300mm wafer production, with single photon sources, detectors, and optical switches fabricated in semiconductor processes at GlobalFoundries. That matters because a million qubits is really a manufacturing and packaging problem, not just a physics problem.
  1 sacra 2 psiquantum 5 psiquantum
• The core technical bottleneck is connecting many photonic processing units with very low loss switching. PsiQuantum’s own technology roadmap centers on Barium Titanate optical switches and integrated photon detectors, which are the pieces that let many smaller chips act more like one large machine.
  2 psiquantum 5 psiquantum
• That contrasts with companies like IonQ and IBM, which commercialize smaller trapped ion and superconducting systems now. Those systems are real products, but PsiQuantum is optimizing for the stage after early access, when customers in chemistry, materials, and finance need fault tolerant machines that can run long calculations reliably.
  1 sacra 3 ionq 4 ibm 6 ionq

The field is heading toward a split between companies selling access to near term quantum hardware and companies trying to own the first truly fault tolerant system. If PsiQuantum can turn photonics and semiconductor scale manufacturing into a working million qubit architecture, it will compete less like a niche hardware startup and more like the builder of a new class of compute infrastructure.