Key Takeaways
- Photonic quantum computers use individual photons as qubits, enabling room‑temperature operation and direct compatibility with existing optical‑communication infrastructure.
- The two main encoding strategies are discrete‑variable (polarization, path, time‑bin) and continuous‑variable (GKP states), both hampered primarily by photon loss.
- Leading companies—PsiQuantum, Xanadu, ORCA Computing, Quandela, QuiX Quantum, Quantum Source, Nu Quantum, Q.ANT, and Sparrow Quantum—are pursuing diverse approaches ranging from fusion‑based cluster states to deterministic single‑photon sources and photonic networking fabrics.
- Compared with trapped‑ion and superconducting platforms, photonics offers scalability via mature silicon‑photonic manufacturing and natural suitability for quantum networking, though gate fidelities and loss remain technical hurdles.
- Industry consensus is that no single modality will dominate; the optimal hardware choice will depend on the target application and system‑level integration requirements.
Overview of Photonic Quantum Computing
Photonic quantum computing encodes quantum information in photons—particles of light—rather than in electronic spins or currents. Because photons are largely immune to thermal decoherence, the processor itself can operate at room temperature, eliminating the need for large dilution refrigerators that superconducting and many trapped‑ion systems require. Only ancillary components such as single‑photon detectors may still need cryogenic cooling, but the overall hardware footprint is markedly lighter and can be integrated with existing fiber‑optic networks used for telecommunications.
How Photonic Qubits Store and Manipulate Information
In a photonic quantum computer, each photon serves as a qubit, and information is impressed onto its physical degrees of freedom. Discrete‑variable (DV) encodings use properties such as polarization, arrival time (time‑bin), or spatial path within an interferometer. Continuous‑variable (CV) approaches, by contrast, encode data in the quadrature amplitudes and phase of coherent light fields, enabling techniques like Gottesman‑Kitaev‑Preskill (GKP) error‑correctable states. Quantum gates are realized with linear‑optical elements—beam splitters, phase shifters, and waveguides—that implement interferometric transformations. Because photons do not naturally interact, two‑qubit gates are probabilistic in linear optics and rely on ancillary photons, measurement‑based feed‑forward, or nonlinear media to achieve determinism.
Core Challenges: Loss and Scalability
The dominant obstacle for photonic platforms is photon loss. A photon that is absorbed, scattered, or Detected cannot be recovered or reset, unlike a trapped ion or superconducting qubit that can be re‑initialized. Loss accumulates with each optical component, making low‑loss waveguides, high‑efficiency detectors, and near‑unit‑efficiency sources essential. Scaling also demands large numbers of indistinguishable photons and sophisticated timing synchronization; approaches such as time‑domain multiplexing (used by ORCA) or fusion‑based cluster‑state construction (PsiQuantum) attempt to mitigate these demands by recycling photons or building entanglement incrementally.
Advantages Over Other Quantum Modalities
Photonics leverages the mature silicon‑photonics supply chain that already produces high‑volume photonic integrated circuits for data‑center interconnects. This foundation offers a clear path to wafer‑scale fabrication, reducing the custom‑process burden faced by nascent superconducting or ion‑trap technologies. Furthermore, photons travel through standard optical fiber with minimal attenuation at telecom wavelengths, creating a natural interface for quantum networks and distributed quantum computing. While trapped‑ion systems currently lead in two‑qubit gate fidelity (>99.5 %) and superconducting circuits excel in qubit density, photonics provides a unique combination of room‑temperature operation, networking readiness, and manufacturability.
Company Spotlight: PsiQuantum
Based in Palo Alto, PsiQuantum pursues fault‑tolerant quantum computing through fusion‑based quantum computing (FBQC) on silicon photonic chips fabricated by GlobalFoundries. The company aims to deliver a utility‑scale, error‑corrected system by the end of the decade, targeting 2029 as a milestone. Recent financing milestones include a $1 billion Series E round in September 2025 (led by BlackRock, Temasek, and Baillie Gifford) valuing the firm at $7 billion, plus substantial government backing for facilities in Chicago and Moreton Bay, Australia. PsiQuantum’s strategy relies on generating small entangled resource states and fusing them via destructive measurements to build large cluster states for computation.
Company Spotlight: Xanadu
Xanadu, headquartered in Toronto, adopted a continuous‑variable approach, encoding qubits in the amplitude and phase of light and targeting GKP error correction for fault tolerance. After going public via a SPAC merger in March 2026, the company reported $302 million in gross proceeds. In June 2025 Xanadu demonstrated on‑chip GKP state generation in a Nature paper, the first time error‑correctable photonic qubits were produced on a silicon‑nitride chip. The firm now reports 12 logical GKP qubits with real‑time error correction and a 60 % reduction in optical loss, projecting up to 500 logical qubits by 2029‑2030.
Company Spotlight: ORCA Computing
ORCA Computing, based in London, uses time‑domain multiplexing to store photons in fiber delay lines, synchronizing them for gate operations without requiring simultaneous active sources. Its PT‑2 system is already deployed at the UK National Quantum Computing Centre and has been used for fiber‑network route optimisation trials with Vodafone. In October 2025 ORCA partnered with NVIDIA on the NVQLink open reference architecture for real‑time quantum‑classical integration. The forthcoming PT‑3 system, slated for 2026 release, targets quantum advantage on optimization and generative‑AI workloads.
Company Spotlight: Quandela and QuiX Quantum
Quandela, a French spin‑out, builds gate‑based photonic processors using semiconductor quantum‑dot single‑photon sources, offering both cloud and on‑premises access. In late 2025 it delivered a 12‑qubit “Lucy” system to the French CEA and partnered with OVHcloud to provide BELENOS (12‑qubit) and CANOPUS (24‑qubit) processors on a sovereign European cloud. QuiX Quantum, from Enschede, the Netherlands, fabricates programmable processors on silicon‑nitride waveguides, leveraging the region’s photonics ecosystem. After a €14 million contract with the German Aerospace Center (DLR QCI) for 8‑ and 64‑qubit machines, QuiX secured €15 million Series A funding in July 2025 to deliver the first single‑photon‑based universal photonic quantum computer in 2026.
Company Spotlight: Quantum Source, Nu Quantum, Q.ANT, and Sparrow Quantum
Quantum Source (Israel) employs a cavity‑QED architecture where single atoms on a photonic chip mediate deterministic photon‑photon interactions, aiming to reduce the overhead of linear‑optical gates. Nu Quantum (Cambridge, UK) focuses on photonic networking, providing an “Entanglement Fabric” that links multiple quantum processors via qubit‑photon interfaces; it raised $60 million in Series A in December 2025 and opened a dedicated R&D facility for distributed trapped‑ion quantum computing in early 2026. Q.ANT (Stuttgart) develops processors on thin‑film lithium niobate (TFLN), promising faster electro‑optic modulation and lower loss than silicon photonics, and also pursues quantum‑sensing applications. Sparrow Quantum (Copenhagen) specializes in deterministic single‑photon sources based on quantum dots, supplying the high‑purity, indistinguishable photons required by fusion‑based and boson‑sampling architectures; its €27.5 million Series A in late 2025 marked the largest Nordic quantum‑technology investment to date.
Comparing Photonics to Trapped‑Ion and Superconducting Approaches
While trapped‑ion devices continue to set the benchmark for gate fidelity, photonic systems face difficulty in direct comparison because their error models differ—loss and probabilistic gates dominate rather than decoherence. Nevertheless, photonics enjoys two structural advantages: (1) the ability to leverage commercial semiconductor fabs for photonic‑integrated‑circuit production, providing a scalable manufacturing base; and (2) inherent compatibility with fiber‑optic networks, facilitating distributed quantum computing and future quantum‑internet architectures. Superconducting platforms, by contrast, excel in qubit density and rapid gate speeds but require extensive cryogenic infrastructure. The choice among modalities will therefore hinge on the specific algorithm, system‑size goals, and integration constraints of the end user.
Outlook and Frequently Asked Questions
Experts agree that no single hardware technology will dominate outright; rather, the quantum‑computing landscape will likely host multiple coexisting platforms, each suited to particular niches. Photonics remains attractive for applications that benefit from low‑latency optical links, such as sensor networks, secure communication, and hybrid quantum‑classical data‑center workloads. Continued progress hinges on reducing photon loss, improving deterministic single‑photon sources, and developing efficient error‑correction codes tailored to photonic encodings. As the ecosystem matures, we can expect tighter co‑design of photonic chips, control electronics, and software stacks—bringing room‑temperature, network‑ready quantum processors closer to practical deployment.