NVision, a quantum technology startup, has entered a partnership with U.S. photonics firm HyperLight to integrate molecular qubits onto optical chips. The announcement follows a peer-reviewed study published in Nature on October 28, which demonstrated a coherence time of 2.2 milliseconds for a single molecular spin. The first prototype chips are slated for delivery by the end of this quarter.
The core engineering challenge is whether a single molecule can preserve its laboratory-grade performance when embedded into a chip. In the study, researchers measured 2.2 milliseconds of coherence using a protective sequence of microwave pulses, compared to 12.2 microseconds with a simpler Hahn echo sequence. This represents a roughly 180-fold improvement, highlighting the importance of the protection protocol.
The experiment, conducted at 4.5 kelvin (about -269°C), used a carbene molecule embedded in a closely matched host crystal. Light was used to prepare and read out the electron spin, while microwaves manipulated it. The results confirm that an individual electron spin can be controlled with high precision, a key building block for future quantum processors.
From Lab to Chip
NVision's approach involves depositing organic molecular films directly onto optical cavities. The company's platform description lists film thicknesses of approximately 20–100 nanometers. This method aims to leverage the molecules' natural interaction with light, enabling connectivity without the need for physical proximity between qubits.
"Because our qubits interact directly with light from the start, we achieve connectivity without proximity," said Ilai Schwartz, NVision's technology chief, in the announcement.
HyperLight brings its thin-film lithium-niobate platform, already used in optical communications, to the partnership. The company offers shared and dedicated wafer runs, custom designs, and bare-die delivery. However, molecular-qubit yields and processor performance on this platform have yet to be demonstrated.
Optical Efficiency and Photon Matching
Optical efficiency remains a critical constraint. The paper reports a photoluminescence quantum yield of 10.8% and a zero-phonon line spectral share of 19%, resulting in a combined emission efficiency of about 2.1% into that narrow line before external collection losses. The measured optical linewidth was 38 MHz, about 5.8 times the lifetime-derived limit, which is essential for ensuring photons from different nodes match sufficiently for entanglement.
NVision's architecture tolerates up to 70% photon loss according to proposed networking protocols, but this is a separate metric and does not improve the measured emission efficiency.
Development Roadmap and Risks
The company's targets extend beyond the current results. NVision aims to demonstrate on-chip performance by the end of 2026 and plans to scale to 1,000 physical qubits before moving toward 1,000 logical, fault-tolerant qubits by 2030. Logical qubits require significant error-correction overhead, so the two counts are not directly comparable.
Several risks remain: film deposition could alter optical linewidths or reduce coherence, and efficient coupling, reproducible fabrication, and gates between separate molecules still need to be proven. The Q4 2026 prototype date is a company target, not a guarantee.
This partnership marks a significant step toward practical quantum computing, but as with all early-stage technologies, the path from laboratory success to commercial viability is long and uncertain. Investors should watch for further technical milestones from NVision and its partners.
