Dirac Labs, a Madison, Wisconsin startup founded in 2023, is field-testing a quantum positioning sensor to replace GPS in defense and aerospace applications where satellite signals can be jammed or spoofed. The system—a 5×5-centimeter cube—uses nitrogen-vacancy centers in diamond as sensing qubits and fuses 70 years of publicly available U.S. magnetic flight data into an onboard navigation map. A former DARPA program manager advising the company estimates wide-scale production is 1–2 years out.
The hardware couples a diamond substrate from Great Lakes Crystal Technologies and staC12 (a University of Chicago spinoff) with a photonic layer, laser, and photodiode that reads fluorescence from the diamond defects. Current compute uses an Nvidia Jetson Orin Nano; Dirac plans to swap it for an FPGA as the design matures. All AI inference runs on-chip, quantized to fit without a cloud link. Manufacturing is happening at university fabs today, with a path to SkyWater for low-volume CMOS wafers as volumes grow.
| Layer / Component | Role | Supplier / Source | Status |
|---|---|---|---|
| Diamond substrate (NV centers) | Sensing qubits — magnetic field detection | Great Lakes Crystal Technologies; staC12 (U. of Chicago spinoff) | Current |
| Photonic layer | Couples light into / out of diamond | Dirac Labs | Current |
| Laser | Excites nitrogen-vacancy centers | Dirac Labs | Current |
| Photodiode | Reads NV fluorescence output | Dirac Labs | Current |
| Edge compute | Runs quantized AI inference on-chip | Nvidia Jetson Orin Nano | Current |
| Edge compute (target) | Runs quantized AI inference on-chip | FPGA (supplier TBD) | Planned |
| Semiconductor fab (prototype) | Low-volume sensor manufacturing | University fabs | Current |
| Semiconductor fab (scale) | Low-volume CMOS wafers | SkyWater Technology | Planned |
The navigation logic is a fusion engine built on the 70-year magnetic dataset. "We have built our own fusion engine, which takes 70 years of magnetic data from flights over different parts of the U.S.," COO Aishwarya Das told EE Times. Recent AI acceleration has let the team quantize navigation models to fit on a small chip, with processing happening locally.
GPS jammers transmit disruptive signals on the same GPS frequency. Spoofers send counterfeit signals carrying false location data that GPS receivers accept as valid. A 2019 RTI International study commissioned by NIST estimated a 30-day GPS outage could cost the U.S. $1 billion per day. Since Russia's invasion of Ukraine, GPS denial has become routine on the battlefield, with jamming spreading into Eastern Europe and the Baltic States. CEO Sanket Deshpande: "Spoofing GPS can be disastrous."
| Threat | Mechanism | Effect on Receiver | Economic Impact |
|---|---|---|---|
| Jamming | Transmits disruptive signals on GPS frequency | Receiver cannot acquire or maintain a position fix | ~$1 billion/day (U.S.) for a 30-day outage — RTI International / NIST, 2019 |
| Spoofing | Broadcasts counterfeit signals carrying false location data | Receiver accepts fabricated position as valid — silent, hard to detect | ~$1 billion/day (U.S.) for a 30-day outage — RTI International / NIST, 2019 |
Current integration partners include aircraft and avionics manufacturers in defense, and international firms working on underground and underwater positioning—domains GPS never reached. Dirac is positioning the sensor for seabed mapping; more than 90% of the world's seabed is unmapped, and undersea cable cuts in the Red Sea, Baltic Sea, and Taiwan Strait have become recurring infrastructure disruption vectors. Deshpande declined to name partners, citing feasibility-demonstration milestones that must close before broader customer engagement.
The bottleneck is manufacturing, not physics. Michael Nayak, the former DARPA program manager advising Dirac, said: "The thing that is difficult about quantum navigation sensors is actually not the quantum part of it. It's more the manufacturing part. We know what the potential is, but actually being able to drive the cost down is not something that you can deliver on until you've done it, the classic assembly-line problem." DARPA's RoQS program and the Defense Innovation Unit's Transition of Quantum Sensing program are both pushing quantum PNT from prototype to operational deployment—institutional tailwinds Dirac can ride, but the sensor-to-production gap remains the binding constraint.
Dirac's approach—on-device magnetic map matching with NV-diamond qubits, FPGA-targeted compute, and CMOS-foundry manufacturing—is the cleanest hardware path to GPS-independent positioning at this form factor. The 1–2 year production window is an adviser estimate, not a shipping date. Evaluate it as a second-source option alongside quantum inertial navigation, not a drop-in replacement. Sensor fusion across modalities remains the safe default until field reliability data exists.