SafeKey LabTalk to usSecurity for the autonomous spacecraft
Security for the autonomous spacecraft.
We give space operators a live picture of the orbital environment, defense for the spacecraft's bus and onboard AI, and a way to prove resilience by attacking it first.
SafeKey Vela
Space Domain Awareness and counter-autonomy for autonomous spacecraft.
Three capabilities, one system. Vela leads with Apsis, our live space domain awareness layer, and extends to onboard defense and adversarial testing for the spacecraft itself.
Pillar 01 · Space Domain Awareness
Apsis
Know what's operating around your satellite.
Real-time orbital tracking and a live common operating picture. Conjunction screening for close-approach and collision warnings. RPO classification that flags suspicious rendezvous and proximity behavior. Driven by public TLEs plus SatNOGS ground-station observations and decoded downlink telemetry.
Status: live on real operational data.
Explore Apsis →
Pillar 02 · Counter-autonomy on the spacecraft
Onboard Defense
Protect the command bus and the onboard AI.
Bus (CSP/CAN) intrusion and anomaly detection for command injection, telemetry spoofing, power and attitude attacks. Onboard-AI protection against adversarial patches, visual prompt injection, supply-chain backdoor triggers, and unsafe-tasking. Safety-envelope enforcement and emergency stop for autonomous actions.
Status: demonstrated against a simulated spacecraft; bus attacks detected with correct type and severity at sub-10-second latency.
Explore Onboard Defense →
Pillar 03 · Adversarial testing
Range
Attack it before an adversary does.
Satellite bus and RF attack suite, plus a six-stage onboard-AI attack chain. Mapped to the SPARTA space-threat framework, producing a resilience scorecard. PIN-gated and audit-logged. Runs only against owned or simulated systems.
Status: built and exercised in our environment.
Explore Range →Air-gapped deployment
Offline-capable for disconnected and tactical-edge environments.
Standards-mapped
SPARTA, MITRE ATLAS, and NIST AI RMF.
Post-quantum crypto
Post-quantum cryptography for command and data links.
Sectors
Space leads. Defense and commercial on the same platform.
Commercial operators, defense space programs, and satellite manufacturers all face the same problem on orbit: hostile, automated behavior they cannot fully predict. We lead with space, and the same architecture stretches across the rest of the customer base.
Sector 01 · Government space
Government space programs
Built for Space Force, NRO, SDA, and prime-integrator programs to run as an independent red and blue layer. Aligned to SPARTA from day one, so the work plugs into the frameworks reviewers already use.
Defense →
Sector 02 · Commercial Space
Constellation operators
Comms, EO, and PNT operators use SafeKey to check that a spacecraft is ready before launch, and to keep checking the fleet once it is up.
Commercial →
Sector 03 · Manufacturers
Satellite builders and primes
Bus manufacturers and primes ship SafeKey-evaluated platforms, so their customers and reviewers have an outside score they can rely on instead of taking the OEM at their word.
Manufacturers →
Posture
Built for how this industry actually buys.
We did not bolt compliance on at the end. SafeKey Lab was designed from day one for the export-controlled, audited environment our customers actually work in.
Export control
USML Cat XI scope. DDTC registration in progress.
Supply chain
Section 889 review underway.
Environmental
On the MIL-STD-810H pathway.
EMI and EMC
On the MIL-STD-461G pathway.
Cybersecurity
CMMC L2 self-assessed (C3PAO pending). IL5 inherited from GCP Assured Workloads. ATO with a 3PAO in progress.
Intellectual property
Patent-pending. 8 patents filed.
Track Record
Validated in the lab. Demonstrated in the field.
We do not ask customers to take our word for any of this. SafeKey has been tested against real captured satellite downlink RF in our own ground station, and put in the hands of a U.S. Army unit through field exercises.
Ground Station / RF
Tested at our Colorado Springs ground station
- Real-RF scorecard, published in full: 4 techniques fully detected, 3 partial, 1 not yet detected, 4 correct holds, not just a headline.
- SDA / orbital monitoring subsystem at TRL 5, running against real TLEs and SatNOGS data.
- Counter-autonomy core at TRL 4, demonstrated in a simulated environment, validated against real captured satellite downlink RF at our Colorado Springs ground station.
- Adversary scenarios replayed through our stack against real captured satellite downlink RF and our CubeSat emulator.
- Mapped to SPARTA, the recognized space threat framework.
In the Field
Exercised with a military unit
- Ground architecture exercised in field exercises with a U.S. Army unit.
- Operated by service members through multiple field exercises, no contractor on the keyboard.
- Briefed to military stakeholders during field exercises.
- Formal research agreement signed on our side, waiting on government execution.
The Company
Built by people who lived the problem.
SafeKey was started by U.S. Army veterans who built and fielded counter-autonomy systems with operational units, then turned the same architecture toward the spacecraft.
Founder & CEO
Sukin Yang
Cornell alum and U.S. Army veteran. While serving overseas, he triaged a large population of service members and their families and simultaneously supported operational missions. After the Army, he built data pipelines scaling 30 years of real estate records, built an AI diagnosis tool at Weill Cornell Medicine that scans MRI to detect breast cancer, and worked on RF and satellite engineering at Cornell Bowers. He built the original counter-autonomy platform end to end and is the inventor behind SafeKey's detection and red/blue team work, with 8 patents filed (patent-pending).
Military Integration
Chance LeBlanc
U.S. Army veteran with experience bridging operational units and emerging technology. Runs government deployments and integration. The reason the system actually got into operators' hands in the field.
Adjacent module
SafeKey Guard. AI and agent security for the ground segment.
Guardrails and red-teaming for AI in the ground segment. Prompt-injection and jailbreak defense, data and secret leakage prevention, AI-agent action approval, and RAG security. OWASP LLM Top 10 testing, mapped to MITRE ATLAS. An optional module for teams deploying AI alongside their space operations, not the core product.
Status: built and functional in our environment.
Get in touch
Let's talk about your spacecraft.
Whether you fly one satellite or a constellation, build them, or buy them, we would like to hear from you. Drop us a line and we will get back to you within two business days.
Or email [email protected]