A satellite in low Earth orbit is a computer hurtling through a vacuum, but its operations are still largely a human-in-the-loop affair. Ground controllers monitor telemetry, diagnose anomalies, and send commands, a process that is both expensive and slow. Orbtrix Space, a 2025-founded startup in Bengaluru, is betting that the next layer of space software will be an AI autonomy stack that runs from the ground station to the satellite itself [Orbtrix Space, September 2026].
Its architecture, which the founders call "mirrored intelligence," proposes to synchronize two core products. DISHA is an AI-enabled ground platform for monitoring constellations and predicting failures. Rigel OS is an onboard operating system designed to let a satellite make real-time decisions without waiting for a command from Earth [Perplexity Sonar Pro Brief, web-grounded]. The goal is straightforward: reduce the need for round-the-clock human oversight and make spacecraft more responsive and resilient.
The Full-Stack Autonomy Bet
Orbtrix is not building a better ground station or a more efficient radio. Its wedge is the software layer that sits between the hardware and the operators. The company lists its target customers as a broad cross-section of the space industry: mission providers, satellite manufacturers, operators, and system integrators [Perplexity Sonar Pro Brief, web-grounded]. For them, the value proposition is operational cost reduction and risk mitigation.
The technical approach is to treat the ground and orbit as a single, distributed system. DISHA would ingest telemetry to build a model of nominal satellite behavior, flagging deviations that could indicate a developing fault. Rigel OS, running on the satellite's limited compute, would be tasked with time-sensitive decisions, like avoiding debris or managing power during an eclipse. The promise is a closed loop where intelligence is mirrored, not merely relayed.
Early Traction and the NVIDIA Stamp
As a pre-seed company with a team estimated at 1-10 employees, Orbtrix's most concrete validation to date is its selection for the NVIDIA for Startups Inception program in February 2026 [LinkedIn, February 2026]. This provides access to NVIDIA's AI tools and ecosystem, a critical resource for a startup whose core differentiator is training and optimizing models for both cloud and edge (space) environments. The company also exhibited at the BSX 2026 industry conference, indicating early efforts to engage with the market [BSX India, September 2026].
The founding team, Aswin Sunil (CEO) and Aaron Steephen (COO), began their roles in November 2025 [Perplexity Sonar Pro Brief, web-grounded]. Public records show the company has begun building technical capacity, with an AI/ML engineer joining in mid-2026 after an internship [Perplexity Sonar Pro Brief, web-grounded].
| Role | Name | Joined |
|---|---|---|
| Founder & CEO | Aswin Sunil | November 2025 |
| Co-founder & COO | Aaron Steephen | November 2025 |
| AI/ML Engineer | Akhila Anil | July 2026 |
| Table: Orbtrix Space's identified core team as of late 2026 [Perplexity Sonar Pro Brief, web-grounded; LinkedIn]. |
The Uphill Path to Orbit
The ambition is significant, and so is the distance to proving it out. The space industry is notoriously conservative, especially when it comes to trusting flight software. Orbtrix must clear several high bars that have grounded similar software promises before.
- The certification gauntlet. Space-grade software requires rigorous verification and validation. Getting a new operating system like Rigel OS qualified for flight on a customer's satellite is a multi-year, expensive process with no guarantee of adoption.
- The data problem. Training effective AI models for anomaly detection requires vast amounts of high-quality telemetry data, much of which is proprietary and closely held by incumbent operators and agencies. Orbtrix's algorithms will only be as good as the data they can access.
- The integration burden. Selling to spacecraft manufacturers means embedding software into a hardware supply chain with long lead times. Selling to operators means integrating with legacy ground systems that were never designed for an AI copilot.
There is also the fundamental question of market timing. While automation is an obvious long-term trend, the immediate budget priorities for many satellite operators may still be launching more assets, not retrofitting existing ones with new AI stacks.
Technical Breakdown and Scale Risks
Architecturally, Orbtrix's two-product approach makes sense. Decoupling ground and edge intelligence allows each to be optimized for its environment,unlimited compute on Earth, severe resource constraints in orbit. The real test will be in the handoff between them. Defining the decision boundary,what the ground decides versus what the satellite can decide autonomously,is a complex systems engineering problem with major safety implications.
At scale, the risks are less about the AI and more about systems integration and liability. A false positive from DISHA could trigger an unnecessary and costly satellite maneuver. A bug in Rigel OS could, in a worst-case scenario, lead to a collision. The company's long-term challenge isn't just building smart software, but building trust in a industry where failure is measured in hundreds of millions of dollars.
For now, Orbtrix Space represents a specific, technically coherent bet on the future of space operations. Its progress will be measured not by funding announcements, but by its first signed commercial contract and, ultimately, by its software flying on a mission where it has real operational authority.
Sources
- [Orbtrix Space, September 2026] Orbtrix Space company website | https://www.orbtrix.space/
- [Perplexity Sonar Pro Brief, web-grounded] Company profile and product details
- [LinkedIn, February 2026] NVIDIA Inception program announcement | https://www.linkedin.com/posts/aswin-sunil-1681341a1_every-orbit-every-planet-absolute-autonomy-activity-7431978156661821440-HSA5
- [BSX India, September 2026] BSX 2026 Exhibitors List | https://bsxindia.com/assets/img-2026/pdf/BSX%202026%20-%20List%20of%20Exhibitors%2005-09-2026..pdf