The problem with managing a fleet of satellites is that the sky is a crowded, noisy, and uncooperative place. Signal fades, traffic spikes, and orbital debris all conspire to drop a packet, and by the time the telemetry tells you it happened, it's already too late. Constellation Space Corp, a Seattle startup founded last year, is betting that the right answer is to stop reacting and start forecasting.
Its product, ConstellationOS, is an ML-native operations platform that ingests a satellite operator's live telemetry and ground data. It then runs continuous forecasts for signal-to-noise ratio, traffic, weather, jamming, and potential collisions, aiming to give operators hours of warning before a link degrades [constellation.space, August 2026]. The software can then route data, isolate failing streams, and orchestrate hand-offs, all within the guardrails of an operator's own policy [constellation.space, August 2026].
The Wedge: From Telemetry to Policy
The company's wedge is straightforward: connect everything, predict the failures, and act within the rules. For an operator running dozens or hundreds of satellites, the sheer volume of telemetry from different manufacturers and ground stations is a headache. Constellation promises a single API and schema to ingest it all [constellation.space, August 2026].
By forecasting critical events like conjunctions or signal degradation, the platform shifts the operator's job from firefighting to strategic oversight. The final piece is the policy engine, which allows the operator to set rules for how the system should autonomously respond to those forecasts, rerouting traffic, powering down non-critical systems, or alerting human controllers, with a full audit trail [constellation.space, August 2026].
The Team and the Backers
The founders bring the kind of pedigrees that make venture capitalists nod thoughtfully. CEO Kamran Majid is a former SpaceX and NASA engineer [constellation.space, August 2026]. CTO Raaid Kabir comes from Blue Origin [Y Combinator]. They are joined by Head of Product Omeed Tehrani and Head of AI Laith Altarabishi [constellation.space, August 2026]. As of the Winter 2026 Y Combinator batch, the team was just four people [Y Combinator], but they are now hiring for roles in software, machine learning, and flight software in Seattle [constellation.space, August 2026].
Their investor list reads like a who's who of strategic and financial heavyweights, including Y Combinator, NVIDIA, OpenAI, Standard Capital, Samsung NEXT, Fellows Fund, and Founders Future [LinkedIn].
The Early-Stage Reality
For all the impressive backing and technical ambition, Constellation Space is still in the earliest phase of commercial proof. The company offers a 30-day shadow pilot, a common tactic for complex enterprise software that needs to prove its value in a live environment without disrupting operations [constellation.space, August 2026]. Its public materials speak of enabling "the world's leading space companies" but do not name specific customers [LinkedIn].
The competitive landscape is also opaque. The company does not name direct competitors in its materials. The primary risk is not a head-to-head feature war, but the immense inertia of large, risk-averse satellite operators who may prefer to build similar capabilities in-house over trusting a startup with mission assurance.
Constellation's answer to that inertia is likely its team's credibility and the sheer complexity of the problem. Building a system that can accurately forecast link quality across a dynamic fleet requires deep domain knowledge and serious AI chops.
Sources
- [constellation.space, August 2026] Constellation Space Corp | ML-native operations for satellite fleets | https://constellation.space/
- [constellation.space, August 2026] Company page | https://constellation.space/company
- [Y Combinator] Y Combinator company profile | https://www.ycombinator.com/companies/constellation-space
- [LinkedIn] Constellation Space Corp LinkedIn page | https://www.linkedin.com/company/constellationspace