Orca Aerospace

Orca Aerospace is building an AI spacecraft operator designed to run onboard the vehicle rather than depend on continuous ground control. The product separates high-level decision-making from execution: the AI selects what the mission should do next, a deterministic planner determines how to perform it, and command guardrails approve or deny actions before they reach the spacecraft.¹

That architecture is the central strength of the concept. It lets Orca keep making mission-level decisions between ground contacts while preserving a deterministic layer for maneuver planning and a separate approval boundary for commands. Y Combinator describes the same division between LLM-based decisions about “what” to do and deterministic tooling for “how” to execute it.²

Public hiring material says the team is building flight software for the onboard operator stack, conducting software-in-the-loop and hardware-in-the-loop testing, and integrating with flight-software platforms customers already use.³⁴ Those statements show an engineering path toward flight deployment, but they do not establish completed flight heritage or independently validated reliability.

Orca does not publish product pricing, a free tier or self-serve purchase path on the reviewed public surfaces. Prospective customers are asked to describe their spacecraft platform and request a briefing.¹

That is consistent with an early aerospace infrastructure product where integration, validation and hardware requirements are likely to be deployment-specific. It also makes economic comparison difficult: buyers cannot evaluate software cost, integration fees, licensing structure or support commitments without direct engagement.

No public API specification or open-source product claim was verified. The product should therefore be treated as contact-led spacecraft infrastructure rather than a generally accessible developer service.

Y Combinator lists Orca Aerospace as an active Fall 2026 company and corroborates the founders, team size, San Francisco location and onboard-autonomy positioning.² The company site also states that Orca is registered for U.S. government and defense contracts and publishes a CAGE identifier.¹ These are useful identity and procurement-context signals, not evidence of flight qualification or mission success.

The explicit separation between AI goal selection, deterministic planning and command guardrails is a positive safety design principle.¹ However, public evidence is still thin on the controls that matter for a mission-critical flight system: fault containment, model update policy, verification and validation, cybersecurity, telemetry handling, safe-state behavior, radiation/hardware qualification and operational incident history are not described in enough detail to assess them independently.

Orca's public privacy policy covers website visitors and people who contact the company. It describes ordinary website/contact-data collection, retention and security, but it does not describe spacecraft telemetry, mission data or customer operational data.⁵ Buyers should not infer a spacecraft-data policy from that document.

Orca is aimed at spacecraft operators and manufacturers that face a real autonomy problem: large constellations, limited ground passes, constrained links or missions where waiting for ground intervention costs time or capability.¹ Government and defense programs are also a plausible fit given the company's stated contract registration and the mission-critical nature of onboard autonomy.

It is a weaker fit for organizations that need a mature, flight-proven autonomy stack with extensive public qualification evidence, teams seeking transparent pricing before vendor engagement, or programs that cannot accommodate an early-stage supplier in a safety-critical flight-software chain.