Four 17-kilogram satellites are quietly rewriting what counts as proof in this industry. Aerospace Corporation’s DiskSat mission, launched on a Rocket Lab Electron in December, has spent the months since descending toward very low Earth orbit (VLEO) at roughly 2 kilometers a month, while engineers commission Enpulsion electric-propulsion systems meant to hold the four vehicles below 300 kilometers once they get there, according to the mission’s flight log as reported by SpaceNews . It is a slow, unglamorous process. Power management, communications, attitude control, thermal design, dispensing and ground operations are all being worked out in real time, on orbit, under real conditions. That is the point. Nobody buys VLEO capability off a spec sheet anymore. They buy flight evidence: proof, generated in orbit, that it actually works.
The bottleneck has moved
For a long time, the hard part of space was building the technology itself: propulsion that fires reliably, power systems that survive eclipse cycles, avionics that don’t brick themselves on day one. That is still genuinely hard, and we dug into it in our look at power, propulsion and orbital compute as spacecraft’s persistent subsystem bottleneck . But for most companies that already have a working prototype, subsystem invention is no longer the binding constraint. The binding constraint now is proving, repeatably and in a way that can be audited later, that the technology performs in the actual operating environment rather than a lab chamber or a simulation.
DiskSat’s mission log is a good illustration of what that proof actually costs. The challenges Aerospace Corporation is logging, power, comms, attitude control, thermal, dispensing, ground ops, are exactly the kind of subsystem-level friction that only shows up once hardware is flying. VLEO is not just a lower altitude on a mission plan. It is a coordinated spacecraft, propulsion and operations architecture, and every piece of that architecture has to earn its place through flight data, not a datasheet.
Government capital is now structured around this gap
That evidence-generation gap, the space between a working prototype and a program of record, is exactly what SpaceWERX’s STRATFI program exists to fund. In its 2026 cycle, STRATFI matched public and private funding into $562.5 million of total awards , including $245 million from private capital, spread across 11 companies working in propulsion, nuclear power, space-domain awareness, thermal-infrared Earth observation (EO), debris removal and orbital logistics. The program has previously backed companies including Impulse Space, K2 Space, Apex, True Anomaly and Astranis, though the 2026 announcement itself did not name which of those six sectors each of this cycle’s 11 recipients falls into.
What matters about the structure is the matching requirement. The government is not simply subsidizing research and development, it is requiring private capital to co-sign the bet, which functions as a second layer of validation stacked on top of whatever technical proof the company has already produced. A similar logic showed up this month when ThinkOrbital secured a $4.6 million US Air Force award . Funding alone doesn’t get a new capability adopted. Someone still has to prove it holds up in the actual operating environment, then hand a mission owner something they can point to and reuse, rather than take on faith.
This is where in-orbit validation fits into the stack: not a nice-to-have demo flight, but the evidence-generation layer between prototype and procurement. Validate a payload or a software capability on orbit, document how it actually performed, and hand a government buyer something closer to a proven module than a pitch deck.
Owning the whole stack compresses time-to-proof
India’s OrbitAID Aerospace offers a clean example of how much faster that evidence can be generated when a company controls more of the mission stack itself. OrbitAID’s planned Q1 2027 demonstration will use a chaser and target spacecraft to inspect, approach, dock with and raise the target’s orbit, in low Earth orbit (LEO). The company built its own docking interface and its own autonomous rendezvous-and-proximity-operations (RPO) software, and says it reached launch readiness in 11 months by keeping technology development and supply chains largely in-house. A follow-on geostationary orbit (GEO) life-extension mission is already under discussion.
Eleven months from concept to launch readiness is fast for a mission carrying this much custom hardware and software. It is fast in large part because OrbitAID is not stitching together a docking system from one vendor, RPO software from another, and mission operations from a third, then hoping the integration seams hold under real conditions. It is the same logic driving standardization elsewhere in the industry: as I’ve written in the context of the Earth-observation market’s own shift toward standardized, catalog-style capability , the fastest way to compress a validation timeline is to stop treating every mission as a one-off bespoke build.
For a company running in-orbit validation as a service, OrbitAID is a useful benchmark less because of the RPO technology itself and more because of what a whole-stack approach produces. Customers do not just want a payload slot. They want the in-orbit result, a complete answer to whether the thing actually works, as the product.
Flight evidence isn’t just a hardware concept anymore
The DiskSat and OrbitAID stories both live in hardware: propulsion, docking mechanisms, power systems. But the same evidentiary standard is now showing up in software. Shield AI, Sedaro and NOVI Space recently completed an in-orbit demonstration of AI-supported autonomous spacecraft operations , in which the autonomy made and executed operational decisions with onboard validation.
That is a meaningfully different claim than saying software is “space-ready.” It is evidence that autonomous decision-making, with the actual decision boundaries and validation logic flying rather than sitting in a simulator, held up in the real operating environment. Autonomous operations are going to need exactly this before government and commercial operators trust them at scale: flight evidence, bounded decision authority and behavior that can be audited after the fact. It also points at a fairly natural opportunity for repeatable in-orbit validation. Mission software can ride alongside hardware payloads on the same flight, generating telemetry and operational outcomes the same way a propulsion system or a docking mechanism would, and successful autonomy functions can graduate into standard, reusable modules the same way a proven thruster does.
Prove once, deploy everywhere
Line up DiskSat, STRATFI, OrbitAID and the Shield AI, Sedaro and NOVI Space demonstration, and a pattern falls out. Flight evidence is turning into a general discipline that spans hardware and software, not a category that only applies to propulsion systems and solar arrays. The industry’s real constraint is not inventing new technology anymore. Plenty of good technology already exists. The constraint is proving that the technology works in orbit, repeatably enough and with enough documentation that the next buyer does not have to re-run the experiment from scratch.
That is the shift in-orbit validation is built around: prove it once, in a mission that produces real telemetry and a real operational record, then make that proof reusable instead of bespoke to a single customer or a single spacecraft.
What to watch
A few threads worth tracking as this plays out over the next few quarters:
- DiskSat’s descent into VLEO. Whether Enpulsion’s electric propulsion can take the four DiskSats from a slow atmospheric descent into controlled, sustained VLEO operation below 300 kilometers.
- STRATFI’s 2026 cohort. Which of this cycle’s 11 awardees move from funded prototype work into actual flight demonstrations, or further, into Phase III procurement.
- OrbitAID’s Q1 2027 demonstration. Whether the chaser-target RPO mission flies on schedule, and whether the company’s claimed LEO-to-GEO commercial pathway turns into a contracted mission.
- Trusted autonomy, beyond the simulator. Whether more operators follow Shield AI, Sedaro and NOVI Space’s lead in demonstrating auditable onboard decision-making, rather than autonomy that has only ever been proven in ground-based simulation.
This is exactly the gap I started SATELYX to fill: the space between a working prototype and a mission owner willing to bet on it, filled with real flight evidence instead of a pitch deck. More at satelyx.com .