By: Paul Tilghman, Chief Technology Officer
In the 1970s, NASA introduced the Technology Readiness Level (TRL) scale, now the universal yardstick for engineering maturity. But TRL reflects an era of bespoke invention. A TRL 9 system may be flight-proven, yet the scale makes no distinction between a one-off prototype and a product that can be manufactured at scale.
To address this gap, NASA introduced the Commercial Readiness Level (CRL) in the early 2010s. While an important step, CRL still evaluates individual technologies in isolation. Neither TRL nor CRL captures supply-chain depth, second and third sources, private-capital viability and demand beyond government subsidies.
What we need is a metric that measures the market itself.
The six-level-scale Commercial Readiness Index (CRI) is exactly such a metric, where CRI 1 reflects a mature technology with no market viability, CRI 3 marks commercial scale-up and CRI 6 represents a mature market.
While elements such as launch have reached high maturity with growing competition, the space economy as a whole – low-Earth orbit in particular – is at a critical inflection point. Today, the space economy is at CRI 3.
AI, particularly Agentic AI, is the last piece of the technology puzzle necessary to form the foundation of a durable CRI 6 space economy. To kick off 2026, I’ll highlight five crucial next steps, highlighting AI itself and the infrastructure required to support it.
- We’ll see the first foundational Agentic AI for space emerge.
- Agentic AI enables astronauts to act as orchestrators rather than operators, overseeing a potentially massive number of complex machines that execute autonomously in LEO, on the Moon and eventually Mars.
- The communication delay from the Moon to Earth (~2.5s round trip) is manageable. But on Mars, 20+ minute communication delays make Earth-based control impractical. This is where agents effectively bring mission control with the mission.
- The same agentic autonomy is exactly what’s required for space traffic management, a growing necessity for orbital safety, and will be foundational to operating spacecraft supporting Lunar and Martian economies.
- Agentic engineering will scale spaceborne manufacturing, science and research efficacy.
- Spaceborne manufacturing, science and research, and associated novel organic and inorganic materials, is seen by many as a foundational pillar of a CRI 6 space economy.
- Even with an emerging crop of science-as-a-service providers, available demand for non-terrestrial discovery exceeds supply.
- Earthbound GPU-accelerated scientific discovery models will be extended to incorporate the microgravity environment, and in combination with agentic engineering tools, improve the efficacy of non-terrestrial science.
AI will only scale the space economy if it is anchored by spaceborne infrastructure that makes intelligence operational in orbit, not just on the ground. Low-power processors will give way to orbital data centers (ODCs), whether as distributed constellations that function as a virtual edge cloud or, eventually, as centralized hyperscale platforms.
- We move beyond the traditional “rad-hard or nothing” paradigm.
- Radiation is unavoidable but no longer defining. Its impact is increasingly managed at the system level rather than bespoke, low-performance hardware. A compute-driven space economy demands both capability and survivability, not a tradeoff.
- Progress is converging on three fronts: advanced shielding that allows terrestrial-grade processors to operate in select orbits; maturing open architectures like RISC-V that embed radiation tolerance directly into logic without costly licensing; and software-driven resilience (containerization, fault isolation and automated recovery) adapted from terrestrial data centers to tolerate transient faults in space.
- The result is a fundamental shift: reliability is defined by system-level resilience, not individual components, enabling scalable orbital computing and autonomous operations.
- The industry pivots toward novel thermal management.
- Space is often assumed to be ideal for cooling, but it’s not cold. It’s empty and requires a thoughtful thermal design.
- As orbital AI grows, thermal management becomes a first-order design problem, with today’s solutions far too large to scale.
- The industry must pivot towards low-cost advanced heat pipes, active fluid loops and high-emissivity materials that make scalable cooling possible. Without them, hyperscale computing in orbit will fail to scale to meet the need of the space economy.
- We see the rise of “third-wave” optical terminals.
- ODCs (especially disaggregated ODCs) require fast, flexible links between nodes, but today’s laser communications can take 10s to low 100s of seconds to establish. This is far too slow for dynamic, multi-constellation networks.
- As carrier hotels, or space-based facilities interconnecting multiple clouds, emerge in medium-Earth orbit, links must be made and broken continuously across heterogeneous systems. Third-wave optical terminals replace old-school mechanical gimbals with non-mechanical beam steering, enabling millisecond target switching through the equivalent of optical phased arrays. The result is a shift from fixed optical pipes to a true dynamic heterogenous network of networks in space.
The capabilities emerging in 2026 – agentic autonomy, orbital computing, high-speed optical networking, scalable thermal systems and system-level radiation resilience – are not incremental upgrades. They are the enabling infrastructure of a self-sustaining space economy. While many of these technologies remain immature by traditional TRL measures, they are precisely the capabilities required to move the industry from government-anchored experimentation to durable commercial scale.
The defining question for the next decade is no longer whether space technologies can work, but whether markets can form, compete and endure. By that measure, 2026 stands as the inflection year when the space economy begins its transition toward CRI 6.