Jyunmi Hatcher

So that's the through line.

The autonomy on Orion is the conservative end of the spectrum, and the AI and perseverance is the experimental end.

Both are being developed by the same agency for the same long-term reason.

Missions are getting farther, longer, and more complex.

The spacecraft has to be able to handle more of its own operations.

Now,

According to Fortune Business Insights, AI in space operations market was valued at 2.3 billion in 2025, and it's projected to reach roughly 15 billion by 2034.

The Brookings Institution, citing McKinsey, estimated the broader economy at $61.3 billion in 2024 and projected it could reach $1.8 trillion by 2035.

The number of satellites in orbit is approaching 15,000 and is expected to reach 100,000 by 2030.

NASA's Earth Observation Archive alone has surpassed 100 petabytes.

At those volumes, no ground team can keep pace without machine learning doing a substantial portion of the analysis.

The honest caveat here is that the AI in this context covers a wide range of capabilities, and not all of them are operating at the same level of maturity.

Orion's autonomy is a largely classical control engineering and deterministic algorithms running on radiation-hardened hardware.

The Perseverance demonstration uses a generative AI model, but only for route planning, only on two specific drives, and only after output had been checked against the digital twin and 500,000 telemetry variables.

The NASA administrator, Jared Isaacman, in the JPL press release framed it carefully.

He says, this demonstration shows how far our capabilities have advanced and broadens how we will explore other worlds.

It's a strong example of teams applying new technology carefully and responsibly in real operations.

Keyword here is carefully.

Nothing about the way NASA is integrating AI into deep space missions resembles consumer grade AI deployment.

The cost of a wrong answer is too high.