- IoT For All
- Posts
- How AI Went From Orbit to the Assembly Line
How AI Went From Orbit to the Assembly Line
And the week's top AI news!
AI For All
September 12, 2025
Hello readers,
Welcome to the AI For All newsletter! Today, we’ll be exploring how an AI approach has tackled high-frequency trading, solar storms and predictive maintenance, and more!
AI in Action: Orbit to Assembly Line
As solar storms and auroras captured public attention in 2024, NASA’s Frontier Development Lab (FDL) was quietly working with AI firms to solve the space weather challenges behind those spectacular skies. In partnership with KX Systems, FDL applied machine learning to real-time data from the ionosphere, solar activity, and Earth’s magnetic field to predict disruptions to GPS and satellite systems up to 24 hours in advance. Their AI approach—originally built for high-frequency trading in finance—was adapted to search for subtle patterns in solar interference that human analysts would miss.
The collaboration between NASA and KX Systems is a clear example of how AI can translate across domains. By retraining their kdb+ analytics engine to process scientific satellite data instead of market movements, the company expanded its software’s capabilities into predictive maintenance, anomaly detection, and early warning systems. This same pattern recognition technology, once used to forecast GPS outages, is now being deployed commercially to monitor and optimize industrial equipment in manufacturing environments.
The partnership is a case study in what happens when public and private innovation ecosystems collide. With NASA’s scientific rigor and KX’s real-time data expertise, the result was a better way to forecast solar storms, but also a commercial-grade AI toolkit with applications far beyond space. These kinds of collaborations are what make AI progress practical, not just theoretical.
🔥 Rapid Fire
NVIDIA to pay $1.5 billion to rent back its own chips?
OpenAI says its business will burn $115 billion through 2029
Goldman Sachs: inevitable AI slowdown could tank S&P 500
Microsoft’s AI chief says machine consciousness is an illusion
OpenAI executives rattled by campaigns to derail for-profit conversion
How thousands of humans train Google’s AI to seem smart
80% of ransomware attacks now use artificial intelligence
How AI and politics hampered the secure open source movement
Patients turn to AI to interpret lab tests, with mixed results
Research shows humans are far superior to AI on this business task
Publishers fear AI summaries are hitting online traffic
California’s latest attempt at AI regulation is inching toward passage
Light-powered chip could make AI 100x more efficient
FTC launches inquiry into AI chatbots of major tech firms
Albania appoints AI bot as minister to tackle corruption
The question all colleges should ask themselves about AI
📖 What We’re Reading
“This year, more manufacturers are scaling AI across the enterprise: compared to 2024, three times as many report scaling 50 percent or more of their AI pilot projects. But while adoption accelerates, measuring impact is still lagging for many organizations. Machine health remains the most common AI use case, yet it ranks seventh when it comes to quantifying impact.
Confidence in AI is high—83 percent say they’re advanced or very advanced in applying it—but many of its most promising applications are still underleveraged or poorly understood. U.S. manufacturers trail European peers in leveraging AI for asset care, and ecosystem fragmentation continues to slow progress.”