For as long as humanity has sent machines into the void, those machines have thought slowly — constrained by the same radiation-hardened chips that kept them alive but left them dependent on distant human minds. Now, at NASA's Jet Propulsion Laboratory, a processor one hundred times more powerful than its predecessors is being tested against the full hostility of space, carrying with it the promise that future spacecraft may finally possess something resembling autonomous judgment. The stakes are not merely technical: when a rover on Mars faces a crisis and Earth is twenty-four minutes away, t
NASA Tests Radiation-Proof Chip That Brings AI to Deep Space
Spacecraft can finally think and act in the moment
Mark
Why does a spacecraft need AI at all? Can't engineers on Earth just send commands?
Mimi
Because of distance. A signal to Mars takes up to 12 minutes to arrive. By the time you get a response back, 24 minutes have passed. If a rover encounters a boulder or a system starts failing, waiting that long could mean losing the mission.
Mark
So this chip lets the spacecraft make decisions on its own?
Mimi
Exactly. It can run AI models that recognize problems and respond immediately. A rover can navigate around obstacles. A satellite can fix errors in its own systems before they cascade into failure.
Mark
What makes this chip different from the ones already in space?
Mimi
Speed, mainly. Current space processors are built to survive radiation, but they're based on old designs. This one is 100 times faster while still being radiation-hardened. It's the first time NASA has had both at once.
Mark
How do you make a chip that can handle radiation?
Mimi
You design it to detect when radiation corrupts data and fix it automatically. You use materials and layouts that are less vulnerable to particle strikes. It's not about immunity—it's about resilience.
Mark
Will this change how we explore space?
Mimi
Fundamentally. Right now, rovers and spacecraft are tethered to Earth by communication delays. This gives them real autonomy. Crewed missions to Mars become safer because the spacecraft can handle emergencies without waiting for help.
Il Polso
- Decades of deep-space computing have been shackled to ancient, radiation-hardened chips — durable enough to survive space, but far too slow to think for themselves in a crisis.
- Communication delays of up to 24 minutes each way make human intervention during emergencies on Mars or the Moon not just impractical, but physically impossible.
- NASA and Microchip Technology are now stress-testing a palm-sized system-on-chip that packs processing, memory, and networking into one unit capable of running real-time AI — subjected to radiation bombardment, violent temperature swings, and electromagnetic assault at JPL.
- The chip must prove it can detect and self-correct data corruption from solar particles on the fly, or risk leaving a spacecraft frozen and helpless at the worst possible moment.
- If certified, the processor will grant rovers, satellites, and crewed mission systems the ability to recognize hazards, respond to anomalies, and manage emergencies without waiting for Earth — a quiet but profound transfer of agency from human to machine.
For as long as humanity has sent machines into the void, those machines have thought slowly — constrained by the same radiation-hardened chips that kept them alive but left them dependent on distant human minds. Now, at NASA's Jet Propulsion Laboratory, a processor one hundred times more powerful than its predecessors is being tested against the full hostility of space, carrying with it the promise that future spacecraft may finally possess something resembling autonomous judgment. The stakes are not merely technical: when a rover on Mars faces a crisis and Earth is twenty-four minutes away, the machine must be wise enough to act alone.
For decades, the computers aboard NASA spacecraft have been a paradox: built to survive the most hostile environment humanity has ever explored, yet so underpowered by modern standards that they could barely think for themselves. The problem was never ambition — it was physics. Radiation, temperature extremes, and the vacuum of space demand chips engineered for endurance, not speed. And so, as smartphones grew exponentially smarter, the machines we sent to Mars and beyond continued to rely on processors designed generations ago.
That constraint is now being directly challenged. Under NASA's High Performance Spaceflight Computing program, Microchip Technology Inc. has developed a system-on-chip that delivers roughly one hundred times the computational power of current deep-space processors. Small enough to rest in a palm, it integrates a processing unit, memory, networking hardware, and input-output interfaces into a single component — the same architectural philosophy that powers consumer devices like iPhones, now adapted for the radiation and vacuum of space.
Engineers at NASA's Jet Propulsion Laboratory in California are currently subjecting the chip to everything space will throw at it: violent temperature swings, intense vibration, and the high-energy solar particles that can silently corrupt data or trigger shutdown sequences that leave a spacecraft adrift and unresponsive. The processor must prove it can catch and correct these errors in real time, without missing a beat.
The implications reach far beyond raw speed. With this level of onboard computing, spacecraft can run artificial intelligence models autonomously — a Mars rover could read terrain and reroute itself without waiting for Earth's guidance; a satellite could diagnose and respond to its own failures; crewed missions could lean on onboard AI when astronauts are out of contact with mission control. Program manager Eugene Swanback at NASA's Langley Research Center described the achievement as the culmination of decades of work in fault-tolerant computing — a legacy honored, with quiet wit, when the team sent their first test message with the subject line 'Hello Universe,' a nod to the 'Hello World' tradition that has greeted every new era of programming.
Once certified, the processor will make its way into satellites, planetary rovers, and deep-space probes. The same design principles are already being adapted by Microchip for automotive and aviation systems on Earth — places where reliability and performance are equally non-negotiable. A technology forged in the extremity of space may, in time, quietly improve the systems that carry us through our own world.
For decades, spacecraft have operated under a constraint that seems almost quaint in the age of smartphones and cloud computing: the computers running them are built on processors that are decades old. NASA has relied on chips designed to survive the radiation and temperature extremes of space, but this durability came at a steep cost in raw power. A spacecraft bound for Mars cannot simply wait for an engineer on Earth to solve a problem. The communication delay alone—up to 24 minutes round-trip—makes real-time human control impossible. Now NASA is testing a processor that could finally close that gap.
The new chip, developed by Microchip Technology Inc. under NASA's High Performance Spaceflight Computing program, delivers roughly 100 times the computational power of the processors currently flying on deep-space missions. It is small enough to hold in your hand, yet it integrates everything a spacecraft needs into a single component: a processing unit, memory banks, networking hardware, and input-output interfaces. This architecture, called a system-on-chip, is already familiar from consumer electronics—it's what powers iPhones and MacBooks. Now it's being adapted for the vacuum and radiation of space.
The testing is underway at NASA's Jet Propulsion Laboratory in California, where engineers are subjecting the chip to the kinds of punishment it will face in orbit and beyond. Extreme temperature swings, intense vibrations, high levels of electromagnetic radiation—all must be survived without degradation. The real threat comes from solar particles, high-energy projectiles that can corrupt data and trigger automatic shutdown sequences that leave a spacecraft helpless. The new processor must prove it can detect and correct these errors on the fly, maintaining operation even under bombardment.
The implications are profound. With this kind of processing power, spacecraft can run artificial intelligence models in real time. A rover on Mars could analyze terrain, recognize hazards, and adjust its path without waiting for instructions from Earth. A satellite could detect and respond to anomalies in its own systems. Future crewed missions to the Moon and Mars could rely on onboard AI to handle emergencies when astronauts cannot reach mission control. The technology essentially grants spacecraft a form of autonomous judgment—not consciousness, but the ability to think and act in the moment.
Eugene Swanback, the program manager at NASA's Langley Research Center, framed the achievement in terms of institutional legacy. The new system builds on decades of work in space-hardened computing, he said, combining fault tolerance and flexibility with performance that was simply unavailable before. The testing team marked the milestone with a small gesture toward computing history: they sent the first test message with the subject line "Hello Universe," echoing the "Hello World" program that became the standard introduction to computer programming.
Once NASA certifies the processor, it will begin appearing in satellites, planetary rovers, and deep-space probes. The same underlying technology is already being adapted by Microchip for use in terrestrial applications—automotive systems, aviation electronics, anywhere on Earth where reliability and performance matter equally. The radiation-hardened version will remain exclusive to space, but the architecture and design principles will ripple outward. A technology born from the constraints of deep space may end up improving the systems we depend on here at home.
Citazioni salienti
Building on the legacy of previous space processors, this new multi-core system is fault-tolerant, flexible, and extremely high-performance.— Eugene Swanback, program manager at NASA's Langley Research Center
Vuoi la storia completa?
Leggi l'originale su tovima.com ↗
Technology