The subject line read “Hello Universe.” In February 2026, engineers at NASA’s Jet Propulsion Laboratory powered up a new spaceflight processor for the first time and sent that two-word message to colleagues, echoing the tradition programmers have followed since the earliest days of computing. But the chip behind that message is anything but traditional. Known as the High Performance Spaceflight Computing processor, or HPSC, it delivered roughly 500 times the performance of the radiation-hardened processors currently flying on NASA spacecraft, according to the agency’s announcement of the milestone.
If that number holds up through further testing, the implications reach far beyond faster number-crunching. A probe orbiting Jupiter or descending through Titan’s orange haze cannot phone home for quick instructions. Radio signals crawl at the speed of light, and round-trip communication with Jupiter takes roughly 35 to 52 minutes depending on orbital positions. With Saturn, the wait stretches past an hour. A spacecraft equipped with HPSC could, in theory, spot a geyser erupting from an icy moon, redirect its cameras, and capture the event before a ground controller on Earth even knows it happened.
What the chip actually is
The HPSC is a radiation-hardened, multicore, 64-bit system-on-chip designed from the ground up for the punishing environment of deep space. It includes eight processing cores and a built-in 240 Gbps Ethernet switch built around Time-Sensitive Networking standards, which essentially gives the chip a high-speed internal highway for shuttling data between onboard instruments, cameras, and communications systems.
Its architecture is also built to handle artificial intelligence and vector computing workloads. In plain terms, that means a future spacecraft could run machine-learning models onboard to classify terrain in real time, identify atmospheric signatures worth investigating, or decide which sliver of science data is most valuable to send home over a narrow communications link, rather than transmitting everything and letting scientists on Earth sort through it later.
To build the processor, JPL awarded a $50 million firm-fixed-price contract to Microchip Technology Inc. Microchip is also investing its own research and development money in the effort, a sign the company sees a commercial market for radiation-hardened multicore chips beyond NASA’s immediate needs. The contract describes HPSC as broadly applicable to future missions spanning Earth science, planetary exploration, and deep-space observation.
Why current space computers are so far behind
The performance gap sounds dramatic, but it makes sense once you understand how old most spaceflight processors are. The workhorse of NASA’s current fleet is the RAD750, a radiation-hardened chip whose design traces back to the early 2000s. It first flew in 2005 and has powered missions from the Mars Reconnaissance Orbiter to the James Webb Space Telescope’s ground systems. The RAD750 was built to survive cosmic rays and solar particle storms, not to win speed contests, and its processing power is roughly comparable to a mid-1990s desktop computer.
That trade-off between reliability and performance has defined spaceflight computing for decades. Radiation can flip bits in a processor’s memory, corrupt calculations, or destroy transistors outright. Hardening a chip against those effects takes years of design work and specialized manufacturing, which means space processors almost always lag several generations behind their commercial counterparts. The HPSC is NASA’s attempt to close that gap without sacrificing the radiation tolerance that keeps missions alive.
The numbers that still need scrutiny
NASA’s own materials present two different performance comparisons. The February 2026 testing announcement cites “approximately 500 times” the capability of current space processors. A separate HPSC program page describes the chip as delivering “over 100 times” the computing capability. The difference likely reflects distinct benchmarks, workloads, or comparison baselines, but NASA has not publicly reconciled the two figures.
Nor has the agency specified exactly which existing processor serves as the yardstick for either claim. The RAD750 is the most obvious candidate, but no published data from the February tests includes a head-to-head comparison against it or any other specific chip. Independent benchmarks, academic papers, or third-party engineering assessments of the HPSC have not appeared in publicly available literature as of May 2026.
That said, the sources behind the claims carry weight. JPL and NASA’s Goddard Space Flight Center, which jointly describe the chip’s architecture in technical documentation from Goddard’s Engineering and Technology Directorate, have long track records of publishing accurate hardware specifications. The performance leap is also plausible on its face: a modern multicore 64-bit SoC built with current semiconductor techniques would naturally outperform a chip designed two decades ago by large multiples.
Critical tests still ahead
As of March 2026, NASA reported that the HPSC was undergoing additional testing focused on power consumption and performance validation. Those results have not been released, and they matter enormously. Deep-space missions operate on tight power budgets. A probe beyond the asteroid belt receives a fraction of the solar energy available near Earth, and every watt consumed by the processor is a watt unavailable for instruments, heaters, or communications. If the HPSC draws too much power at full capacity, mission designers may have to throttle it, erasing some of the performance advantage.
Thermal management poses a parallel challenge. In the vacuum of space, a processor cannot shed heat through convection the way a laptop fan does on Earth. Heat must radiate away, a slower process that constrains how hard a chip can work continuously. Fault-tolerance data, another essential metric for any radiation-hardened component, also remains absent from public disclosures.
No specific mission has been named as the first flight opportunity for HPSC. NASA’s contract and program documents describe the chip in general terms as supporting future missions, but they do not identify a launch date, a spacecraft integration timeline, or a target destination. Missions in NASA’s pipeline that could benefit from this kind of onboard computing power include Dragonfly, the rotorcraft lander headed for Titan with a planned launch in 2028, and future flagship-class missions to the outer solar system. But no public statement has linked HPSC to any of them.
What it would take to change exploration
The practical promise of HPSC is not raw speed for its own sake. It is autonomy. Every major discovery in planetary science involves some element of luck and timing: Cassini flying through a plume jetting from Enceladus, Curiosity spotting unusual rock formations on Mars. In each case, the spacecraft captured something unexpected, but the science team on Earth had to wait for data to arrive, analyze it, and then send new commands back. That cycle can take hours or days.
A spacecraft with enough onboard intelligence to recognize what it is seeing and react in real time could compress that cycle to seconds. It could prioritize a fleeting event over a routine observation, reallocate instrument time on the fly, and deliver higher-value science data per bit transmitted. That is the vision NASA has outlined for HPSC, and the 500-times performance figure is the quantitative foundation for it.
Three milestones will determine whether that vision becomes operational reality: publication of independent benchmark comparisons against current flight processors, disclosure of power and thermal performance under simulated space conditions, and the naming of a specific mission that will carry the HPSC. Until those arrive, the chip represents a significant engineering achievement in the lab. What it means for actual exploration is a story still being written.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
