3 Sources
[1]
Intel's new space-grade Starfire chip is a Panther Lake SoC that puts an 18A CPU into orbit -- chip designed for the US government leverages Intel 3 for the GPU
Intel has unveiled Starfire, a space-grade system-on-chip designed for the U.S. government that pairs eight CPU cores and a three-tile NPU built on its Intel 18A node with an Intel 3 graphics tile, all in one Foveros package. Intel published the Starfire sell sheet, listing two versions that draw 10 W and 35 W and reach up to 45 and 75 TOPS, respectively, rated to run between -55 and 125 Celsius. Both SKUs share the same layout of four Intel 18A P-cores, four low-power efficiency cores, a three-tile NPU also on 18A, and a four-core Xe GPU with 64 execution units built on Intel 3. The Low Power part runs its P-cores at 1.0 GHz, efficiency cores at 850 MHz, and the GPU between 800 MHz and 1.0 GHz. The Performance part clocks the P-cores to 3.1 GHz, efficiency cores to 2.1 GHz, and the GPU to 2.0 GHz. Both carry 12 PCIe Gen4 lanes, support LPDDR5 or DDR5, and are rated for a 10-plus year lifetime. Intel builds the CPU and NPU on 18A and the GPU on the older Intel 3, the same node division it used for Clearwater Forest, the 288-core Xeon that stacks 18A compute tiles on Intel 3 base tiles. Smaller transistors hold less charge per stored bit, which makes leading-edge silicon more prone to radiation-induced bit flips, so committing 18A to orbit leans on RibbonFET and design-level hardening rather than a mature, inherently more tolerant node. The market Starfire is targeting has run on BAE Systems' RAD750 for two decades. That radiation-hardened PowerPC part clocks 110 to 200 MHz, carries 10.4 million transistors, and is built on 150nm or 250nm lithography, per public specifications, and it flies on the Mars rovers, Kepler, and Fermi, among more than 150 spacecraft. BAE's multi-core RAD5545 and the Microchip-built processor NASA is developing to reach 100 times the throughput of current spaceflight chips are the more recent step up. Starfire's up to 75 TOPS and dedicated NPU put it in a different bracket, built for on-orbit AI inference rather than telemetry and control. Intel lists the radiation data, covering total ionizing dose, single-event latch-up, and single-event effects, as characterization in process, so the part isn't radiation-qualified yet, and it notes the specs are subject to change. Intel Government Technologies is handling Starfire, with samples in Q3 2026 and a pitch of market-competitive pricing and domestic manufacturing. Intel Foundry is the only U.S.-based maker of leading-edge logic, holds Trusted Foundry status, and has tied its 18A and packaging roadmap to Pentagon programs including RAMP-C and SHIP, though 18A yields aren't expected to reach industry-standard levels until 2027. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.
[2]
Intel's Starfire SoC combines 18A CPU, Intel 3 GPU, and NPU for space-grade computing
Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. Forward-looking: Intel is bringing its 18A process node to space with two Starfire chips designed for onboard AI processing and computing in harsh environments. Intel lists two SoC variants: one optimized for low-power applications and another focused on higher performance. If everything goes smoothly, the company expects to deliver samples in the third quarter. Starfire features an eight-core CPU built on the 18A node, with each chip combining four performance cores and four low-power efficiency cores. The two models operate at different clock speeds, with the low-power version prioritizing efficiency and the performance model targeting higher compute capability. Performance cores run at 1.0 GHz on the low power model while the efficiency cores operate at 850 MHz, keeping the entire package within a 10W TDP. The performance version increases clock speeds, with the four performance cores running at 3.1 GHz and the efficiency cores reaching 2.1 GHz. It carries a 35 W TDP for systems capable of supporting higher power consumption. The chips also include a dedicated graphics tile. Both versions feature a GPU tile built on Intel 3, with four Xe cores and a total of 64 execution units. On the low power model, the Xe cores operate between 800 MHz and 1.0 GHz. The performance model increases that frequency to up to 2.0 GHz, providing additional headroom for graphics and compute workloads in orbit. Both variants also feature an NPU. The low power Starfire delivers up to 45 TOPS of INT8 AI performance, while the performance version reaches up to 75 TOPS. Combined, the CPU, GPU, and NPU in a single space-qualified chip allow satellites and other systems to process more data and run AI workloads onboard, reducing reliance on ground-based processing. The operating limits are designed for space. Starfire is rated to operate at junction temperatures ranging from -55°C to 125°C. The documentation notes that the chips carry radiation-hardening certifications for total ionizing dose, single-event latch-up, and single-event effects. These measures are intended to reduce damage from long-term radiation exposure and short bursts of high-energy particles. Radiation is a major challenge for any processor deployed in space. Ionizing radiation can pass through a chip and flip bits in memory or logic circuits, potentially corrupting data or causing errors. Over time, accumulated radiation exposure can degrade devices, while single events can trigger immediate faults. For this reason, space-grade chips undergo radiation hardening and qualification before deployment, including processors derived from commercial designs. By integrating an 18A CPU, an NPU, and an Intel 3 GPU into a single SoC, Intel is bringing newer process technology to systems that have traditionally relied on older nodes. Intel is targeting Starfire at satellites and other spacecraft that require real-time data processing, onboard AI capabilities, and strict size, weight, and power constraints.
[3]
Intel Brings 18A Silicon To Orbit With Starfire, A Space-Grade SoC Rated For 125°C And Radiation
Intel has unveiled its latest SoCs, codenamed Starfire, that offer space-grade survivability and are based on the 18A process technology. Intel Rolls Out Space-Grade "Starfire" Chips With 8 Cores, 18A Tech, & 10+ Years of Lifetime Space-grade platforms require extra engineering as they are built to withstand various aspects of space, such as exposure to radiation and extreme temperatures. Plus, these chips need to offer utmost levels of stability as space systems need to retain a steady operation throughout their life-cycle. As such, Intel has new space-grade chips ready for the market. Codenamed Starfire, the SoCs are manufactured in the United States for the US Government, and Intel claims that they offer market-competitive pricing. Some of the main highlights that Intel Starfire SoCs have on offer include: * Space-Grade Survivability * Low-Size, Weight, Power * Advanced AI Performance * Multi-chip Foveros Package The Intel Starfire SoCs will come in two flavors, a low-power SKU and a Performance SKU. Both SKUs feature the same 8-core configuration with 4 P-Cores and 4 LPE cores. The clock speeds for the Low-Power SKU are rated at 1 GHz (P-Cores) and 850 MHz (LPE Cores), while the Performance SKU is rated at 3.1 GHz (P-Cores) and 2.1 GHz (LPE Cores). These chips are essentially based on the Panther Lake 4Xe3 SKU. For the iGPU, both SKUs feature 4 Xe3 iGPU cores, which are once again clocked at up to 1.0 GHz for the Low-Power and 2.0 GHz for the Performance variant. The NPU is also based on the Intel 18A node, while the iGPU is based on the Intel 3 node. The whole chip offers up to 45 TOPs on the Low-Power SKU and up to 75 TOPs on the Performance option. The Low-Power chip is rated at 10 Watts while the Performance chip is rated at 35W. Besides the core configuration, the chips offer radiation protection such as TID, SEL, and SEE, operate between -55 °C and 125 °C, offer 12 PCIe Gen4 lanes, support LPDDR5/DDR5 memory, and come backed by a 10+ year warranty. News Source: @x86deadandback Follow Wccftech on Google to get more of our news coverage in your feeds.
Share
Copy Link
Intel has unveiled Starfire, a space-grade system-on-chip designed for the US government that combines an eight-core CPU and NPU built on its 18A process with an Intel 3 GPU. The chip delivers up to 75 TOPS for on-orbit AI inference while operating in temperatures from -55°C to 125°C. Samples are expected in Q3 2026, positioning Intel to compete in a market dominated by older processors like BAE Systems' RAD750.
Intel has introduced the Intel Starfire SoC, a space-grade system-on-chip designed specifically for the US government that marks a significant leap in onboard AI processing for spacecraft and satellites
1
. The chip combines an Intel 18A CPU with an Intel 3 GPU and a dedicated NPU, all integrated into a single Foveros package2
. This architecture enables the chip to deliver up to 75 TOPS of AI performance on the performance variant, positioning it for on-orbit AI inference rather than traditional telemetry and control tasks1
.Source: TechSpot
The Intel Starfire SoC comes in two configurations: a low-power variant drawing 10W and a performance variant consuming 35W
3
. Both versions feature eight cores, split between four P-Cores and four LPE cores. The low-power model runs its P-Cores at 1.0 GHz and efficiency cores at 850 MHz, delivering up to 45 TOPS of INT8 AI performance2
. The performance version clocks the P-Cores to 3.1 GHz and efficiency cores to 2.1 GHz, reaching up to 75 TOPS1
.Intel built the CPU and NPU on its cutting-edge 18A process, while the GPU uses the older Intel 3 node—the same node division employed for Clearwater Forest, Intel's 288-core Xeon processor
1
. This decision to deploy the 18A process in space is notable because smaller transistors hold less charge per stored bit, making leading-edge silicon more susceptible to radiation-induced bit flips . Intel is relying on RibbonFET technology and design-level hardening rather than using a more mature, inherently radiation-tolerant node.Both Starfire variants feature a Xe3 iGPU with four cores and 64 execution units. The low-power model operates the GPU between 800 MHz and 1.0 GHz, while the performance version increases that frequency to 2.0 GHz
2
. The chips are designed for extreme temperature operation, rated to function at junction temperatures ranging from -55°C to 125°C3
.The market for space-grade computing has been dominated by BAE Systems RAD750, a radiation-hardened PowerPC processor that clocks between 110 to 200 MHz and carries 10.4 million transistors built on 150nm or 250nm lithography
1
. The RAD750 has flown on over 150 spacecraft, including the Mars rovers, Kepler, and Fermi, for two decades. Starfire's up to 75 TOPS and dedicated NPU place it in a different performance bracket entirely, built for sophisticated AI workloads rather than basic telemetry.Both Starfire SKUs support 12 PCIe Gen4 lanes and LPDDR5 or DDR5 memory, with a rated lifetime exceeding 10 years
1
. The radiation-hardened chip includes protection for total ionizing dose, single-event latch-up, and single-event effects, though Intel notes that radiation data characterization is still in process and the chip isn't fully radiation-qualified yet1
.
Source: Tom's Hardware
Related Stories
Intel Government Technologies is managing the Starfire program, with samples expected in Q3 2026 and a commitment to market-competitive pricing and domestic manufacturing
1
. Intel Foundry holds Trusted Foundry status and is the only US-based manufacturer of leading-edge logic, having tied its 18A process and packaging roadmap to Pentagon programs including RAMP-C and SHIP1
. However, 18A yields aren't expected to reach industry-standard levels until 2027, which could affect production timelines.The Starfire chips are essentially based on the Panther Lake 4Xe3 SKU, manufactured in the United States specifically for the US government
3
. By integrating advanced process technology into satellites and spacecraft that require real-time data processing under strict size, weight, and power constraints, Intel is positioning itself to capture a growing segment of space infrastructure that demands more computational capability for autonomous operations and data analysis in orbit2
.Summarized by
Navi
[1]