A 28nm FD-SOI FPGA has run continuously for more than a year in a 10 Gbps inter-satellite laser link. The significance here is not Lattice itself, but that rad-hard programmable devices are moving from peripheral telemetry and control roles to core control positions in high-throughput links such as optical communication terminals.
Lattice announced on November 16, 2026 that its Nexus-series rad-hard FPGA has been operating in orbit for more than a year in a TESAT optical communication terminal, in a 10 Gbps space-based laser communication application. Andreas Hanselmann, who leads optoelectronics and development at TESAT, cited rad-hard performance, power consumption and long-term reliability as the three selection criteria, and said the device's on-orbit behavior met all three. Lattice senior vice president for aerospace and defense Jim Tavacoli framed his comments around core control functions for optical links and onboard processing, emphasizing low power, high reliability and autonomy. It is important to distinguish that these are statements from the vendor and the customer; the on-orbit duration and link rate were published by both parties, while specific bit error rates and single-event metrics were not disclosed. The genuinely verifiable hard information is the process: the Nexus platform is based on 28 nm fully depleted silicon-on-insulator, and that node choice itself points to a tradeoff between low power and radiation tolerance.
TESAT has carried optical communication terminals on GEO and LEO missions for more than a decade, serving broadband constellations and institutional programs, and its Backnang, Germany site has more than 1,200 employees developing, assembling and testing satellite communication systems and equipment. For this terminal electronics redesign, Lattice's solution had to satisfy three constraints at once: radiation tolerance, low power and board footprint. The official line is that the Nexus platform suits terminals requiring years of maintenance-free operation. Lattice also lists its space heritage as several thousand FPGAs launched to date, and says space-grade versions of Certus-NX, CertusPro-NX and CrossLink-NX are all available, along with the Radiant design software and radiation characterization data. The informational value here lies in product-line coverage: from control to interface, Lattice has built out a full family of space-grade programmable devices rather than a single-point part.
Put this back on the desk of a domestic engineer, and the direct conclusion is: this is a reference selection path, not a replicable supply-chain path. A 28 nm FD-SOI rad-hard FPGA running for a full year in a 10 Gbps laser link shows that low-power programmable devices can handle onboard high-speed link control, which has architecture-level reference value for teams working on optical terminal electronics and onboard data-processing boards. The problem is device availability: Lattice's space-grade line depends on its own process and radiation characterization data, and domestic projects have no guarantees on export controls or lead times, while the Radiant toolchain also creates lock-in. In the short term, such devices are better suited as architecture references than as production selections; in the medium term, the gap in domestic rad-hard FPGAs at 28 nm and below, and in supporting radiation characterization databases, is the key variable determining whether substitution is possible.
For domestic engineers working on high-speed interfaces and boards, the value of this news is not direct part substitution but validation of an architectural judgment: the control plane of an onboard 10 Gbps-class laser link can be handled by low-power programmable devices without resorting to high-power processors. Teams working on optical terminal electronics and onboard processing boards can use this to recheck their power and footprint budgets. But on the supply chain, Lattice's space-grade devices are not stable for domestic projects, and the Radiant toolchain and radiation characterization data create de facto lock-in, so production substitution is not feasible in the short term. What really bears watching is progress in domestic rad-hard FPGAs on 28 nm FD-SOI or comparable processes, and whether radiation characterization databases are publicly available, which determines whether alternatives can enter the selection list.
Daily Briefing · Weekly Digest · Monthly Deep Dive