PolarFire® FPGA Ethernet Sensor Bridge Rev 2.0 Expands Edge AI Connectivity
The new PolarFire FPGA Ethernet Sensor Bridge Rev 2.0 provides a smaller, more scalable platform for connecting high-speed image sensors to NVIDIA® edge AI systems. The updated platform is 60% smaller than the first generation, supports up to four MIPI CSI-2® cameras, adds NVIDIA Jetson™ camera connector compatibility and incorporates on-board optical latency measurement circuitry.
TechBytes Newsletter: July 2026 Edition
The July 2026 edition of FPGA and SoC TechBytes is now available, with new technical insights and application examples, including: VectorBlox™ Accelerator SDK 3.0 model optimization and compression; a DDR-free, Zephyr®-first PolarFire® SoC architecture; new SDI, HDMI®, DisplayPort™ and video-processing IP capabilities; hardware-based clock-glitch detection and configurable tamper response; and new medical imaging and space-system examples.
VectorBlox™ 3.0 Makes Edge AI More Efficient on PolarFire® FPGAs and SoCs
The new VectorBlox™ Accelerator SDK 3.0 makes it easier to deploy efficient AI inference on PolarFire® FPGA and SoC FPGA platforms. VectorBlox 3.0 adds support for sparse neural networks and advanced model compression, allowing the accelerator to skip unnecessary zero-valued operations and reduce compute, memory and bandwidth requirements.
The integrated tool flow supports familiar AI frameworks and provides model optimization, compilation, simulation and deployment capabilities, helping developers increase inference performance while maintaining the low power, security and reliability advantages of PolarFire devices. VectorBlox 3.0 can deliver up to 2× more AI performance using the same FPGA resources for supported sparse workloads.
Upcoming FPGA Webinars
Mission-Ready FPGAs: Enabling the Next Generation of Defense Systems
This webinar explores how FPGA technology can support increasingly software-defined, connected and autonomous defense platforms. The live webinar is scheduled for September 15, 2026, at 9:00 a.m. PDT.
FPGA-Accelerated Motor Control With PolarFire® SoC for Intelligent Motion Systems
FPGA-based parallel processing enables deterministic, low-latency multi-axis motor control for robotics and industrial automation. PolarFire SoC motor-control solutions support parallel Field-Oriented Control (FOC), control loops as fast as 1 µs and Time-Sensitive Networking (TSN)-synchronized motion control over industrial Ethernet.
Coming Soon—PolarFire® 2 FPGAs: Advancing Compute and Power Efficiency
Building on the proven power and thermal efficiency of today’s PolarFire® FPGA and PolarFire SoC FPGA families, PolarFire 2 FPGAs will deliver significantly greater compute capability while maintaining a strong focus on power efficiency. A new development toolset will also simplify FPGA acceleration, making it easier for algorithm and application developers to take advantage of FPGA performance without requiring deep FPGA design expertise.
Explore FPGA and SoC Solutions
PolarFire FPGAs
- Cost-optimized, lowest power in mid-range FPGA class
- 250 Mbps to 12.7 Gbps transceivers
- 100K to 500K programmable Logic Elements (LEs) and up to 33 MB RAM
- Military-grade security and exceptional reliability
- Available as PolarFire Core FPGAs with the same features and pin compatibility without transceivers or PCIe®
PolarFire SoC FPGAs
- Embedded, coherent RISC-V CPU cluster and a deterministic L2 memory subsystem
- 25K to 460K programmable LEs with 12.7 Gbps transceivers
- Up to 50% lower power than alternatives
- Integrated DDR3/4, LPDDR3/4 controller and PHY
- Available as PolarFire core SoC FPGAs with the same features and pin compatibility without transceivers or PCIe
Radiation-Tolerant PolarFire FPGA and SoC FPGA
- Based on flight-proven, radiation-tolerant, power-efficient PolarFire FPGA fabric
- SoC FPGA with embedded, coherent RISC-V CPU cluster and a deterministic L2 memory subsystem
- No configuration upsets
- 100 krad Total Ionizing Dose (TID)
- Package options qualified for space flight applications