New LaunchJul 14, 2026, 07:12 AM
Microchip Launches VectorBlox 3.0 SDK for AI on FPGAs
AI Summary
Microchip Technology has launched the VectorBlox™ 3.0 Accelerator Software Development Kit (SDK) and associated CoreVectorBlox IP, designed to simplify FPGA-based AI implementation and accelerate time-to-market. This new SDK leverages sparsity-based model compression, acquired from Neuronix, to reduce compute demands and power consumption while maintaining accuracy, crucial for power-constrained and mission-critical environments like aerospace and defense. The integrated toolchain streamlines optimization, compilation, and deployment of convolutional neural network (CNN) models on PolarFire® FPGA and SoC-based platforms, and is offered to developers free of charge.
Key Highlights
- Microchip released VectorBlox™ 3.0 Accelerator Software Development Kit (SDK).
- SDK simplifies FPGA-based AI implementation and speeds time-to-market.
- Features sparsity-based model compression from Neuronix acquisition.
- Reduces compute demands and power consumption for AI inference.
- Enables efficient execution of vision-based CNN models on PolarFire FPGAs.
- VectorBlox 3.0 SDK and CoreVectorBlox IP are available free of charge.
- Validated by Planetek Italia for onboard AI in space (AI-eXpress-1 satellite).
- Validated by AIKO for autonomous payload operations with clear_CHARLES suite.
Price Impact
More from MCHP