AI Processor IP + software stack built to keep pace with your model roadmap.
One IP foundation
Inference. 1 – 6400 TOPS.
One architecture
Low latency. Power efficient.
One toolchain
Any chip. Your models.
The Problem
Multiple accelerator IPs. Multiple toolchains. And NPUs that can't keep up with the models you need to run.
There's a better way to build.
The Solution
Licensable Processor IP for end-to-end inference. One toolchain across your chip line.
The Opportunity
The next billion AI users won't connect to a datacenter. They'll run inference locally—on silicon built for it.
Multimodal AI on PCs and mobile. Voice, gesture, and personalization. Privacy and latency—solved.
Edge servers, smart printers, inspection, security, and analytics—at the edge where data lives.
Multimodal perception and decision-making. Sensor fusion, navigation, and manipulation.
Vision, radar, and LiDAR processing with ASIL B/D support. From in-cabin monitoring to full ADAS.
Here's how it's built.
For Architects
Single-batch performance. Lower power. Deterministic timing. Simpler code.
Optimized for single-batch latency. Immediate inference for real-time responsiveness.
No round trips to separate vector units or DSPs. Data stays local, memory bandwidth drops.
Data movement is instruction-encoded—no routing decisions, no contention. Every cycle is predictable.
Single instruction stream. No partitioning code between NPU, DSP, and CPU. One codebase.
For Developers
Compile most models. Extend the rest in C++.
Graph Compiler auto-compiles hundreds of models. Import ONNX from PyTorch or TensorFlow and run.
No waiting on Quadric for new ops. Python via ChiPy™ or C++ extensible with LLVM compiler included.
Instruction Set Simulator (ISS) gives cycle-accurate performance prediction before tape-out.
One processor handles pre-processing, inference, and post-processing. No partitioning across NPU, DSP, CPU.
Stay Updated
May 29, 2026
Vision-language-action models dominated discussion at this year's Embedded Vision Summit. Every team building autonomous systems is targeting them. But the hardware question nobody is answering directly: can your NPU actually run one? We benchmarked Pi-0.5 to find out.
May 19, 2026
Whisper runs end-to-end on Chimera. Encoder and decoder compiled as native GPNPU kernels, INT4 weights, FP16 attention, top-1 token match against the float32 reference. Scales to four cores with a flag, no recompile.
Apr 8, 2026
AI agents hallucinate when they hit domain-specific silicon knowledge like the Chimera GPNPU. We built an MCP server to address that. Here's what we've learned about where MCP works well and where it gets complicated.
Jan 14, 2026
Tripling product revenues, comprehensive developer tools, and scalable inference IP for vision and LLM workloads, position Quadric as the platform for on-device AI.
Jan 14, 2026
Quadric today announced that TIER IV, Inc., of Japan has signed a license to use the Chimera AI processor SDK to evaluate and optimize future iterations of Autoware, open-source software for autonomous driving pioneered by TIER IV.
Dec 11, 2025
Former EVP/GM of Synopsys IP Division Appointed Independent Board Member BURLINGAME, Calif., Dec. 11, 2025 /PRNewswire/ -- Quadric® today announced the appointment of Joachim Kunkel as an independent member of the Board of Directors of the company. Kunkel joins a revamped Quadric Board in anticipation of an imminent closing of Quadric's Series C fund raising.
Get the latest on AI silicon, product updates, and industry insights delivered to your inbox.
Join Our Mailing ListIn datacenters, yes. But also in pockets, dashboards, factories, and devices you haven't imagined yet. We built the silicon that gets it there.
