The Ximplic inference IP
One energy-efficient inference block. It does the math where the data already sits, so always-on devices run AI without draining the battery. Proven in software and on FPGA today, ready to license into your SoC.
A licensable inference block that computes in memory
One block, a de-risked path
Our proprietary flow takes you from a trained model to silicon-ready IP: evaluate it in software, map any model onto it, then license the same validated design.
Evaluate
VyzoraRun any AI model on a virtual copy of the chip and measure power up front. Software only, no hardware.
Learn moreMap
VextylThe compiler maps any trained ONNX model onto the memory array, quantized and checked against the hardware.
Learn moreSilicon
SilvenA hardened, drop-in macro. On the roadmap.
On the roadmapTwo ways it integrates into the SoC
From a low-risk drop-in to an in-place memory upgrade. Option A is available and FPGA-tested today. Option B is on the way.
External Accelerator
- — Sits next to the host SoC as separate IP
- — Standard AXI bus connection
- — Zero changes to host silicon
In-Place Memory Upgrade
- — Replaces passive on-chip SRAM
- — Same footprint, memory now computes
- — Zero data movement
Start the evaluation, then the design-in
Share your target workloads and platform. We'll scope a software evaluation under NDA, then walk the same path you just read into your SoC: map your model, integrate over AXI.
Email info@ximplic.com · Groningen, The Netherlands