32.8 TOPS INT8 · 512 GFLOPS FP16 · Native FFT / DCT / STFT / Wavelet
GF 22FDX · 2.5W · OEM Silicon IP
No signal transform tax. No off-chip memory bottleneck. No ecosystem lock-in.
Proprietary MAC architecture with unstructured sparsity engine. ~40% effective throughput uplift on sparse models.
FFT · DCT · STFT · Wavelet — all first-class silicon. No CPU offload. No latency penalty.
Half-precision inference engine. Handles mixed INT8/FP16 pipelines natively — ideal for transformer and vision models.
32 MB SRAM (MAC output) + 32 MB eMRAM (model weights). Entire MobileNet V2 fits on-chip.
Skip zero MACs at hardware level — no software pruning required. Works with any model out of the box.
Standard AXI4 + AXI4-Lite + APB interfaces. Drops into any SoC without custom glue logic.
End-to-end compiler toolchain that maps standard neural network graphs directly to AdaptCore's proprietary compute engine — no manual kernel writing required.
| Chip | INT8 TOPS | FP16 | Sparsity | Signal Transforms | Power |
|---|---|---|---|---|---|
| AdaptCoreOUR IP | 32.8 | 512 GFLOPS | Unstructured | FFT · DCT · STFT · Wavelet | 2.5W |
| Google Edge TPU | 4 | ✗ | ✗ | ✗ | 2W |
| Hailo-8 | 26 | ✗ | ✗ | ✗ | 2.5W |
| Apple ANE (A17) | 35 | FP16 (locked) | ✗ | ✗ | ~5W |
| NVIDIA Jetson Orin NX | 50 | 1.6 TFLOPS | Struct 2:4 | ✗ | 10–25W |
| Qualcomm Hexagon 8G3 | ~45 | ✗ | ✗ | ✗ | 5–8W |
✗ = not supported at this power point · Figures at typical inference workload
Native signal processing unlocks applications no competitor can match at this power.
Industrial inspection, surveillance, robotics. INT8 MAC array delivers 32.8 TOPS for real-time multi-stream object detection and segmentation — all under 3W.
5G RAN and radar signal processing. Native 64-unit FFT engine computes full OFDM demodulation in silicon — no separate DSP co-processor required.
MRI reconstruction, CT filtering, ultrasound AI. Wavelet and DCT engines handle imaging transforms natively — eliminating the reconstruction CPU entirely in portable devices.
Always-on ASR, noise cancellation, voice activity detection. STFT-native acceleration processes audio features in silicon — enabling continuous listening at under 500mW.
IEEE P1735 encrypted delivery. Zero RTL exposure. Three tiers to match your programme.
Share your use case — datasheet + integration guide within 48 hours.