Platform roadmap

Our goal is to provide the industry with an NPU reference implementation, based on RISC-V standards, that is easy to adopt and integrate.

The initial Coral NPU release included our RISC-V compliant scalar core and the vector execution unit. A future release will introduce the RISC-V matrix execution unit that will provide a complete, end-to-end, open-source stack for a fully RISC-V compliant NPU. Below is a high-level roadmap for planned feature additions and enhancements to the platform.

Note that this roadmap is a tentative plan and is subject to change.

	Milestone 2 Released: Q4 2025	Milestone 3 Coming in 2026
Focus	RISC-V vector (RVV) execution engine launch	Premier reference accelerator for Gemma
Highlights	For low-power wearable devices, adding always-on ambient sensing. Neural encoding of sensor data, acts as a trigger for high-power domain or tethered device. Scalar core achieves open standards ISA compliance. Introduces new RISC-V vector execution unit, integrated with scalar core for RVV specification-compliant SIMD. MLIR compiler support, from LiteRT (formerly TensorFlow Lite) frontend.	For low-power wearable devices, adding always-on ambient sensing. Neural encoding of sensor data, acts as a trigger for high-power domain or tethered device. Scalar core achieves open standards ISA compliance. Improves the RVV vector execution unit to support larger models and transformer architectures. MLIR compiler support, targeting LiteRT and JAX frontends.
Specifications	RISC-V 32I ISA 4-stage CPU RISC-V 32V ISA vector unit with VLEN = 128 Supported numerics: int8, int16 Area = approx. 0.2 mm^2 in TSMC 12nm (N12) Performance = 800 MHz clock x 16 MACs per cycle	RISC-V 32I ISA 4-stage CPU RISC-V 32V ISA vector unit with VLEN = 128, 256, or 512 Supports larger memory capacity, up to 4 GB maximum Supported numerics: int8, int16, FP32 (FP16, BF16) Area = approx. 0.2 mm^2 in TSMC 12nm (N12) Performance = 800 MHz clock x 16 MACs per cycle
Reference architectures	CNNs: Lyra (audio) MobileNet (image)	Gemma3 270M (LLM) CNNs: Lyra (audio) MobileNet (image)
Target model size	Less than 1 MB	Less than 1 GB
Example use cases	Is someone talking Keyword detection Person or object detection	Is someone talking Keyword detection Person or object detection Who is the person What is the object LLM transcription/translation Token encoding for tethered device