Specifications
Product
Huashan A2000 family — high-performance, all-scenario automotive AI chip platform
NPU
Jiushao NPU with large-core architecture for real-time AI large-model inference
Precision Support
FP16, FP8 floating point plus INT4 / INT8 / INT16 integer compute
Automation Levels
L2 through L5 driving automation, perception and decision-making
CPU Architecture
Arm-based AI computing platform (Arm success-library partner)
Safety Positioning
Claims highest industry safety levels with training-to-deployment inference consistency
Toolchain
BaRT — end-to-end development from model training to deployment
Workload Focus
Real-time inference for large models with reduced latency on complex computational tasks
Regulatory Status
Passed US Department of Commerce and Department of Defense review (January 2026)
Global Availability
Cleared for international sales and deployment outside China
Company
Black Sesame Technologies, founded 2014, listed on HKEX (2533.HK)
Delivery Model
SoC plus self-developed IP cores, algorithms and support software as a full-stack solution
Partners
FUTU-RE (Japan market implementation support agreement, May 2026)
Volume Status
Extensive mass-production track record across prior Huashan generations
Target Customers
Automotive OEMs and Tier-1 suppliers
Overview
The Black Sesame Technologies Huashan A2000 is a high-performance automotive AI chip platform designed for next-generation AI models across L2 to L5 driving automation. Its Jiushao NPU uses a large-core architecture specifically built for real-time inference of large models, cutting latency on the heavy perception and decision-making workloads that define modern ADAS and autonomous stacks.
Numerically the platform supports FP16 and FP8 floating point alongside INT4, INT8 and INT16 integer formats, which allows a single SoC to serve both accuracy-critical perception networks and aggressively quantised transformer workloads. Black Sesame pairs the silicon with BaRT, its own AI toolchain, giving a continuous path from model training through to on-vehicle deployment rather than a separate embedded conversion flow.
In January 2026 the Huashan A2000 passed review by the US Department of Commerce and Department of Defense, clearing it for global sale and deployment — a significant milestone for a Chinese automotive AI chipmaker and the reason this platform is now reaching European, Japanese and North American design pipelines. Black Sesame is listed on the Hong Kong Stock Exchange under 2333.HK lineage and has signed Japanese market implementation support agreements to localise integration.
QS Compute supplies Huashan A2000 platform silicon, evaluation hardware and integration support for automotive Tier-1 and robotics programmes. Contact us to discuss quantities, thermal and functional-safety requirements and toolchain onboarding.
Key Benefits
A mass-production automotive AI SoC with a documented US clearance for global deployment; large-core NPU tuned for real-time large-model inference rather than legacy CNN workloads; wide precision range from FP16 down to INT4 in one device; a single-vendor full stack covering IP cores, algorithms and the BaRT toolchain; existing OEM and Tier-1 mass-production history.
Applications
ADAS and automated driving domain controllers; L2-L5 perception and decision stacks; robotics and autonomous machine vision; intelligent cockpit AI; industrial autonomous vehicle programmes.
Request a Quote — BLACK SESAME HUASHAN A2000 — JIUSHAO-NPU AUTOMOTIVE AI SOC PLATFORM
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →