BUDAPEST, Hungary, May 14, 2026 /PRNewswire/ -- aiMotive, a leading provider of automotive AI technologies, today announces the official launch of aiWare5, its latest-generation automotive NPU IP. Following a successful soft launch at CES and a significant extension of the aiWare ecosystem through a strategic relationship with Socionext to meet the demands of global Tier 1 and OEM partners, aiWare5 is now available for general licensing. While aiWare has always been designed to accelerate any automotive AI workload, this latest release also fully supports the latest aiDrive automated driving software stack, which was recently announced to be available for general licensing.

Unlike many more general-purpose NPUs on the market, aiWare was designed from the ground up specifically for automotive inference. aiWare5 delivers many significant enhancements over the production-proven aiWare4 by maintaining industry-leading efficiency while significantly scaling up the massive throughput required for next-generation systems. While aiWare4 has already demonstrated its capability by delivering full L2+ driving solutions on cost-effective platforms like the Nextchip APACHE6, aiWare5 introduces many new features and upgrades, including:

• Enhanced support for future leaps in AI workloads such as LLMs, Vision Transformers, and SSMs
• Support for dynamic FP8 scaling
• Further extending leadership in tooling with a unique GPU-optimized emulator enabling faster-than-real-time bit-accurate modeling prior to first silicon

This unique combination of performance scalability and ease of use in system modeling enables faster design cycles from edge sensor processors to centralized high-performance vehicle compute platforms.

As the world's first ISO 26262 ASIL B-certified NPU IP, aiWare provides a safety-first foundation for production automotive programs. Developed as a Safety Element out of Context (SEooC), it is backed by comprehensive safety documentation and compliance support, simplifying integration into ASIL B and higher-compliance systems.

Beyond hardware, aiWare5 is complemented by industry-leading development tools. The unique aiWare GPU-optimized emulator enables large-scale software-in-the-loop (SiL) validation of the final AI-based ADAS software stack without target silicon. Combined with aiWare Studio – which enables offline neural network optimization and performance estimation within 5% of final silicon – development teams can complete the majority of workload optimization before hardware availability, significantly reducing development timelines and integration risk.

The architecture's deterministic design further supports certification efforts and ensures predictable, production-ready performance across safety-critical automotive deployments.

"aiWare5 represents a major step forward in scalable, production-ready automotive AI acceleration," said Marton Feher, SVP of Semiconductor Engineering. "From day one, we have dedicated ourselves to automotive-native inference. While others adapt general-purpose designs, our focus on the specific rigors of the vehicle environment allows us to provide OEMs and semiconductor partners with an unrivaled, safety-certified, and future-proof NPU IP solution."

aiWare5 is available now for immediate licensing and IP delivery, suitable for integration into next-generation automotive SoCs, dedicated AI accelerator coprocessors, or chiplets.

For more information, visit: www.aimotive.com

Contact:

Bence Boda
Marketing Director
bence.boda@aimotive.com

Logo - https://prnewswire.vietreader.com/uploads/posts/2026-05/aimotive-announces-aiware5-delivering-unrivalled-flexibility-and-scalability-for-l2-to-l4-automotive-ai-workloads-1.jpg

Source: aiMotive