Imagine a brain for a machine that is not only incredibly fast and energy-efficient but also built like a digital fortress. Instead of one giant chip, it uses a Lego-like system of smaller specialized pieces that can be rearranged to fit different tasks. This allows devices to think and react instantly without needing to send data to a distant cloud server.
Current edge-AI hardware often forces a trade-off between high performance and low energy consumption, while lacking the rigorous safety certifications required for automotive and aerospace use.
A complete European HW/SW stack featuring a RISC-V multicore chip with two tightly coupled chiplets and a HW/SW Design Space Exploration tool.
If you are an autonomous vehicle manufacturer dealing with strict safety standards like ISO 26262 — this project developed a secure AI hardware and software stack that ensures real-time processing with high safety integrity levels.
If you are an avionics systems integrator dealing with DO-178C certification requirements — this project developed a reconfigurable processing platform that provides high security and safety for flight-critical AI applications.
If you are a medical device developer dealing with the need for low-power, secure patient monitoring — this project developed an Edge-AI chiplet system that maximizes inferences per watt while keeping sensitive data encrypted.
Based on available project data, specific pricing is not mentioned, but the project aims to democratize access to these resources as a service for SMEs to large organizations.
The project utilizes multi-chiplet integration and silicon packaging, which are standard industrial methods for scaling high-performance semiconductor production.
Based on available project data, the project develops a purely European HW/SW stack, but specific licensing terms for the RISC-V based designs are not detailed.
The platform includes a dedicated software stack, middleware, and AI libraries designed to take full advantage of the underlying hardware accelerators.
The project period runs from 2023-02-01 to 2027-04-30, suggesting the technology will reach maturity toward the end of this window.
The consortium is heavily industry-weighted with 13 industrial partners (52% ratio), including 8 SMEs. This strong commercial presence, combined with 6 universities and 5 research centers across 9 countries, indicates a high probability of commercial translation and a focus on practical market needs rather than pure theory.
Contact POLYTECHNEIO KRITIS in Greece for partnership inquiries.
Contact us to identify the specific chiplet IP available for licensing from the REBECCA consortium.
