Along with the partnership, Renesas also invested into EdgeCortix’s most recent funding round. Through both the investment and the partnership, this relationship will provide Renesas with unique access to EdgeCortix’s technology.
EdgeCortix is a start-up founded to deliver high-speed, low-power, AI-focused processors with a software-first approach for edge applications. The company’s suite of proprietary software and AI chip products accelerate AI inference tasks at very low power with low latency.
To date, EdgeCortix has brought three products to market including: MERA, its multiple hardware platform compatible compiler software framework; Dynamic Neural Accelerator (DNA), its scalable, run-time reconfigurable neural network processor intellectual property (IP); and the power-efficient SAKURA AI co-processor device.
Collectively, these form an integrated solution, especially suited for real-time edge AI. EdgeCortix claims that its solutions can deliver up to 50x better energy efficiency and over 10X better performance versus conventional GPUs and CPUs.
Renesas has sort to establish a leadership position in edge and endpoint AI offering a broad portfolio of embedded computing solutions complemented by its Reality AI Software and a larger number of internal and partner solutions for AI in industrial, smart home, automotive, and other applications.
Partnering with EdgeCortix will support Renesas in streamlining and unifying its overall AI/ML (Machine Learning) developer experience across its entire MCU and MPU portfolio and enable a seamless back-end layer for compilation that supports heterogeneous architectures.
“Renesas saw the potential for AI/ML at the edge and endpoint very early on, and we have moved swiftly to solidify our position in this fast-growing segment,” said Sailesh Chittipeddi, Executive Vice President and General Manager of Renesas’ Embedded Processing, Digital Power and Signal Chain Solutions Group. “The combination of our embedded processors and Reality AI Software along with EdgeCortix solutions will enable our customers to seamlessly compile their AI code with support for heterogeneous architectures, reducing their development risk, time and cost.”
Working with EdgeCortix provides Renesas the ability to leverage its existing IP and hardware, while simultaneously tapping into EdgeCortix’s expertise in key technology areas, including Apache TVM1, compiler development as well as heterogeneous framework design.
This partnership also enables Renesas to access specialised hardware expertise that EdgeCortix has developed leveraging Arm as well as multiple other architectures including Renesas’ DRP-AI accelerator.
