Microchip, another acquisition

Microchip Technology recently acquired Neuronix AI Labs to extend the capabilities of its energy-efficient, AI-enabled edge systems deployed on field-programmable gate arrays (FPGAs), according to reports. Neuronix AI Labs provides neural network sparse optimisation technology that reduces power, size and computation for tasks such as image classification, object detection and semantic segmentation, while maintaining high accuracy, according to the profile.

　　Microchip's mid-range PolarFire? FPGAs and SoCs already lead the industry in low power consumption, reliability and safety features. The acquisition of this neural network sparse optimisation technology will enable Microchip to develop cost-effective, large-scale edge-deployed components designed for computer vision applications on systems with cost, size and power constraints, and to enable AI to exponentially grow /ML processing power on low- and mid-range FPGAs.

　　"This acquisition will increase the power efficiency of our FPGAs and SoCs deployed in intelligent edge systems utilising AI/ML algorithms," said Bruce Weyer, corporate vice president of Microchip's FPGA Business Unit. "Neuronix technology combined with our VectorBlox design flow improves neural network performance efficiency and delivers outstanding GOPS/watt performance in our low-power PolarFire FPGAs and SoCs. System designers will now be able to build and deploy small-scale hardware that was previously difficult to build due to size, thermal or power constraints."

　　Neuronix sparse optimisation technology will allow non-FPGA designers to leverage powerful parallel processing capabilities using industry-standard AI frameworks without requiring in-depth knowledge of the FPGA design flow.Neuronix AI intellectual property, combined with Microchip's existing compiler and software design suite, enables the implementation of AI/ML algorithms on customisable FPGA logic without requiring resistor Translation Level (RTL) knowledge or an in-depth understanding of the underlying FPGA architecture. It also allows CNNs to be dynamically updated and upgraded without reprogramming the hardware.

　　"Neuronix AI Labs has been focused on producing best-in-class neural network acceleration architectures and algorithms that transform user expectations for size, power, performance and cost," said Yaron Raz, executive officer, Neuronix AI Labs. "Joining the Microchip team provides us with a unique opportunity to expand and align an FPGA product portfolio that sets the industry standard for energy efficiency."

　　Previously reported, Microchip Acquires SerDes Pioneer VSI

　　Previously, Microchip Technology acquired Korea-based VSI to leverage its Automotive SerDes Alliance (ASA) open standards technology for in-vehicle networking (IVN).

　　The deal adds VSI ASA Motion Link (ASA-ML) SerDes technology to Microchip's Ethernet and PCIe automotive networking portfolio to support next-generation software-defined vehicles as BMW moves to ASA technology. Terms of the deal have not been disclosed.

　　Last month, BMW announced that they would be moving to standardised ASA-ML to begin production of their vehicle architectures and now also video architectures. This provides link layer security and scalability to support 2 Gbps to 16 Gbps line rates and is being updated to support Ethernet-based architectures.

　　The ASA has more than 145 members, including 11 automakers such as BMW, GM, Ford, Stellantis and Hyundai-Kia Motors, as well as suppliers, semiconductor and image sensor vendors, and test and compliance organisations.

　　The market for automotive radar, camera and LiDAR modules is expected to more than triple in size between 2022 and 2028, reaching $27 billion in revenue, according to the Yole Group. This is driven by ADAS, in-vehicle monitoring, safety and convenience features (e.g., 360-degree surround view, electronic mirrors), and multi-screen digital cockpits for next-generation software-defined vehicles (SDVs).

　　These require more highly asymmetric raw data and video links and higher bandwidth, making current proprietary serialiser/deserialiser (SerDes)-based solutions no longer sufficient.

　　The Automotive SerDes Alliance (ASA) was formed in 2019 and published an open standard ASA Motion Link (ASA-ML) specification to address this issue, and VSI has been a pioneer in developing connectivity devices that comply with the standard.

　　Mitch Obolsky, vice president of automotive products, networking, Microchip, said, "This acquisition brings VSI's knowledgeable team, its market appeal, and ASA Motion Link technology and products to Microchip's broad automotive networking portfolio to better serve the ADAS megatrend that we are focusing on. ADAS megatrends." and Data Centre business units.

　　"Camera and display connectivity is one of the fastest growing and sizable IVN markets as the industry converges around the three main IVN pillars: Ethernet, PCIe and ASA Motion Link. With VSI, Microchip can now offer products across all three pillars and provide our customers with automotive safety, microcontroller, motor control, touch and power management solutions to support their next-generation software-defined vehicle architectures."

　　Steve Kang, executive officer of VSI, said, "Microchip Technology is a well-established and trusted marketplace for automotive networking, known for its automotive quality and strong supply chain, and our team is excited to join them in addressing the growing market for ADAS and digital cockpit connectivity. " .

　　"VSI is in the ASA-ML product development space and a company that brings products to market. Our standards-compliant chipsets are being evaluated by automakers around the world. We recently conducted a proof-of-concept with BMW to demonstrate ASA-ML and our product readiness. This acquisition brings together two organisations with a shared commitment to advancing technology through innovation. We look forward to successfully deploying our solutions in production vehicles in the coming years."