Processor/DSP/FPGA-2025 World Electronics Achievement Awards-EE Times

RIVAI Technologies (Shenzhen) Co., Ltd. （简称RIVAI Technologies）

Product：LingYu

Recommended reasons： I. Key Features Fully Self-Developed RISC-V Architecture: China’s first entirely self-developed high-performance RISC-V server processor. Heterogeneous "Dual-Core in One Chip" Design: Combines a 32-core CPU with an 8-core LPU to balance traditional computing and AI workloads. Advanced Interconnect Technology: Supports CXL 2.0 and PCIe 5.0 protocols, enabling up to 8-way interconnect. Enterprise-Grade Reliability: Incorporates RAS features (e.g., dynamic tuning and dedicated management cores) to meet data center stability demands. II. Key Specifications Architecture: RISC-V ISA with a 32-core CPU + 8-core LPU heterogeneous design. Process Technology: Advanced node with Design-Technology Co-Optimization (DTCO) for enhanced power efficiency. Memory Support: DDR5 high-speed memory interface. I/O Interfaces: PCIe 5.0 and CXL 2.0, supporting 8-way interconnect.. Scalability: Compatible with 8x GPU direct-attach, 6x 400Gbps high-performance NICs, and 20+ NVMe all-flash storage configurations. III. Target Applications Designed for next-gen workloads: machine learning, recommendation systems, vector databases, and 671B open-source LLMs (e.g., DeepSeek). IV. Competitive Advantages Independenc: Fully self-developed RISC-V architecture. Superior Power Efficiency: Hardware-software co-design and process optimization reduce TCO. Interconnect Leadership: Pioneering CXL 2.0 and PCIe 5.0 support (8-way) outperforms mainstream PCIe 4.0 server chips. Workload-Optimized: Integrated LPU accelerates AI inference natively, eliminating the need for external GPUs/FPGAs.

More Vote

GOWIN Semiconductor Corp.

Product：GW5AT-LV15CS130

Recommended reasons：

More Vote

AMD

Product：AMD Versal Premium Series Gen 2

Recommended reasons： The AMD Versal™ Premium Series Gen 2, an adaptive SoC platform designed to deliver the highest levels of system acceleration for a wide range of workloads. Versal Premium Series Gen 2 will be the FPGA industry’s first devices featuring Compute Express Link (CXL^®) 3.1 and PCIe^® Gen6 as well as LPDDR5X memory support in hard IP. These next-generation interface and memory technologies access and move data rapidly and efficiently between processors and accelerators. CXL 3.1 and LPDDR5X help unlock more memory resources faster to address the growing real-time processing and storage demands of data-intensive applications in data center, communications, test and measurement markets. With support for CXL 3.1 and PCIe Gen6, the industry’s fastest host interfaces, Versal Premium Gen 2 devices enable industry-leading, high-bandwidth host CPU-to-accelerator connectivity. PCIe Gen6 offers a 2-4X faster line rate compared to competing FPGAs with PCIe Gen4 or Gen5 support, while CXL 3.1 running PCIe Gen6 provides double the bandwidth of competing devices with CXL 2.0 at similar latencies, as well as enhanced fabric and coherency capabilities. AMD Versal Premium Series Gen 2 adaptive SoCs accelerate memory bandwidth for faster data transfers and real-time responsiveness with the fastest LPDDR5X memory connectivity available, at up to 8533 Mb/s. This ultra-fast, enhanced DDR memory enables up to 2.7X faster host connectivity over comparable competitive devices with LPDDR4/5 memory. Connectivity to CXL memory expansion modules enable up to 2.7X more total bandwidth than LPDDR5X memory alone.

More Vote

NXP Semiconductors

Product：i.MX 93 Applications Processor

Recommended reasons： i.MX 93 applications processor is NXP’s first i.MX MPU to integrate the scalable Arm Cortex-A55 core. Optimizing performance and power efficiency for Industrial, IoT and automotive devices, i.MX 93 processors are built with NXP’s innovative Energy Flex architecture, it is designed with multi-core heterogeneous architecture, featuring Cortex-A55 dual core for high performance efficiency, and at the same time integrates Cortex-M33 core for real time control. With the creative core combination, the MPU can provide flexible power and performance tuning with dedicated clocks and supply voltages. Aside from the main cores, i.MX 93 also delivers efficient machine learning (ML) acceleration by integrating Arm Ethos™-U65 microNPU, providing significant performance improvement in broad AI application scenarios such as image recognition, classification, anomaly detection. Powered with NXP’s eIQ Toolkit, i.MX 93 enables developers to create more capable, cost-effective and energy-efficient ML applications. i.MX 93 also provides advanced security with integrated EdgeLock^® secure enclave. It is NXP’s pre-configured, self-managed automomous security subsystem enabling platform level security and trust. It supports functions such as key management, cryptographic services, trust provisioning, remote attestation, secure boot, RNG, tamper detection, secure clock, and also support memory with inline ECC. A rich set of peripherals are also provided within i.MX 93 including Gb Ethernet with TSN, which is essential for industrial networking, MIPI-DSI and MIPI-CSI to support display and vision applications, and rich set of interfaces such as UART, SPI, I2C, I3C, CAN-FD, to support various applications. With all these innovative features, the i.MX 93 applications processor is suitable for next generation industrial applications such as industrial control, gateway, PLC in factory automation, as well as consumer electronics, home entertainment and smart appliances in IoT applications. The MPU is also supported by NXP’s product longevity program and ensures at least 15 years manufacturing and production.

More Vote

Renesas Electronics

Product：R-Car V4H

Recommended reasons： Targeting the high-volume Level 2+ and Level 3 markets, the Renesas R-Car V4H is tailored for central processing of Advanced Driver Assistance (ADAS) and Autonomous Driving (AD) solutions. With 34 TOPS (trillion operations per second) of deep learning performance, it enables high-speed image recognition and processing of surrounding objects from automotive cameras, radar and LiDAR. Thanks to its highly integrated performance, the R-CarV4H allows customers to develop cost-competitive single-chip ADAS electronic control units (ECUs). These ECUs support autonomous driving systems for Level 2+ and Level 3, including full NCAP 2025 functionality. The dual V4H configuration enables seamless upgrades and fail-over operation of the system performance required for Level 3 systems.The R-Car V4H also supports surround view and auto-parking functionality with excellent 3D visualization for realistic rendering.The R-Car V4H development process delivers ASIL-D system capabilities for all safety-related IP. The signal processing part of R-Car V4H is expected to achieve ASIL-B and D targets in the real-time domain. Engineers using the R-Car V4H family for automotive development will be able to take full advantage of the new SDV development environment offered by Renesas. For example, the R-Car Open Access (RoX) SDV platform provides the various core components required by automotive developers, including hardware, operating system (OS), software and tools, to efficiently develop next-generation vehicles with safety and continuous software upgrades. In addition, the platform is equipped with a software development kit (SDK) that enables developers to validate and optimize their models and test their applications in the cloud, making the development of “software-defined vehicles” even easier.

More Vote

Altera Corporation

Product：Altera Agilex™ 5 FPGA and SoC

Recommended reasons： The Agilex™ 5 product families are mid-range FPGAs and SoCs that are optimized for applications requiring high performance, lower power, and small package sizes: • Up to 2X better fabric performance/watt (Agilex™ 5 D-Series vs. competing 7nm FPGAs) • 2.5X Higher Performance (Agilex™ 5 E-Series vs. prior generation Cyclone^® V devices) • 50% Lower Power (Agilex™ 5 E-Series vs. prior generation Cyclone^® V devices) • 37X Peak INT8 TOPS (Agilex™ 5 E-Series vs. prior generation Cyclone^® V devices) Agilex™ 5 FPGAs are the first FPGA products with logic fabric infused with AI tensor blocks. The AI Tensor blocks enable implementation of AI models for faster inference, supporting advanced autonomous systems, real-time analytics, and edge AI inference applications. The AI Tensor blocks also support traditional variable precision DSP functions and provide optimized performance for complex DSP and AI applications. This advancement allows Agilex™ 5 FPGA to achieve up to 37X peak INT8 TOPS improvement, further optimizing AI workloads and digital signal processing tasks. Developers can use industry leading tools such as FPGA AI Suite and DSP Builder for seamless AI and DSP design. Agilex™ 5 FPGAs are also the first FPGAs to feature ARM v8.2 with asymmetric multi-core application processors, including dual A76 and dual A55 cores, and the first hard processor with L3 cache, TSN, USB3.1, and DMA. These peripherals enable customers to integrate more processing subsystems and features into the Agilex 5 device with the benefits of: • Increased system performance • Reduced total system power consumption by offloading the HOST CPU functions to the embedded processor subsystem within the FPGA. • Reduced board size • Reduced system cost Agilex™ 5 devices are the first Mid-Range FPGAs and SoCs to support LPDDR5 external memory interfaces. It can support external DDR5 memory data rates up to 4,000 Mbps and LPDDR4/5 memory data rates up to 4,267 Mbps. These high-performance DDR memory interfaces are crucial for applications requiring extensive data processing and transfer, making Agilex 5 devices ideal for high-performance computing, data-intensive tasks, and memory-intensive AI functions. Agilex™ 5 devices further enhance connectivity with high-speed transceivers capable of running data rates up to 28.1 Gbps and support for PCIe 4.0 x8, providing superior bandwidth for networking and data-intensive applications. This makes Agilex™ 5 products particularly suited for applications requiring high-speed connectivity and data exchange. Agilex™ 5 devices also integrate the MIPI D-PHY interface for high-speed video and imaging applications. Agilex™ 5 devices also offer Altera’s highest level of security with the latest SDM (Security Device Manager). The SDM is designed to safeguard data integrity and prevent unauthorized access by supporting advanced features such as Bitstream Encryption, Physical Unclonable Function (PUF) Key Storage, Black Key Provisioning, and Post-Quantum Cryptography (PQC). These capabilities and advanced features make the Agilex™ 5 product families ideal for mid-range FPGA applications across the edge and core. The applications span across many market segments, including wireless and wireline communications, video and broadcast equipment, industrial automation, test and measurement, medical electronics, and data centers.

More Vote

ASR Microelectronics Co., Ltd.

Product：ASR1903 5G RedCap chip platform

Recommended reasons： ASR Microelectronics believes that RedCap products have the potential to benefit a wide range of industries. From an application perspective, both the industrial and consumer markets present significant demand opportunities, including areas such as the industrial internet, power industry, video surveillance, vehicle networking, and wearable devices. The ASR1903 RedCap chipset platform from ASR adopts an innovative architecture that differs from traditional industry solutions which use separate baseband and RF transceivers. Inheriting ASR’s innovative chip design from its 4G products, the ASR1903 integrates traditionally discrete components—including RF, baseband, and memory—into a single chip. It also supports two software platforms: RTOS and OpenWRT, meeting the diverse needs of 5G IoT and wearable terminals. This high level of integration simplifies PCB design for customers and brings significant cost advantages. Leveraging its forward-looking strategy, ASR has taken the lead in foundational chip technology R&D and has successfully achieved mass commercialization. Based on the 5G RedCap platform, commercial terminals powered by ASR1903 now include mobile broadband devices, MiFi, IP cameras, power system DTUs, trackers, and modules. To date, partners include more than 20 module manufacturers, 6 MiFi brands, and 3 tracker brands.

More Vote

Mobileye Global Inc.

Product：EyeQ™6 High (EyeQ™6H)

Recommended reasons： The Eye™Q6 High (Eye™Q6H) is Mobileye’s latest-generation system-on-chip (SoC) designed to power L2+ and higher-level driving automation systems. EyeQ6H SoC is a major leap on two fronts: a seamless path to full autonomy, and a design that enables greater synergy between added functionalities. It’s a design that consolidates safety-critical functions and additional ECUs—like parking applications and visualization—into a single system-on-chip, essentially producing a centralized interconnected system. This gives automakers a wide range of choices in the components they wish to add, as they gravitate toward designs that offer more flexibility, centralization, faster time to market, and new revenue streams. It’s the EyeQ6H that forms the foundation of this shift. When it comes to autonomy, this setup is what takes the industry further, to a more software defined focus. Automakers build on core ADAS safety features while transitioning to more advanced driver assistance. They can start with Mobileye Surround ADAS, powered by a single EyeQ™ 6H, move up to Mobileye SuperVision™ with two EyeQ™ 6H SoCs—enabling point-to-point hands-free driving on most roads—and scale to Mobileye Chauffeur™ or Mobileye Drive™, unlocking hands-off, eyes-off driving for consumer or commercial offerings respectively. A fully scalable path to autonomy, built to evolve.

More Vote

All Winner Technology Co., Ltd.

Product：A733High-Performance Octa-Core AI Platform Chip

Recommended reasons： A733 is a high-performance platform processor series specifically designed for mid-to-high-end tablets and interactive display applications.The A733 integrates Dual-Core Arm^® Cortex^®-A76 and Hexa-Core Arm^® Cortex^®-A55 CPUs, along with the BXM-4-64 MC1 GPU, a 3 TOPS NPU, and support for LPDDR4, LPDDR4x, LPDDR5 DRAM and UFS/eMMC storage. The A733 supports a rich set of peripheral interfaces, High-speed interfaces includings PCIe 3.0, USB 3.1 Gen2 DRD, GMAC, Video output interfaces includings RGB, LVDS, MIPI-DSI, E-ink (electronic paper display), and eDP 1.4b/DP 1.4. It enables dual independent or synchronized display output, supporting up to 8K@24fps decoding for H.265/VP9/AVS2 video formats and up to 4K@30fps encoding for H.265/H.264. 1.1.1Key Competitive Advantages: 1.High performance & computing power: A76×2 + A55×6, up to 2GHz 2.3 Tops NPU for AI applications (AISR, gesture control, AI-ISP, etc.) 3.Supports high-speed storage (UFS 3.0) and large-capacity LPDDR5 (up to 16GB) 4.2K@120Hz high refresh rate with versatile display interfaces (Dual MIPI-DSI/Dual-LVDS/RGB/E-ink/eDP 1.4 & DP) 5.Advanced video codec support: Decoding: 4K@60fps (VP9/H.265/AVS2), 4K@30fps (H.264) Encoding: 4K@30fps (H.265/H.264) 6.4K + 2K dual-screen independent display 7.High-speed interfaces: PCIe 3.0 & USB 3.1

More Vote

Witmem

Product：Witmem CIM Chip WTM8600

Recommended reasons： With the advancement of edge-side AI technology, AI/VR/XR glasses are no longer simple display devices but have evolved into intelligent terminals equipped with on-board AI processing capabilities. As a result, consumer expectations regarding the AI performance of such devices have significantly increased. On one hand, the required AI computing power has risen from an initial 1–2 TOPS to approximately 10 TOPS in order to support local inference for larger and more complex algorithm models. On the other hand, display technologies continue to improve, with VR devices now commonly featuring resolutions approaching 2K or 4K and refresh rates ranging between 90Hz and 120Hz. However, mainstream solutions based on the traditional von Neumann architecture have increasingly revealed performance limitations due to the growing volume of data being transferred back and forth, which negatively impacts both latency and power consumption. To overcome this challenge, Witmem Technology has developed the world's first display enhancement coprocessor chip—WTM8600—based on 3D CIM architecture, with sampling scheduled for 2024. This solution offers significant advantages in terms of computing power, energy efficiency, and cost-effectiveness. While typical market solutions provide less than 10 TOPS of computing power, the WTM8600 delivers up to 24 TOPS. When integrated with standard control chips and running multiple algorithms—including AI super-resolution and AI frame interpolation—the system maintains efficient computing performance while keeping power consumption within approximately 1W (depending on the specific algorithm). By offloading tasks from the main control chip and reducing overall power consumption, the WTM8600 significantly enhances visual performance, doubling the display frame rate from 60FPS to 120FPS and upgrading resolution from 2K to 4K. Product Specifications: • Computing Power: Up to 24 TOPS @ INT8 • Computing Energy Efficiency Ratio: >40 TOPS/W • Storage-Computation Array Size: 32MB (supports storage of multiple and larger AI models) • Computing Precision: Supports 8-bit, 10-bit, and 12-bit fixed-point operations • Package and Dimensions: FCBGA package, 15mm x 15mm, utilizing 3D packaging technology • Supported AI Models: AI super-resolution, AI frame interpolation, resampling, YoloV5, ResNet, and others

More Vote

Axera Semiconductor Co., Ltd.

Product：Axera Neutron Mixed-precision NPU

Recommended reasons： Dynamically selects numerical precision, based on varying computational requirements, our Axera Neutron NPU employs a multithreading, heterogeneous, and multi-core design that tightly integrates memory with processingunits. This architecture delivers fundamental efficiency gains through two key mechanisms: dramatically reducing computational overhead by optimizing for neural network operations, and minimizing unnecessary data transfers through memory hierarchy design. Additionally, Axera Neutron NPU has native compatibility with mainstream AI models—including those based on Transformer and convolutional neural networks (“CNN”) architectures—ensures largescale seamless deployment of AI models across edge and endpoint devices. This dual capability of high efficiency and broad compatibility establishes our technology as a versatile hardware foundation for next-generation AI applications.This mixed-precision approach can provide several times the computational power of traditional NPUs within the same chip area while also greatly reducing chip costs and power consumption, making it more favorable for deploying quantized models at edge and on devices. Such flexibility and optimized performance also allow our NPU to cater to a wide array of edge and endpoint AI applications, such as smart visual perception, smart vehicle, AI-powered home hub, and industrial automation, where power efficiency and cost-effectiveness are critical success factors.

More Vote

Texas Instruments

Product：AM275x-Q1 Audio MCU

Recommended reasons： AM275x-Q1 automotive audio MCU helps reduce the number of components required for an automotive audio amplifier system by integrating TI's vector-based C7x DSP core, Arm cores, memory, audio networking and a hardware security module into a single, functional safety-capable SoC. The C7x core, coupled with a matrix multiply accelerator, together form a neural processing unit that processes both traditional and edge AI-based audio algorithms. These automotive audio SoCs are scalable, allowing designers to meet memory and performance needs, from entry-level to high-end systems, with minimal redesign and investment.

More Vote