Altera’s 25G Holoscan Sensor Bridge Demo Wins Embedded Computing Design Best-in-Show Award
Altera, widely recognized as one of the world’s leading providers of field-programmable gate array (FPGA) solutions, has announced that its innovative 25GbE Holoscan Sensor Bridge demonstration received the prestigious Best-in-Show Award in the Artificial Intelligence and Machine Learning category at embedded world 2026. The award was presented by Embedded Computing Design, a well-known industry publication that evaluates cutting-edge embedded technologies and systems displayed at the global event.
The recognition highlights Altera’s efforts to advance high-performance data connectivity and sensor processing for next-generation AI applications. The winning demonstration showcased a powerful solution powered by Agilex 5 SoC FPGA, illustrating how high-bandwidth sensor data can be efficiently transmitted and processed in modern edge AI environments.
Recognition at Embedded World
Embedded World is one of the most influential gatherings in the embedded technology sector, bringing together semiconductor companies, hardware manufacturers, software developers, and system integrators from around the globe. Each year, the Best-in-Show Awards recognize the most impressive innovations unveiled during the event. Products and demonstrations are judged by the editorial team of Embedded Computing Design based on several criteria, including engineering innovation, product performance, design excellence, and potential impact on the embedded technology market.
Receiving the award in the AI and machine learning category places Altera’s solution among the most notable technology demonstrations at the event. The recognition reflects not only the strength of the company’s FPGA technology but also its ability to address emerging challenges in data-intensive computing environments.
Introducing the Holoscan Sensor Bridge
At the center of the award-winning demonstration is Altera’s Holoscan Sensor Bridge reference design, which functions as a high-performance bridge between camera sensors and AI processing platforms. The solution enables seamless conversion of MIPI camera input streams into high-speed 25-gigabit Ethernet output, allowing sensor data to be transmitted quickly and efficiently to AI compute engines such as GPUs.
This architecture is particularly important in modern AI systems where massive amounts of sensor data must be processed in real time. Applications such as robotics, autonomous vehicles, industrial automation, and smart surveillance depend on the ability to capture, transport, and analyze high-resolution visual data with minimal delay.
The Holoscan Sensor Bridge demonstration uses multiple 4K resolution cameras operating at 60 frames per second (4Kp60). These cameras generate extremely large data streams that must be transferred to computing platforms capable of running AI models. The FPGA-based bridge ensures that this data can move rapidly from the camera interface to the processing platform without creating bottlenecks.
Role of the Agilex 5 SoC FPGA
The demonstration relies on the processing power and flexibility of Altera’s Agilex 5 SoC FPGA family. These advanced chips combine traditional FPGA programmable logic with integrated system-on-chip (SoC) processing capabilities. This combination enables developers to build complex systems that require both flexible hardware acceleration and software-controlled processing.
The Agilex 5 platform includes high-speed transceivers, programmable logic blocks, and dedicated digital signal processing (DSP) resources optimized for AI and high-performance workloads. These features allow developers to build custom data pipelines capable of handling extremely high bandwidth while maintaining low latency.
In the Holoscan Sensor Bridge design, the Agilex 5 FPGA handles multiple tasks simultaneously, including sensor aggregation, real-time data processing, and high-speed network transmission. By performing these operations directly in programmable hardware, the system avoids the delays often associated with purely software-based solutions.
High-Bandwidth Data Streaming for Edge AI
One of the biggest challenges facing modern AI deployments is the need to process enormous volumes of data generated by sensors at the edge of the network. Cameras, radar sensors, lidar systems, and other devices continuously produce streams of information that must be analyzed almost instantly in order to make real-time decisions.
Traditional computing architectures often struggle to handle this type of workload because they rely heavily on centralized processing. As a result, data must be transported long distances before it can be analyzed, introducing latency that can affect performance.
Altera’s FPGA-based architecture addresses this problem by enabling data pre-processing and aggregation directly at the edge. The Holoscan Sensor Bridge allows sensor streams to be combined, filtered, and transmitted efficiently before reaching the main AI processing platform.
This approach ensures that GPUs and other AI accelerators receive high-quality, well-organized data streams that are ready for immediate analysis. The result is faster system response times and improved overall efficiency.
FPGA Advantages in AI-Attached Architectures
FPGAs are becoming increasingly important in AI-attached architectures because they offer a unique combination of flexibility and performance. Unlike fixed-function chips, FPGAs can be reprogrammed to support new algorithms, protocols, or data processing tasks. This adaptability is particularly valuable in rapidly evolving AI environments.
In edge AI systems, FPGAs are often used to perform tasks such as sensor data aggregation, signal conditioning, data compression, and protocol conversion. By handling these operations before data reaches the main AI compute engine, FPGAs reduce the workload on GPUs and CPUs.
The result is a more efficient computing architecture that can scale to support larger numbers of sensors and higher data rates.
Expanding Holoscan Ecosystem Capabilities
Altera’s Holoscan Sensor Bridge demonstration also highlights the company’s broader strategy for supporting high-performance AI infrastructure. According to the company, it is currently the only FPGA vendor offering FPGA-based Holoscan solutions that scale from 10-gigabit Ethernet to 100-gigabit Ethernet connectivity.
In addition to these capabilities, Altera provides a range of production-ready networking solutions, including SmartNICs and data processing units (DPUs) designed for high-performance computing, cloud infrastructure, and data center deployments. Some of these solutions support network speeds of up to 400 gigabits per second, enabling extremely high throughput for demanding workloads.
This ecosystem of programmable hardware solutions is intended to help developers build scalable AI platforms capable of supporting a wide range of data-intensive applications.
Leadership Perspective
Venkat Yadavalli, head of Altera’s Business Management Group, emphasized the importance of the recognition received at Embedded World. According to Yadavalli, the award highlights the company’s commitment to advancing FPGA-based technologies that address the growing demand for real-time data processing.
He explained that modern AI systems depend on the ability to move massive volumes of sensor data quickly and efficiently to compute platforms. With Agilex 5 devices, developers can achieve the deterministic performance and flexibility needed to support these workloads while maintaining extremely low latency.
Yadavalli also noted that the Holoscan Sensor Bridge reference design demonstrates how programmable FPGA architectures can serve as a critical link between high-bandwidth sensors and powerful AI compute platforms.
Applications in Emerging Technologies
The technology demonstrated in the Holoscan Sensor Bridge has broad potential across a variety of industries. Robotics is one of the most obvious areas where high-speed sensor data processing is essential. Autonomous robots rely on cameras and other sensors to navigate complex environments and make decisions in real time.
Autonomous vehicles represent another major application. Self-driving systems must process enormous volumes of visual and sensor data to detect obstacles, track objects, and maintain safe navigation. Efficient data transfer from sensors to AI processing units is a critical component of these systems.
Industrial automation is also benefiting from advances in high-speed sensor connectivity. Smart factories increasingly rely on machine vision systems and AI analytics to monitor production lines, identify defects, and optimize manufacturing processes.
In each of these scenarios, the ability to transport high-resolution sensor data with minimal latency can significantly improve system performance and reliability.
Agilex 5 for Edge and Embedded Systems
Beyond the Holoscan Sensor Bridge demonstration, the Agilex 5 FPGA and SoC families are designed to support a wide range of embedded and edge computing applications. These devices offer industry-leading performance-per-watt, making them well suited for systems where energy efficiency is critical.
The chips support the latest memory standards, including LPDDR5 and DDR5, which enable higher data throughput and improved system responsiveness. They also feature high-speed serial transceivers that allow developers to implement advanced networking and connectivity solutions.
Security is another important focus for the Agilex 5 platform. The devices incorporate post-quantum cryptography-ready secure boot capabilities, helping protect systems against emerging cybersecurity threats.
Power Efficiency and Performance Balance
Many embedded applications require a careful balance between performance and power consumption. Systems deployed at the edge often operate in constrained environments where cooling and power resources are limited.
Agilex 5 devices address this challenge by delivering strong computational performance while maintaining efficient power usage. This combination makes them suitable for midrange applications that demand high throughput without the overhead of large data center hardware.
By integrating programmable logic, embedded processing cores, and AI-optimized DSP resources, Agilex 5 devices enable developers to accelerate a wide range of workloads while maintaining flexibility for future updates.
The Best-in-Show Award for the Holoscan Sensor Bridge demonstration represents an important milestone for Altera as it continues to expand its role in the AI and embedded computing markets. As edge AI applications continue to grow in complexity, solutions capable of managing large volumes of sensor data will become increasingly critical.
Through innovations such as the Holoscan Sensor Bridge and the Agilex 5 FPGA platform, Altera is positioning itself to help developers build scalable, high-performance AI systems that can operate efficiently across a wide range of industries.
The recognition at Embedded World highlights the growing importance of programmable hardware in modern computing architectures and underscores the role that FPGA technology will play in enabling the next generation of intelligent systems.
Source link: https://www.businesswire.com
