Intel and Altera expand their collaboration to simplify sensor integration in physical AI

Intel Corp. and Altera Corp. announced they are jointly developing flexible, real-time "sensor-bridge" reference systems for edge AI and physical AI applications, including robotics. The idea is to combine Intel's compute, connectivity and software with Altera's FPGAs (field-programmable gate arrays). The aim is to help manufacturers manage the growing complexity of connecting, processing and acting on real-world data in increasingly sensor-heavy environments.

According to the companies, edge and physical AI systems may need to aggregate data from different sensors and interfaces while meeting timing, performance, power and cost requirements. Some applications also demand very high bandwidth or flexible form factors, with input/output (I/O) located far from the compute board. In those cases, sensor data must be moved quickly into memory so the host Intel processor can use it for edge, robotics and other workloads.

The companies plan to develop modular reference systems that combine Altera's programmable hardware with a consistent Intel compute and software foundation. The goal is to give original equipment manufacturers (OEMs) more flexibility as applications, sensors and workloads evolve, and to help them move from prototype to deployment more efficiently. The resulting solutions will be offered to developers through Intel's Open Edge Platform software development kit (SDK) and an Altera SDK.

Farhad Shafai, head of business solutions and marketing at Altera, said Altera's FPGAs provide sensor connectivity, data processing and deterministic control, and remain programmable as sensors and AI models evolve. Dan Rodriguez, corporate vice president and general manager of Intel's Edge Computing Group, said customers need a simpler way to turn real-world data into timely action.

The announcement focuses on three markets. In robotics, Intel's compute, connectivity and robotics software will be combined with Altera FPGAs to address the sensing, timing and real-time processing requirements of intelligent machines. In healthcare, they are developing a reference system aimed at the demanding latency and real-time requirements of advanced medical equipment. In aerospace and defense, they are collaborating on next-generation mission systems, focusing on size, weight and power for workloads where AI and signal processing converge.

For context, the article notes that Altera, headquartered in San Jose, California, and founded in 1983, describes itself as the largest pure-play FPGA supplier. Intel acquired it for $16.7 billion in 2015 and sold 51% of the company to Silver Lake last year.

The announcement should be treated with caution: it consists of stated plans and goals ("plan to," "aim to"), with no concrete products, dates or performance figures. A reader commented on the page itself that it is unclear what competitive advantage Intel plus Altera would offer over combinations such as AMD plus Xilinx or other high-end processor-and-FPGA pairings. The article does not answer that question.