Imagination debuts E-Series GPUs for edge AI

Imagination Technologies has launched its E-Series graphics processing units (GPUs), a next-generation family of processors designed to redefine how graphics and AI workloads are handled at the edge. Combining cutting-edge performance with energy efficiency, the E-Series marks a major step forward in on-device intelligence for markets like automotive, mobile, and industrial computing.

Unveiled ahead of availability in Autumn 2025, the E-Series GPUs bring together powerful graphics capabilities and deeply integrated AI acceleration. With scalable performance from 2 to 200 TOPS (INT8/FP8), these GPUs are engineered to meet the increasingly complex demands of edge AI systems—delivering both compute density and efficiency.

AI and Graphics Integration at the Core

At the heart of the E-Series are two breakthrough technologies: Neural Cores and Burst Processors.

• Neural Cores offer up to 200 TOPS of AI processing power, supporting a range of low-precision formats like INT8 and FP8. These cores deliver up to 400% more AI performance than Imagination’s previous D-Series, enabling advanced capabilities for natural language processing, computer vision, and autonomous driving.
• Burst Processors bring a 35% average improvement in power efficiency by optimizing how data moves through the GPU pipeline and memory system—critical for devices operating under tight power and thermal constraints.

“Edge AI is evolving rapidly, but system designers still struggle with the balance of performance, power, and programmability,” said Phil Solis, research director at IDC. “Imagination’s E-Series offers a unique solution by blending AI acceleration with proven GPU flexibility.”

A Future-Proof Platform for Edge Intelligence

The E-Series stands out for its programmable architecture, allowing it to support both evolving AI algorithms and next-gen graphics features like ray tracing. Developers can program the GPUs using industry-standard tools such as OpenCL, oneAPI, Apache TVM, and LiteRT, or leverage Imagination’s own optimized libraries and graph compiler.

Kristof Beets, VP of product management at Imagination, emphasized the company’s focus on smart architectural integration:

“The blending of AI components directly into the GPU allows us to efficiently balance workloads and gives us a unique edge—especially in verticals like automotive where both real-time compute and rich graphics are required.”

Efficient, Scalable, and Built for Complexity

Designed for real-world, multitasking environments, E-Series GPUs are capable of simultaneously handling multiple graphics and AI workloads. The platform now supports up to 16 hardware-backed virtual machines, enabling secure and efficient resource allocation with zero overhead. This is particularly relevant for automotive systems and high-performance consumer devices, where multitasking and quality-of-service (QoS) are critical.

The multicore variants also provide flexibility: designers can opt for raw performance or use extra cores to separate compute and rendering tasks, tailoring GPU configurations to specific device needs.

Applications Across Devices and Industries

From industrial edge computing to mobile devices and vehicle autonomy, E-Series GPUs provide a unified solution for on-device intelligence. Use cases include:

• Automotive: Advanced driver-assistance systems (ADAS), cockpit visualization, and autonomous perception
• Consumer devices: On-device voice and vision AI, immersive user interfaces
• Industrial: Smart cameras, robotics, and predictive maintenance
• Mobile and Desktop: AI-enhanced graphics, language processing, and gaming

Already Licensed, Coming in 2025

The first E-Series GPU IP has already been licensed ahead of its general availability in Autumn 2025. Imagination Technologies is actively developing automotive, consumer, mobile, and desktop variants tailored to their respective performance, power, and integration needs.

“The E-Series places the GPU at the center of edge AI and graphics,” said Tim Mamtora, chief of innovation and engineering at Imagination. “It’s not just about performance—it’s about building a flexible, future-proof platform that enables system designers to do more with less.”

With the E-Series, Imagination Technologies is not just evolving the GPU—it’s reimagining what edge computing can be.