What is the primary innovation of the E-Series architecture for edge computing?

The E-Series architecture represents a significant innovation by converging graphics, compute, and AI capabilities onto a single, flexible architecture with one programmable software stack. This design allows for the concurrent execution of graphics and AI workloads, marking a new era for edge system design. It provides advanced graphics acceleration, acts as a flexible co-processor for AI, and serves as a general-purpose accelerator capable of handling evolving AI models, enhancing overall system efficiency and versatility.

When is the Imagination E-Series silicon expected to be released to the market?

The E-Series was officially announced in 2025, with the first silicon expected to be available to system designers by the end of 2026. This timing positions it as a forthcoming solution for advanced edge system designs. Its introduction is anticipated to bring significant advancements in integrating graphics, general-purpose compute, and AI processing capabilities into a unified and efficient architecture, thereby enabling more sophisticated applications at the edge.

What are the key performance specifications and capabilities of the E-Series architecture?

The E-Series architecture boasts significant performance capabilities, offering up to 32 TOPS Int8 per core at 1GHz. This represents a substantial increase in raw compute performance compared to previous generations, specifically four times that of the D-Series. A key feature is its ability to run graphics and AI workloads concurrently, providing advanced graphics acceleration for user interfaces and games, while also serving as a flexible co-processor and general-purpose accelerator for evolving AI models at the edge.

← Back to front page

Robotics, Hardware & InfrastructureWednesday, September 30, 2026

E-Series GPU IP: The First Step Towards Converged Acceleration

Original reporting by Semiconductor Engineering

E-Series refers to a groundbreaking new architecture designed to integrate graphics, general compute, and artificial intelligence workloads onto a single, flexible silicon design and unified software stack. Announced in 2025 with first silicon expected by the end of 2026, this innovation from Imagination Technology signals a significant shift in edge system design, promising unparalleled efficiency and versatility for a wide range of applications.

This new series dramatically boosts processing power, delivering up to 32 TOPS Int8 per core at 1GHz, quadrupling the raw compute performance of its predecessor, D-Series. Critically, E-Series enables the concurrent execution of demanding graphics and AI tasks, a capability increasingly vital for modern edge devices, from advanced gaming systems to sophisticated user interfaces powered by real-time AI.

Redefining Edge Capabilities

For system designers, E-Series offers a multi-faceted solution. It provides robust graphics acceleration for high-fidelity visuals while simultaneously acting as a flexible co-processor, capable of offloading dedicated AI accelerators during peak demand. Furthermore, its general-purpose nature ensures it can reliably run the evolving AI models of the future, offering a degree of future-proofing in a rapidly changing technological landscape. This article delves deeper into the architectural innovations, power efficiency, and real-world performance data across gaming, neural rendering, and language models that position E-Series as a pivotal advancement in intelligent edge computing.

Imagination’s E-Series IP represents a pivotal moment for edge computing, moving beyond specialized accelerators to offer a truly integrated solution for complex, real-world demands. By merging graphics, general-purpose compute, and dedicated AI capabilities onto a single, flexible architecture with a unified software stack, E-Series promises to simplify development and deployment for system designers. Its ability to deliver substantial AI compute while concurrently managing advanced graphics workloads, all within edge power envelopes, signals a significant leap in on-device processing versatility.

Reshaping the Edge

The implications of this convergence extend far beyond raw performance metrics. E-Series paves the way for a new generation of intelligent edge devices, fostering a future where sophisticated AI can operate seamlessly alongside rich user interfaces and demanding computational tasks without compromise. This integrated approach will accelerate the development of more autonomous systems, highly interactive XR experiences, and smart infrastructure that processes data locally with greater efficiency and privacy. As AI models continue to evolve and diversify, E-Series offers a robust, future-proof foundation, enabling adaptable and powerful edge-native intelligence across industries.

Frequently asked questions

What is the primary innovation of the E-Series architecture for edge computing?: The E-Series architecture represents a significant innovation by converging graphics, compute, and AI capabilities onto a single, flexible architecture with one programmable software stack. This design allows for the concurrent execution of graphics and AI workloads, marking a new era for edge system design. It provides advanced graphics acceleration, acts as a flexible co-processor for AI, and serves as a general-purpose accelerator capable of handling evolving AI models, enhancing overall system efficiency and versatility.
When is the Imagination E-Series silicon expected to be released to the market?: The E-Series was officially announced in 2025, with the first silicon expected to be available to system designers by the end of 2026. This timing positions it as a forthcoming solution for advanced edge system designs. Its introduction is anticipated to bring significant advancements in integrating graphics, general-purpose compute, and AI processing capabilities into a unified and efficient architecture, thereby enabling more sophisticated applications at the edge.
What are the key performance specifications and capabilities of the E-Series architecture?: The E-Series architecture boasts significant performance capabilities, offering up to 32 TOPS Int8 per core at 1GHz. This represents a substantial increase in raw compute performance compared to previous generations, specifically four times that of the D-Series. A key feature is its ability to run graphics and AI workloads concurrently, providing advanced graphics acceleration for user interfaces and games, while also serving as a flexible co-processor and general-purpose accelerator for evolving AI models at the edge.

Intro and outro generated by Printing Press AI from the source article above. Always consult the original reporting for verbatim quotes and primary sources.