With the rapid growth of edge AI and high-performance computing demands, the division of roles within the processor industry is beginning to shift. Recent moves from the dominant CPU IP supplier has increased industry attention on the openness and ecosystem neutrality of the supply chain. 

Against this backdrop, the value of RISC-V CPU IP is becoming more evident. It offers chipmakers greater autonomy and flexibility, while also helping to foster a more open and stable ecosystem.  At the same time, SoC design is evolving from a focus on standalone compute performance to system-level optimization. The collaboration between RISC-V CPUs and GPUs is a key enabler for AI PCs, embodied intelligent robots, automotive, industrial and other intelligent edge devices.

The recently introduced K3 processor from SpacemiT provides a representative example of this trend in practice.

Advancing RISC-V CPUs: From General Compute to High-Performance Platforms

SpacemiT is a RISC-V-based CPU chip design company. Its K3 chip, launched this year, is built on SpacemiT’s in-house RISC-V CPU IP cores (X100), running at up to 2.4GHz. Its single-core performance is comparable to mainstream high-performance CPU architectures, delivering around 130K DMIPS of general-purpose compute. In addition, the chip integrates up to 60 TOPS INT4 of AI compute, enabling on-device inference for models at the 30B parameter scale.

This design highlights an important direction for RISC-V in high-performance SoCs: not only serving as a general-purpose compute core, but evolving into a foundational platform that integrates AI and system-level capabilities.

See full product specs in the K3 brief

GPU + RISC-V: A Key Combination for Complete Platforms

In real-world deployments, CPU and AI performance alone are not sufficient to support a complete system.  Many applications require more than compute—they demand graphical interfaces, multimedia processing, and full operating system support. This shifts the focus of RISC-V SoC design from delivering compute power to enabling complete system functionality. 

Against this backdrop, the combination of GPU and RISC-V CPU is emerging as a practical approach to building high-performance SoCs.  In the case of K3, the platform integrates the Imagination BXM-4-64 GPU IP. This enables the system to complement its CPU and AI capabilities with full graphics support.  This combination delivers value across multiple dimensions:

• Enabling full graphics environments for Linux-based systems. SpacemiT has announced the availability of Ubuntu on its K3 and K1 chips, meaning developers worldwide can benefit from seamless hardware-software integration.
• Supporting standard graphics and compute APIs, including Vulkan.
• Providing compatibility with open-source graphics stacks.
• Accelerating highly parallel workloads like camera stitching or object detection.

Software ecosystem support is the key factor that determines whether performance can translate into real usability. With RVA23 supported in the SpacemiT CPU, the K3 aligns with the mass of the RISC-V software ecosystem and, when used with Imagination GPU drivers (that are already optimised for RISC-V hosts), the result is robust operating system compatibility and support for graphics applications. 

The Value of a Mature Ecosystem

As the RISC-V ecosystem continues to expand, software readiness and system support are becoming critical success factors.  With over 30 years of development, Imagination GPUs are widely deployed across over 11Bn devices, including multiple platforms based on RISC-V CPUs. We have proven integration experience and a mature software ecosystem across embedded and computing markets. We offer tangible benefits  for RISC-V SoC developers, whether it's reducing system design complexity or accelerating time to market.

Looking Ahead: Towards a Collaborative RISC-V Ecosystem

As RISC-V continues to expand into high-performance and AI-driven domains, its ecosystem is rapidly maturing.  The ability to effectively combine compute, AI, and graphics will become a key differentiator for platform competitiveness. A GPU IP with proven performance, efficiency, and ecosystem support is essential to the future growth of RISC-V-based systems.  From an industry perspective, designs like K3 highlight the potential of RISC-V and Imagination GPUs in high-performance SoCs and point toward a future shaped by heterogeneous, system-level innovation.