Building Smarter, Faster: How Arm Compute Subsystems Accelerate the Future of Chip Design

If you’ve built a product on Arm technology, there’s a good chance you’re already harnessing the power of Arm Compute Subsystems (CSS). 

From the smartphone in your pocket to the servers running your favorite apps, Arm CSS enables:

• Faster time-to-market
• Optimized performance
• Reduced risk

Arm CSS isn’t a single, static product; it’s a modern way to deliver Arm’s technology portfolio, purpose-built for key markets like mobile, infrastructure, and automotive. With more than 16 design wins and several repeat orders, momentum is accelerating rapidly. In this blog, we’ll break down what Arm CSS actually is, why it matters, and how it powers innovation across industries.

AI Is Transforming Everything – Starting with Silicon

From LLMs in the cloud to edge inference, AI is redefining how silicon must be built. Chipmakers are facing rising complexity, shrinking timelines, and pressure to innovate at unprecedented speed. 

That’s why we introduced Arm CSS. Since our initial launch in early 2023, Arm CSS has enabled partners to cut time to first silicon by up to 12 months, save tens of millions of non-recurring engineering (NRE) costs and shift engineering focus from integration to true innovation and differentiation. 

By combining optimized compute and system IP into pre-integrated, market-tuned subsystems, Arm CSS simplifies design cycles while delivering best-in-class performance. Whether partners are building for the cloud, the edge, or safety-critical systems, Arm CSS enables them to go faster, with more flexibility and a future-ready foundation for AI. 

Arm CSS Variants 

Cloud & Data Center: Arm Neoverse CSS

For hyperscale cloud providers, delivering optimal compute at scale is everything. Performance is critical, but energy efficiency and thermal design power (TDP) directly impact operational expenditure, as do manageability, reliability and ease of deployment. All this in a competitive environment where falling behind on any given generation of improvements simply isn’t an option. Infrastructure SoCs are some of the physically largest built, with high core counts and matching scale-ups of memory, interconnect and associated peripherals.

Arm Neoverse CSS is designed with this in mind:

• Tens – to hundreds – of identical highly performant and efficient Neoverse CPU cores
• Arm’s uniquely scalable CMN mesh interconnect and key system control components
• Delivered as a pre-integrated, pre-verified RTL package with configuration options to further optimize PPA 

Neoverse CSS accelerates the development of high-performance infrastructure solutions and can cut time to first silicon by over a year.  With Neoverse CSS, partners can focus on delivering performance and efficiency at scale without having to manage complexity.

Companies including ADTechnology, Microsoft, Neureality and XSight Labs and others are all leveraging Neoverse CSS to build a new generation of silicon for the AI-defined cloud data center and network build-outs capturing headlines – and capital – across the globe.

Mobile: Arm Lumex CSS 

From mobile solutions to AR/VR wearables, today’s consumer devices demand peak performance, power-efficiency in battery-constrained systems and robust differentiation. Arm Lumex CSS (coming soon) enables SoC designers and OEMs to deliver on all three, faster.   

• Featuring modular and highly configurable components, Lumex CSS gives chip designers the freedom to tailor solutions to a wide range of form factors and workloads. Arm can work with partners to create customized physical cluster implementations according to customers’ product goals.
• The latest Armv9 CPU cores combined with Mali GPU cores, optimized physical IP (including ready-to-tape-out GDSII implementations), and fast cache memories accelerate time-to-market and deliver best-in-class performance on the latest cutting-edge process nodes.
• Partners benefit from performance-leading reference points, shortening product cycles and ability to scale across device tiers, while consumers will gain more responsive, power-efficient and immersive experiences.

Arm Lumex CSS is delivered in a way that fits the needs of our mobile partners for their solutions using either component level configurations or utilize the CSS to its full capability.

Our current generation of CSS for Client is being adopted by leading mobile partners worldwide such as Samsung Flip 7 smartphone, enabling them to hit flagship performance targets while accelerating time-to-market and reducing costly silicon spins. With CSS as a robust starting point, partners can concentrate their R&D efforts on tailoring features, and driving differentiated innovations that elevate the end-user experience.

Automotive: Arm Zena CSS  

Automotive OEMs are under intense pressure to differentiate with AI-driven unique user experiences and advanced safety features, while also accelerating development cycles to keep pace with an increasingly competitive landscape. Meeting these demands requires a trusted compute foundation that reduces integration burden and frees resources for innovation. Nearly every major automaker already relies on Arm in their vehicle platforms, making it the natural choice for next-generation automotive compute.

Built on Armv9 Automotive Enhanced (AE) technology, Arm Zena CSS delivers:

• A pre-integrated, safety-capable platform with high-performance compute, safety island, and runtime security engine that cuts chip design by up to 12 months and accelerating software readiness by up to two years.
• A scalable compute foundation from ADAS, central compute, and IVI use cases for both monolithic and chiplet-based silicon designs, maximizing reuse across platforms and performance tiers.
• A standardized compute platform with seamless integration of accelerators and partner-specific logic, enabling the automotive ecosystem to build on a proven foundation while tailoring SoCs for evolving AI-capable workloads.

With Zena CSS, partners can bring AI-defined vehicles to market faster and with greater confidence, supported by a safety-certified, industry-trusted compute platform. Leading OEMs – including top EV manufacturers – and major silicon providers have already licensed Zena CSS or are in advanced stages of engagement. 

IoT and edge AI: The Armv9 Edge AI Platform 

From smart thermostats to advanced vision-based robotic systems, the IoT landscape is vast and varied, requiring speed, scalability and flexibility. For industrial-grade AI at the edge, the Armv9 Edge AI Platform brings together the Arm Cortex-A320 CPU and Ethos-U85 NPU, able to run models with more than 1 billion parameters, to deliver scalable performance for real-time inferencing and decision-making capability in power- and space-constrained environments.

This platform, with Armv9 security and AI compute features, gives developers a faster path to deploying AI at the edge, backed by Arm’s extensive ecosystem.

The road ahead

The future of compute is becoming more distributed, more specialized and more central to every industry. From cloud to edge, a new generation of silicon innovation is emerging – much of it built on Arm. Arm CSS is the realization of our platform-first vision for the AI era — delivering compute that’s optimized not only for performance, but flexibility, scalability, and real-world impact. 

As workloads evolve, so do our platforms. Arm CSS is helping partners accelerate time-to-market, achieve significant performance and efficiency gains, and unlock scalable innovation across markets. With the most flexible, power-efficient compute platform on the planet, we’re creating long-term value for the ecosystem and powering the ideas defining tomorrow.

Find out more about Arm CSS for your market today – and start your journey to a future built on Arm.