The Lumex Compute SubSystem ( CSS), an upcoming mobile platform from Arm, has been unveiled. The brand-new Shoulder C1 CPU and Mali-G1 GPU architectures are the foundation of the new system, which enables strong, real-time on-device AI for the upcoming smartphones and tablets.
The Shoulder v9.3-based C1 CPU cores that will be offered in four different bundles will be available for the entire wireless market. The C1-Premium offers the same performance in a smaller, more effective offer while the flagship model is the C1-Ultra, promising a 25 % single-thread performance improvement over the Shoulder Cortex-X925. The C1 Pro, which promises a 16 % sustained performance increase over the Cortex-A725, and the C1 Nano, which is 26 % more effective than the Cortex-A520, are next.
embedded content ]
We have the new Mali-G1 GPU on the pictures front. This fresh architecture, which was developed for 3nm procedure networks, is scalable from 10 to 24 components and offers a 20 % improvement in overall design performance over the past generation. With the introduction of a new Ray Tracing Unit (RTUv2 ), which Arm claims can deliver up to twice the performance of its predecessor, a significant focus has been put on ray tracing.
These new CPUs will support Scalable Matrix Extension 2 ( SME2 ) as part of the Lumex platform. Arm claims that this innovative instruction set can significantly improve AI performance by up to five times and significantly lower overhead for speech-based workloads like voice assistants and live translation. Additionally, Arm is making sure that designers can benefit from this fresh AI energy with its KleidiAI books, which are already included in popular frameworks like PyTorch and Microsoft’s ONNX Runtime. This implies that some well-known apps will be able to use SME2 motion without having to change much of their code.
The Finger Lumex platform, which features its new C1 and Mali-G1 core, is expected to release maybe in 2026 for the first smartphones and tablets. Additionally, according to Arm, similar laptop computer technology will be available under the” Shoulder Niva” label.
KitGuru asserts that Arm is also strongly focused on AI, just like most chip manufacturers are at this point.  ,