The same day, Apple revealed its new iPhone 17 rangeChip Designer Arm introduces its next generation of processors grouped in its new Arm Lumex platform. These processors will probably go to premium Android phones to come in 2026 – and will improve their AI capacities without drainage of batteries of devices faster, says Arm.
“It’s quite surprising to see how there was a kind of insatiable performance requested [by our customers]And many things are around AI, as well as certain graphic workloads, “said Chris Bergey, principal vice-president and managing director of Arm customer customers.
Arm processors have traditionally been the centerpieces of holistic puce systems, which feed smartphones. For example, the high -end central treatment unit of ARM, the Cortex X925, published last year, is presented in the Samsung Exynos 2500 chipset, which propelled the Galaxy Z Flip 7 and the Chipset Dimensité 9400 of Mediatek, found in the Oppo Find X8 Pro.
It is likely that the new Arm chips, only or gathered in the company’s Arm Lumex platform, will supply Android Premium phones and other devices next year. But it faces more competition because Qualcomm (which used ARM processors in older chips such as the Snapdragon 8 Gen 3) moved to its internally designed in its latest silicon – which has been the subject of its own technological license confrontation in recent years between the two flea companies.
The new CPU range of Arm has changed with a new denomination paradigm. The successors of the X925 are two different chips: the most high-end C1-UTRA, which has 25% larger performance on its predecessor, and the most powerful C1-Premium (no improvement in given performance). The successor to the A725 chip is the C1-Pro, which has 12% more efficient.
These CPUs also benefit from an evolution of their chip architecture called scalable matrix extension version 2, which allows better performance of AI.
The LUMEX ARM platform, including its generation of C1 chips.
The new Arm Lumex platform combines these chips with the new GPU Mali G1-UTRA (which says Arm 20% of better performance and twice the tracing of the rays as a predecessor) for a system that can be plugged into larger chipsets. The final result: up to 5x IMP improvement of AI performance, 4.7 times the lower latency for speech -based workloads (think of live translation) and the faster 2.8x audio generation, says Arm.
The objective is faster performance without increasing the battery leakage when performing AI tasks of this tax treatment power. For example, a Yoga Tutor demonstration application running new ARM chips saw an increase of 2.4x of the vocal text to give faster comments to users, the company said in a press release.
As a supplier of chips and technologies, it ultimately belongs to phonémakers using Arm silicon to decide the quantity (or little) of the progress they incorporate into their devices. But as generative AI demand like chatgpt increases, the same goes for the reader so that this feature works on phones as effectively as possible, rather than relying on slower responses going to and from the Cloud. So what will consumers see with devices performing C1-UTRA fleas and new Arm technology?
“I think what they will see is the possibility of running an incredible AI on disk and doing so with significant energy savings, significant performance increases and also third -party support,” said Bergey. “Not just parts, but also third -party devices.”
Look at this: Everything announced during Apple’s iPhone 17 event in 7 minutes
