Reported by 1 source

The short version

  • Xiaomi has launched the first device featuring Arm's new Mali G2-Ultra NX graphics processor, which integrates an AI accelerator for gaming tasks.
  • The technology enables features similar to PC-level upscaling and frame generation, aiming to improve visual fidelity while maintaining low power usage.
  • Full implementation of all neural graphics features will roll out gradually, with frame generation capabilities expected in early 2027.

A significant shift in mobile graphics processing is underway as Arm introduces its first GPU equipped with a dedicated AI gaming accelerator. The technology debuted in mainland China with the launch of the Xiaomi 18 Fold, which utilizes the custom Xring O3 chip containing the new Mali G2-Ultra NX processor. This development represents the culmination of five years of engineering effort by Arm to integrate neural rendering capabilities directly into mobile graphics hardware. The goal is to allow smartphones to render games at higher resolutions and frame rates using less power than traditional methods require.

The core innovation lies in how the GPU handles graphical data. Rather than relying solely on raw processing power to draw every pixel, the new system uses artificial intelligence to predict and generate image details. Arm executives describe this approach as bringing tensor-like capabilities into the GPU shader environment. This allows the device to perform tasks such as upscaling lower-resolution images to appear sharper, generating intermediate frames to smooth out motion, and reconstructing lighting effects. These techniques mirror advancements previously seen in high-end personal computers, particularly those utilizing Nvidia’s DLSS technology.

News Journal

Performance metrics associated with the new chip suggest substantial improvements over previous generations. Arm claims an 85 percent performance increase compared to the GPU used in Xiaomi’s prior devices, alongside a 14 percent boost over its own previous flagship in non-AI gaming scenarios. When all neural features are active simultaneously, the system can theoretically achieve up to four times the frame rate of standard rendering. However, industry observers note that this multiplier does not translate directly to perceived speed, as half of the gain comes from generating artificial frames between existing ones, a process that can introduce input latency.

Power efficiency remains a central focus of this architectural change. Arm’s Chris Bergey, executive vice president of Edge AI, emphasized that the new GPU is designed for sustained gaming sessions without excessive heat or rapid battery drain. The entire phone system is expected to consume no more than 2.5 watts during operation, with only 1.5 watts allocated specifically to the graphics processor. This efficiency allows for longer playtimes at high settings, such as 1080p resolution at 60 frames per second for up to three hours, a significant improvement over previous limitations where such performance might last only thirty minutes.

The rollout of these capabilities will occur in stages rather than all at once. Initial software support will focus on neural super sampling, which enhances image clarity by reconstructing pixels. Frame generation technology is scheduled for release in early 2027, followed shortly by neural denoising features that improve visual quality in low-light or complex lighting scenarios. This phased approach allows developers and device manufacturers to optimize their systems for each new feature as it becomes available.

While the Xiaomi 18 Fold marks the first consumer launch, Arm indicates that this technology will expand beyond Chinese markets and flagship smartphones. Bergey suggested that devices with larger screens, such as tablets and Chromebooks, will also adopt the Mali G2-Ultra NX in the coming year. He hinted that some of Google’s upcoming hardware might utilize this graphics tech, signaling a broader industry adoption. The Xiaomi Pad 9 Pro Max tablet, which shares the same Xring O3 chip, already demonstrates the potential for higher performance on larger displays, showcasing gameplay at 3.2K resolution and 120 frames per second.

The implications for mobile gaming extend beyond raw performance numbers. By offloading traditional rendering tasks to AI algorithms, developers can create more visually complex games that run smoothly on standard hardware. This shift could reduce the need for specialized gaming phones or external cooling accessories, making high-fidelity graphics accessible to a wider audience. Arm positions this not as an overclocking feature for enthusiasts, but as a mainstream advancement in how mobile devices process visual information.

Despite the optimism surrounding these developments, challenges remain regarding software compatibility and user experience. The introduction of frame generation may affect input lag, a critical factor in competitive gaming. Additionally, the full benefits will only be realized as game developers update their titles to support Arm’s neural graphics APIs. Until then, early adopters may see limited improvements depending on the specific games they play. The industry is now watching to see how quickly third-party developers integrate these new tools into their workflows.

This technological leap underscores a broader trend in computing where AI assists in fundamental hardware functions. As mobile devices continue to bridge the gap with desktop performance, the integration of neural accelerators into GPUs represents a pivotal moment for portable entertainment. The success of Arm’s strategy will depend on whether these efficiency gains can be maintained across diverse device types and operating systems without compromising battery life or thermal management.

Looking ahead, the availability of these chips in global markets remains to be confirmed, though Arm executives have indicated that international releases are planned for next year. The transition from Chinese exclusivity to worldwide distribution will test the scalability of this new graphics architecture. If successful, it could redefine expectations for mobile gaming quality and set a new standard for power-efficient high-performance computing on small devices.

Sources behind this briefing

Go to the original reporting

  • The Verge↗First Xiaomi, then the world: why Arm might give phone gaming a huge graphics boost