Arm, the British chip and software design powerhouse, staged its Arm Everywhere China showcase in Shanghai this week, unveiling a suite of new AI‑focused technologies aimed at reshaping graphics on mobile devices. The company’s vision is to build a complete ecosystem for AI‑native computing, anchored by the upcoming AI‑centric compute platform CSS for Mobile 2 and the freshly announced Arm Mali G2‑Ultra NX GPU. This GPU is engineered around neural‑accelerated rendering, promising higher visual fidelity while keeping heat and power draw to a minimum.
While Arm did not name specific handset models that will ship with the Mali G2‑Ultra NX, the predecessor Mali G1‑Ultra already powers smartphones from brands such as Xiaomi, Vivo and OnePlus, suggesting that the new silicon will soon appear in a similar tier of devices. Deyan Lazarov, a senior member of Arm’s GPU product team within the Edge AI division, explained that the G2‑Ultra NX has been in development for roughly two and a half years. Its primary goal is to give mobile developers the ability to harness compute‑heavy effects—most notably ray tracing—without the prohibitive power and thermal costs that have traditionally kept those features confined to PCs and consoles.
"We’re seeing a real uptick in interest for ray tracing on phones," Lazarov said. "It was a slow burn at first, but now more studios are experimenting with it and asking for hardware that can support it efficiently." The new GPU integrates tightly with popular development tools.
Arm has already demonstrated the chip’s capabilities using Unreal Engine’s MegaLights feature, and it partnered with Sumo Digital to produce a demo title called *Neural Dawn*. The demo showcases how the Mali G2‑Ultra NX can deliver desktop‑class visual quality on a handheld form factor. Lazarov emphasized that the chip is meant for games that are "mobile‑first but want to look great—like desktop titles," as well as cross‑platform experiences that need to deliver a consistent look and feel whether they run on a console at home or on a train abroad.
Among the early adopters are titles such as Tencent’s *Arena Breakout Infinite*, NetEase’s *Where Winds Meet* and Infold Games’ *Infinity Nikki*. These games will benefit from the chip’s ability to balance performance, visual fidelity and battery life—a triad that is notoriously hard to achieve on battery‑powered devices. "Phones have very different thermal envelopes and power budgets compared to consoles," Lazarov noted. "They can’t dump heat the way a PC can, and they’re always running off a battery, so efficiency is paramount." Arm’s answer to those constraints is a hybrid rendering pipeline that blends conventional rasterisation with neural‑network‑based up‑scaling and frame generation.
By offloading portions of the image reconstruction to dedicated neural accelerators, the GPU workload is reduced, which in turn lowers overall power consumption. "That extra efficiency frees developers from many of the trade‑offs they normally have to make," Lazarov explained. "They can choose to reinvest the saved budget into sharper textures, higher frame rates, longer battery life, or any combination of those." To make the technology accessible, Arm is taking an open‑software stance. The company has released a toolbox that includes Neural Super Sampling (NSS), Neural Super Sampling and Denoising (NSSD) and Neural Frame Rate Upscaling (NFRU).
These modules can upscale lower‑resolution images, add extra rays for better lighting, and synthesize intermediate frames, all through AI reconstruction. Lazarov illustrated the process with a two‑frame example: in the first frame the GPU renders only 25 % of the pixels while the neural accelerator fills in the remaining three‑quarters via NSS; in the second frame the entire image is produced by the neural engine alone, with no traditional rendering work at all. Nvidia’s competing DLSS 5 technology also uses AI to enhance image quality, but Arm positions its approach as fundamentally different.
"Our networks are not generative; they’re convolutional, meaning they don’t hallucinate new content—they simply know how the final image should look based on real training data," Lazarov said. Arm has made all of its models publicly available on GitHub and Hugging Face, allowing developers to fine‑tune them with datasets that reflect their unique art styles.
This openness ensures that studios retain full control over their intellectual property and can avoid the "black‑box" pitfalls associated with Nvidia’s proprietary solution. Lazarov contrasted the two philosophies: "DLSS is a closed, proprietary system—you can’t see what it’s been trained on or retrain it for your own game.
It’s a take‑it‑or‑leave‑it proposition. Our approach is take‑it‑and‑make‑it‑work‑for‑you." Another key distinction is the mobile‑first design of Arm’s solution. The networks have been engineered to operate within the strict power, bandwidth and memory limits of smartphones. "If you tried to run DLSS on a phone, you’d instantly blow past the power envelope," Lazarov warned.
Arm also highlights substantial reductions in memory traffic: the Mali G2‑Ultra NX cuts DRAM accesses by roughly 70 %, a major win because external memory reads are a primary source of power draw on mobile SoCs. In practical terms, the new chip can deliver four‑times the performance and efficiency of its predecessor, while keeping thermal output low enough for thin, fan‑less devices.
Performance benchmarks illustrate the leap. Running *Neural Dawn* on the previous generation GPU capped at about 40 fps under ideal conditions and dropped to 15 fps in typical scenarios. The G2‑Ultra NX, by contrast, comfortably sustains 60 fps, delivering a fluid experience even with AI‑enhanced rendering active. For developers, Arm is providing Unreal Engine plug‑ins now and plans to add Unity support later, having already announced a partnership with Unity China.
An SDK is also available for studios that use custom engines, offering a streamlined integration path. Lazarov said that retrofitting an existing shipped title with Arm’s AI‑native graphics can take as little as three months, enabling rapid upgrades without a full remake. The *Neural Dawn* demo itself was built by a small team of fewer than 20 people in under 18 months, underscoring how Arm’s tools can accelerate development cycles. "Our tech lets developers quickly and easily boost the visual quality of existing games, improve efficiency, or pursue any creative direction they choose," Lazarov concluded.
"It also empowers them to ship groundbreaking new titles on a much tighter schedule."