Arm, the British chip and software design powerhouse, used its Arm Everywhere China showcase in Shanghai to announce a suite of new AI‑focused technologies aimed at reshaping graphics on mobile devices. The centerpiece of the announcement is a complete AI‑native computing ecosystem that includes the second‑generation AI‑native compute platform CSS for Mobile, as well as the freshly unveiled Arm Mali G2‑Ultra NX GPU.
This GPU is built around dedicated neural processing units and is marketed as delivering desktop‑class visual fidelity while keeping power draw and heat generation far lower than traditional mobile graphics solutions. While Arm did not disclose the exact product line‑ups that will ship with the Mali G2‑Ultra NX, the predecessor Mali G1‑Ultra has already found its way into smartphones from brands such as Xiaomi, Vivo and OnePlus, suggesting that the new chip could appear in a similar range of high‑end Android handsets. Deyan Lazarov, a senior engineer on Arm’s GPU team within the Edge AI division, explained that development on the G2‑Ultra NX has spanned roughly two and a half years. The goal, he said, was to give mobile developers a practical path to leverage compute‑heavy techniques—most notably real‑time ray tracing—without sacrificing battery life or thermal headroom.
"Ray tracing has started to gain traction in mobile games," Lazarov noted. "It moved slowly at first, but we’re now seeing a growing number of titles that want to enable it. Our hardware is designed to make that feasible." The new GPU works hand‑in‑hand with Unreal Engine features such as MegaLights, and Arm has already partnered with Sumo Digital to produce a demo title called Neural Dawn that showcases the chip’s capabilities.
Lazarov described the target audience for the Mali G2‑Ultra NX as developers who adopt a "mobile‑first" philosophy but still aspire to deliver visuals that rival those on PCs or consoles. He added that many games are built to be cross‑platform from the outset, and players expect a seamless experience whether they are gaming on a console at home or on a train with a phone. "The challenge for phones is their thermal envelope and power budget," he said. "Unlike a console, a phone can’t dump heat as easily and runs on a battery, so we need to be clever about how we use the GPU." Arm’s answer is to blend neural processing with the conventional rendering pipeline.
By offloading portions of the image generation to dedicated AI accelerators, the GPU workload is reduced, which in turn lowers power consumption and improves overall system efficiency. "That extra efficiency means developers don’t have to make as many compromises," Lazarov explained.
"They can choose to reinvest the saved budget into higher visual fidelity, smoother frame rates, longer battery life, or any combination of those goals." To make this vision attractive to creators, Arm is taking an open‑source stance. The company is releasing a toolbox that includes Neural Super Sampling (NSS), Neural Super Sampling and Denoising (NSSD), and Neural Frame Rate Upscaling (NFRU). These technologies can upscale lower‑resolution images, add extra rays for better lighting, and generate additional frames using AI‑driven reconstruction. In a typical workflow, the GPU might render only a quarter of the final image while the neural accelerators reconstruct the remaining three‑quarters via NSS.
In a more extreme mode, the entire frame can be produced by the AI engine, eliminating traditional rasterisation for that frame entirely. Arm’s approach stands in stark contrast to Nvidia’s DLSS 5, which has sparked debate because its generative neural networks can sometimes diverge from a developer’s artistic intent.
Lazarov emphasized that Arm’s models are convolutional rather than generative; they do not invent new content but instead learn how existing content should look. All of Arm’s networks—resolution upscalers, frame‑rate generators, and denoisers—are publicly available on platforms such as GitHub and Hugging Face.
This openness allows developers to fine‑tune the models with data from their own games, preserving artistic style and intellectual‑property control. "DLSS is a closed, black‑box solution," Lazarov said. "You can’t see what it’s been trained on, nor can you retrain it with your own assets.
With Arm, you get the code, you can retrain it, and you retain full control over the output." Another key differentiator is the mobile‑first design philosophy. Arm’s networks are engineered to respect the strict power, memory‑bandwidth, and thermal limits of smartphones. By contrast, attempting to run DLSS on a phone would quickly exceed those budgets, making it impractical. Performance metrics underline the efficiency gains.
Arm claims the Mali G2‑Ultra NX delivers up to four times the performance and four times the efficiency of its predecessor, while slashing DRAM traffic by roughly 70 %. Since external memory accesses are a major source of power draw on mobile SoCs, this reduction translates directly into longer battery life.
In practical terms, the difference is striking. When pushing last year’s GPU to its limits, the demo game Neural Dawn could only reach about 40 fps under ideal conditions and dropped to roughly 15 fps in typical play.
The new G2‑Ultra NX, however, runs the same title smoothly at 60 fps, delivering a fluid experience without overheating. Arm is also investing in developer tooling. Plug‑ins for Unreal Engine are already available, and a Unity integration is on the roadmap, complemented by a recent partnership with Unity China.
For studios using custom engines, Arm offers a comprehensive SDK that streamlines integration without requiring a dedicated plug‑in. Lazarov highlighted that retrofitting existing games with Arm’s AI‑native graphics can be surprisingly quick.
"If you have a shipped title and you want to add our solutions, you can be ready to ship an updated version in as little as three months," he said. This rapid turnaround is possible because the underlying AI models are open and can be retrained to match a game’s unique visual style, eliminating the need for extensive re‑authoring. The Mali G2‑Ultra NX is already slated to power several upcoming titles, including Tencent’s Arena Breakout Infinite, NetEase’s Where Winds Meet, and Infold Games’ Infinity Nikki.
These games will serve as early showcases of how mobile‑first AI reconstruction can deliver console‑like graphics without compromising the handheld experience. In summary, Arm’s mobile‑first AI reconstruction strategy offers an open, developer‑friendly alternative to Nvidia’s proprietary DLSS 5. By integrating neural accelerators directly into the graphics pipeline, providing publicly accessible models, and optimizing for the stringent constraints of smartphones, Arm aims to empower developers to create richer, more efficient games that run beautifully on today’s mobile hardware.