Arm, the British chip and software design powerhouse, staged its Arm Everywhere China showcase in Shanghai to announce a suite of fresh AI‑computing innovations aimed at the mobile market. The centerpiece of the announcement is a new AI‑native computing stack, dubbed CSS for Mobile 2, together with the Arm Mali G2‑Ultra NX graphics processor. This GPU is built around dedicated neural‑compute units, promising higher graphical fidelity while keeping heat and power draw to a minimum. Arm positions its on‑device AI‑enhanced graphics as a more open and developer‑friendly alternative to Nvidia’s controversial DLSS 5, arguing that its method aligns better with the original intent of game creators.
While Arm did not disclose specific product partners, its predecessor, the Mali G1‑Ultra, already powers devices from Xiaomi, Vivo and OnePlus, suggesting that the G2‑Ultra NX could soon appear in a similar range of smartphones and tablets. Deyan Lazarov, a 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 design goal is to give mobile developers access to heavyweight rendering techniques—most notably ray tracing—without sacrificing the battery life or thermal envelope required for handheld devices. "There’s a clear upward trend in mobile titles adopting ray tracing," Lazarov noted.
"It was slow to get off the ground, but now more games are actively looking to enable it." The new GPU, as reported by GamesIndustry.biz, can leverage Unreal Engine features such as MegaLights, and Arm has already collaborated with Sumo Digital to produce a demo title called Neural Dawn that showcases the chip’s capabilities. Lazarov described the target audience as "mobile‑first games that still want desktop‑level visual quality" and cross‑platform experiences that deliver a consistent look whether the player is on a console at home or a phone on a train. Among the early adopters are titles like Arena Breakout Infinite (Tencent), Where Winds Meet (NetEase) and Infinity Nikki (Infold Games). He emphasized the unique constraints of phones: limited thermal headroom, strict power budgets and constant reliance on battery power, which differ dramatically from the generous cooling and power supplies of consoles and PCs.
To address these constraints, Arm integrates neural processing directly into the traditional graphics pipeline. By offloading portions of the rendering workload to dedicated AI accelerators, the GPU can render fewer pixels and let the neural engines reconstruct the missing data.
This reduces both GPU cycles and overall power consumption, delivering system‑level efficiency gains. "The result is new efficiency that eases the trade‑offs developers usually face," Lazarov explained, allowing them to allocate saved resources to higher visual fidelity, smoother performance, longer battery life, or a blend of all three. Arm backs this vision with an open‑source software stack and a collection of developer tools.
Over the past two years the company has worked closely with the ecosystem to make AI‑enhanced graphics approachable. "Our aim was to create a developer‑friendly AI workflow that unlocks massive efficiency gains," Lazarov said. "Those gains can then be reinvested in whatever the developer values most—be it sharper images, higher frame rates, or longer battery life." Key components of Arm’s offering include Neural Super Sampling (NSS), Neural Super Sampling and Denoising (NSSD) and Neural Frame‑Rate Upscaling (NFRU).
In practice, NSS can render only a quarter of a frame on the GPU and let the neural accelerators reconstruct the remaining three‑quarters. In an even more extreme scenario, the entire frame can be generated by the AI engine, eliminating traditional rasterisation altogether. This AI‑driven reconstruction is fundamentally different from Nvidia’s DLSS 5, which relies on generative networks that can introduce content that diverges from the developer’s original artistic intent. "Our networks are convolutional, not generative," Lazarov clarified.
"They don’t imagine new content; they simply learn how existing content should look. We trained them on real‑world data and made the models publicly available on GitHub and Hugging Face." Because the models are open, developers can fine‑tune them with data from their own games. If a studio has a distinctive art style, they can capture representative frames, retrain the network, and ensure the upscaled output matches their visual language exactly—preserving intellectual property and creative control. In contrast, DLSS is a closed‑source black box: developers cannot inspect the training data, nor can they retrain the model for a specific aesthetic.
They must accept the vendor‑provided results or abandon the technology. Arm’s mobile‑first philosophy also means the neural networks are engineered to respect the strict power, bandwidth and memory limits of smartphones. "DLSS would simply blow past a phone’s budget even if it were technically supported," Lazarov said. "Our solution is built from the ground up to stay within those constraints." Performance metrics underline the efficiency gains.
Arm reports up to four‑times the performance and efficiency of the previous generation, alongside a 70 % reduction in DRAM traffic—a major source of power draw on mobile chips. In a real‑world test, the older GPU could only push Neural Dawn to about 40 fps at its limit and dropped to roughly 15 fps under normal conditions. The G2‑Ultra NX, however, runs the same demo smoothly at 60 fps. On the tooling side, Arm currently provides Unreal Engine plug‑ins and plans to broaden support to Unity and other engines.
A recent partnership with Unity China signals that Unity integration is on the roadmap, but for now the focus remains on Unreal. For studios using custom engines, Arm offers an SDK that streamlines integration without needing a dedicated plug‑in.
Lazarov highlighted that existing games can be upgraded with Arm’s AI graphics in as little as three months, enabling developers to refresh shipped titles quickly. The Neural Dawn demo itself was built by a team of fewer than 20 people in under 18 months, showcasing how Arm’s technology can accelerate development cycles while delivering cutting‑edge visual quality.
In summary, Arm’s Mali G2‑Ultra NX and associated AI‑native graphics suite aim to give mobile developers a transparent, open, and power‑efficient path to high‑fidelity rendering—offering an alternative to Nvidia’s proprietary DLSS approach that respects both artistic intent and the tight constraints of handheld devices.