Arm, the British chip‑design and software powerhouse, recently staged its Arm Everywhere China showcase in Shanghai, using the occasion to announce a suite of fresh AI‑computing solutions aimed at the mobile market. The centerpiece of the announcement is a new AI‑native compute platform called CSS for Mobile 2, paired with the Mali G2‑Ultra NX graphics processor – a GPU built around neural‑accelerator cores that promises higher visual fidelity while keeping heat and power draw to a minimum. Arm positions this on‑device AI‑enhanced rendering as a more transparent and developer‑friendly alternative to Nvidia’s controversial DLSS 5 technology, and as a realization of the original vision many game creators had for real‑time up‑scaling.
While Arm did not disclose the exact devices that will ship with the Mali G2‑Ultra NX, the predecessor Mali G1‑Ultra has already found its way into phones from Xiaomi, Vivo and OnePlus, suggesting a similar rollout path for the newer silicon. Deyan Lazarov, a senior member of Arm’s GPU product team within the Edge AI division, explained that the company has spent roughly two and a half years shaping the G2‑Ultra NX to give mobile developers access to compute‑heavy features such as ray tracing. “We’ve seen a steady rise in mobile titles experimenting with ray tracing,” Lazarov noted. “It was a slow start, but now more games are actively looking to enable it.” The new GPU is designed to work hand‑in‑hand with popular development tools like Unreal Engine.
As reported earlier by GamesIndustry.biz, the Mali G2‑Ultra NX can leverage Unreal’s MegaLights system, and Arm collaborated with Sumo Digital to produce a demo game called *Neural Dawn* that showcases the chip’s capabilities. Lazarov described the target audience as “mobile‑first games that still want a desktop‑class visual experience,” as well as cross‑platform titles that aim for a consistent look and feel whether the player is on a console at home or a smartphone on a train. A handful of upcoming titles have already signed up to use the new hardware, including Tencent’s *Arena Breakout Infinite*, NetEase’s *Where Winds Meet* and Infold Games’ *Infinity Nikki*. Lazarov emphasized that smartphones face unique constraints: they cannot dissipate heat like a console or PC and must operate continuously on battery power.
To address these limits, Arm’s solution blends neural processing with the conventional rendering pipeline, offloading a portion of the GPU’s workload to dedicated neuro‑accelerators. This not only cuts the GPU’s power consumption but also improves overall system efficiency, giving developers more headroom to allocate resources toward visual quality, performance or battery life. Arm is backing this vision with an open‑source software stack and a collection of developer tools.
“We’ve been collaborating with the ecosystem for the past couple of years to make AI‑driven graphics as developer‑friendly as possible,” Lazarov said. The goal is to let developers reap large efficiency gains and then decide how to reinvest the saved budget—whether that means crisper textures, higher frame rates, longer battery life, or a mix of all three.
Key components of Arm’s AI‑graphics suite include Neural Super Sampling (NSS), Neural Super Sampling and Denoising (NSSD) and Neural Frame‑Rate Upscaling (NFRU). In practice, the technology can upscale lower‑resolution images, cast additional rays for richer lighting, and generate extra frames 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 neuro‑accelerators reconstruct the remaining three‑quarters using NSS.
In the second frame, the GPU does no traditional rendering at all; the entire frame is produced by the neural engines, demonstrating a fully AI‑reconstructed output. Nvidia’s DLSS 5 also uses AI to add detail, but it has drawn criticism for its “black‑box” nature, which can sometimes diverge from a developer’s artistic intent.
Arm’s approach, by contrast, relies on convolutional networks that do not generate new content but rather learn how existing content should look. “Our networks are trained on real‑world data and are publicly available on GitHub and Hugging Face,” Lazarov explained.
“Developers can download a network, fine‑tune it with data from their own game’s art style, and then use it to upscale or insert frames that match their unique visual language. This preserves IP control and avoids the risk of an opaque system altering the look of a game without consent.” Lazarov highlighted that DLSS is a closed, proprietary system—essentially a black box that cannot be retrained by end users. Arm’s solution, however, is open and customizable, allowing developers to adapt the models to their specific aesthetic needs.
Moreover, Arm built its AI stack with mobile constraints in mind. Power, bandwidth and memory budgets are baked into the design, making the technology viable on smartphones where DLSS would quickly exceed available resources. Power efficiency is a cornerstone of the Mali G2‑Ultra NX.
Arm reports up to a four‑fold increase in performance and efficiency, alongside a 70 % reduction in DRAM traffic—a major source of power draw on mobile devices. Fewer memory accesses translate directly into longer battery life.
In a side‑by‑side test, the previous‑generation GPU could only push *Neural Dawn* to 40 fps under ideal conditions and dropped to 15 fps in typical usage, whereas the new chip comfortably sustains 60 fps. On the tooling side, Arm currently offers Unreal Engine plug‑ins and plans to expand support to other engines, with Unity integration on the roadmap.
A recent partnership with Unity China signals that Unity support is forthcoming. For studios using custom engines, Arm provides an SDK that streamlines integration without requiring a dedicated plug‑in. Lazarov also pointed out that existing games can benefit from the AI‑native graphics stack with relatively short development cycles.
“If you have a shipped title and want to integrate our solutions, you could have a new version ready in as little as three months,” he said. 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 and enable smaller teams to deliver high‑quality experiences quickly.
In summary, Arm’s mobile‑first AI reconstruction strategy offers an open, customizable alternative to Nvidia’s DLSS 5, tailored to the power‑ and thermal‑limited environment of smartphones. By merging neural up‑scaling, denoising and frame‑rate generation with traditional rendering, the Mali G2‑Ultra NX aims to give developers the flexibility to enhance visual fidelity, maintain performance and extend battery life—all while keeping the underlying AI models transparent and adaptable to each game’s unique artistic direction.