Arm Details Its Mobile‑First AI Reconstruction Approach, Diverging from the "Black Box" Model of DLSS 5

Arm, the British powerhouse in semiconductor and software design, recently staged its Arm Everywhere China showcase in Shanghai, unveiling a suite of fresh AI‑computing solutions aimed at reshaping the mobile graphics landscape. Central to the announcement is a comprehensive AI‑native ecosystem that includes the second generation of its AI‑centric compute platform, CSS for Mobile 2, and a brand‑new graphics processor, the Arm Mali G2‑Ultra NX GPU. This GPU is built around neural‑compute blocks, promising a noticeable uplift in visual fidelity while keeping heat output and power draw dramatically lower than conventional solutions.

While Arm did not disclose the exact devices that will ship with the Mali G2‑Ultra NX, its predecessor, the Mali G1‑Ultra, has already found its way into smartphones from manufacturers such as Xiaomi, Vivo, and OnePlus. 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 tools to harness compute‑heavy techniques—most notably real‑time ray tracing—without sacrificing the battery life or thermal envelope that mobile devices demand. "Ray tracing has been gaining traction on phones," Lazarov noted.

"It was a slow burn at first, but now an increasing number of titles are asking for it, and we wanted to make sure the hardware can keep up." This sentiment aligns with recent reports from GamesIndustry.biz, which highlighted that the new GPU will let developers tap into Unreal Engine features such as MegaLights. Arm even partnered with Sumo Digital to produce a demo title, *Neural Dawn*, that showcases the G2‑Ultra NX’s capabilities. Lazarov described the target audience for the chip as "mobile‑first games that still want to look like desktop experiences," as well as cross‑platform titles that aim to deliver a consistent visual quality whether the player is on a console at home or a train in London. He cited upcoming games that will leverage the new hardware, including Tencent’s *Arena Breakout Infinite*, NetEase’s *Where Winds Meet*, and Infold Games’ *Infinity Nikki*.

One of the biggest challenges for mobile graphics, Lazarov explained, is the stark difference in thermal and power budgets compared with consoles or PCs. "A phone can’t dump heat the way a console can, and it’s always running off a battery," he said.

To address this, Arm has fused neural processing directly into the traditional rendering pipeline. By offloading portions of the workload to dedicated neuro‑accelerators, the GPU can render fewer pixels and let the AI reconstruct the missing data, dramatically lowering both compute demand and power consumption. "That translates into new efficiency that eases the constraints developers normally face," Lazarov added.

"It means they don’t have to make as many trade‑offs between visual quality, performance, and battery life." To further lower the barrier to adoption, Arm is pursuing an open‑software strategy. Over the past two years the company has been collaborating with the broader ecosystem to create developer‑friendly tools that expose the AI‑driven gains without locking anyone out.

The core of Arm’s offering consists of three technologies: Neural Super Sampling (NSS), Neural Super Sampling and Denoising (NSSD), and Neural Frame Rate Upscaling (NFRU). Together they enable image upscaling, additional ray generation, and the synthesis of new frames through AI‑based reconstruction.

Lazarov illustrated the process with a two‑frame example. In the first frame, the GPU renders only a quarter of the final image; the remaining three‑quarters are filled in by the neuro‑accelerators using NSS.

In the second frame, the GPU does virtually no traditional rendering at all—every pixel is produced by the AI system, delivering a fully reconstructed frame. Nvidia’s competing solution, DLSS 5, also relies on AI to add detail, but it has attracted criticism for its "black‑box" nature and for sometimes diverging from a developer’s artistic intent.

Arm’s approach, by contrast, is deliberately transparent. "Our networks are not generative; they’re convolutional," Lazarov explained. "They don’t imagine new content—they simply know how existing content should look based on real‑world data.

All of our models are publicly available on GitHub and Hugging Face, so developers can inspect, modify, and even retrain them on their own art style. This openness means a studio with a distinctive visual direction can capture data from its own game, fine‑tune the network, and ensure the upscaled frames retain the exact aesthetic they designed. There is no risk of losing intellectual‑property control, a concern that often accompanies proprietary solutions. Lazarov contrasted this with Nvidia’s closed‑source model, describing DLSS as a "take‑it‑or‑leave‑it" proposition.

Arm’s solution, he said, is "take it and make it work for you." Moreover, because Arm designed its networks with mobile constraints in mind—power, bandwidth, and memory limits—the technology can actually run on smartphones, whereas a DLSS‑style system would instantly exceed a phone’s budget. Power efficiency is a cornerstone of the Mali G2‑Ultra NX. Arm reports up to four‑fold gains in both performance and efficiency, alongside a 70 % reduction in DRAM traffic, which is a major source of power draw on mobile devices.

Less external memory access translates directly into longer battery life and cooler operation. In practical terms, the difference is stark. When pushing the previous‑generation GPU to its limits, *Neural Dawn* could only achieve around 40 fps under ideal conditions and fell to 15 fps in typical gameplay.

The G2‑Ultra NX, however, comfortably sustains 60 fps, delivering a smoother experience without sacrificing visual fidelity. From a developer tooling perspective, Arm currently offers Unreal Engine plug‑ins and has a roadmap that includes Unity support, a partnership with Unity China already announced. For studios that rely on custom engines, Arm provides an SDK that streamlines integration, allowing teams to embed AI‑native graphics without waiting for a dedicated plug‑in. Lazarov emphasized that existing titles can also benefit from the new tech.

"If you have a game already shipping, you can integrate our solutions and ship an updated version in as little as three months," he said. This rapid turnaround was demonstrated by *Neural Dawn*, which was built by a team of fewer than 20 people in under 18 months, leveraging Arm’s AI stack to accelerate development and achieve high‑quality graphics quickly. In summary, Arm’s mobile‑first AI reconstruction strategy offers an open, efficient, and developer‑centric alternative to Nvidia’s DLSS 5.

By embedding neural accelerators into the rendering pipeline, providing publicly available convolutional networks, and delivering tools that respect the tight power and thermal budgets of smartphones, Arm aims to empower developers to create games that look great on mobile without the compromises that have traditionally plagued the platform.