Arm, the British chip and software design powerhouse, recently staged its Arm Everywhere China showcase in Shanghai, unveiling a suite of fresh AI‑computing innovations aimed at reshaping mobile graphics. Central to the announcement is the AI‑native compute platform CSS for Mobile 2 and the brand‑new Mali G2‑Ultra NX GPU, a processor engineered around neural‑accelerated rendering.
Arm says this on‑device AI solution can lift visual fidelity while keeping power draw and heat generation far lower than traditional approaches, positioning it as a more transparent alternative to Nvidia’s controversial DLSS 5 technology. While Arm did not disclose exact product partners, the predecessor Mali G1‑Ultra has already found its way into smartphones from Xiaomi, Vivo and OnePlus, suggesting the G2‑Ultra NX could appear in a similar range of devices soon. Deyan Lazarov, a senior member of 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 chip’s primary goal is to give mobile developers access to heavyweight graphics features—most notably ray tracing—without the prohibitive energy costs that have traditionally limited such effects to desktop‑class hardware.
"We’re seeing a growing appetite for ray tracing on phones," Lazarov noted. "It moved slowly at first, but more titles are now experimenting with it, and our hardware is built to support that momentum." Recent collaborations, such as the partnership with Sumo Digital that produced the demo game Neural Dawn, illustrate how the new GPU can harness Unreal Engine tools like MegaLights to deliver console‑grade visuals on a handheld platform. Lazarov described the target audience as developers who adopt a "mobile‑first" philosophy but still aim for desktop‑level aesthetics.
He also highlighted cross‑platform titles that need to deliver a consistent experience whether played on a console at home or on a commuter train. Games slated to showcase the Mali G2‑Ultra NX include Tencent’s Arena Breakout Infinite, NetEase’s Where Winds Meet, and Infold Games’ Infinity Nikki.
One of the core challenges of mobile graphics is the strict thermal and power envelope. Unlike consoles or PCs, phones cannot dissipate large amounts of heat and must constantly run off battery power. To address this, Arm has fused neural processing units (NPUs) directly into the rendering pipeline.
By offloading portions of the workload to these NPUs, the GPU can operate at reduced power while still delivering high‑quality frames. "When the GPU only has to render a fraction of the scene and the NPU fills in the rest, you get a new level of efficiency," Lazarov explained.
"Developers can spend less time making trade‑offs between performance, visual fidelity and battery life, and more time focusing on the experience they want to create." The suite of AI‑driven techniques Arm is releasing includes Neural Super Sampling (NSS), Neural Super Sampling and Denoising (NSSD), and Neural Frame Rate Upscaling (NFRU). Together they allow the system to upscale lower‑resolution images, generate additional rays for better lighting, and synthesize intermediate frames, all through AI reconstruction rather than brute‑force rasterization.
In a practical example, Lazarov described two consecutive frames: in the first, the GPU renders only 25 % of the image while the NPU reconstructs the remaining three quarters via NSS; in the second, the GPU does no traditional rendering at all, and the entire frame is produced by the neural accelerator. This approach dramatically cuts GPU usage, leading to lower power draw and less traffic to DRAM—Arm reports a 70 % reduction in memory bandwidth, a major source of energy consumption on mobile devices. Arm’s open‑source stance sets it apart from Nvidia’s DLSS 5, which has drawn criticism for its black‑box nature.
"Our networks are convolutional, not generative, so they don’t hallucinate new content," Lazarov emphasized. "We train them on real‑world data and publish the models on GitHub and Hugging Face, allowing developers to fine‑tune them with their own art styles. This preserves IP control and lets studios customize the upscaling to match their unique visual language." By contrast, DLSS is a proprietary system that cannot be retrained or inspected, forcing developers into a take‑it‑or‑leave‑it scenario. Moreover, DLSS was designed with high‑end PCs in mind and would quickly exceed the power and bandwidth budgets of a smartphone, according to Lazarov.
Arm’s models, however, are explicitly built for mobile constraints, balancing performance, bandwidth and energy consumption. Performance figures underscore the advantage: when pushed to its limits, the previous‑generation Mali GPU could only run Neural Dawn at around 40 fps under ideal conditions and dropped to 15 fps in typical use.
The Mali G2‑Ultra NX, on the other hand, comfortably sustains 60 fps, delivering a smoother, more responsive experience. To help developers adopt the technology, Arm has released Unreal Engine plug‑ins and is planning support for Unity and other engines. An SDK provides a streamlined integration path for custom engines as well. Lazarov noted that retrofitting an existing shipped title with Arm’s AI graphics can take as little as three months, enabling rapid upgrades to visual quality or efficiency.
The rapid development cycle of Neural Dawn—built by fewer than 20 people in under 18 months—demonstrates how Arm’s tools can accelerate game production. "Our tech lets studios quickly enhance existing games or create groundbreaking new experiences without massive overhead," Lazarov concluded. "We’re putting the power of AI‑native graphics in the hands of developers, wherever they are, and wherever their players are playing."