Arm may be developing a new GPU in bid to challenge Nvidia

Rumor mill: It seems that everyone wants to challenge Nvidia’s grip on the AI chip market, and Arm is no exception. The UK-based semiconductor design company has become skilled in using its business strategies and industry partnerships to work toward this goal. Now, it looks like Arm is taking a more direct approach by potentially developing a GPU that could compete with Nvidia’s products. Although the report from unnamed sources leaves many details unclear, if true, it aligns with Arm’s commitment to becoming a significant player in the AI field.

Sources report that Arm is working on a new GPU at its development center in Ra’anana, Israel, where a team of up to 100 engineers has been assembled for the project. Although the exact focus of their work remains somewhat ambiguous, speculation suggests that the team is developing a graphics architecture that could be used for video rendering or AI training and inferencing.

However, many aspects of the endeavor remain uncertain. It is unclear whether the company will produce a discrete GPU that other companies could license or an enhanced integrated graphics solution for its desktop and laptop SoCs (system on a chip). Arm has not commented on these speculations.

With the success of Qualcomm’s Snapdragon X and the anticipated arrival of desktop form factors, there is a potential market for a discrete GPU specifically designed for the Arm architecture. Qualcomm’s Snapdragon X series, particularly the Snapdragon X Elite, represented a significant step forward in Arm-based PC chips, competing against Apple’s M2 Max and Intel’s Core i9-13980HX, while consuming less power.

Arm’s current GPU offerings include the Mali and Immortalis series, which are integrated into numerous SoCs for smartphones and mobile devices. Much of this development takes place at the Ra’anana center, where hardware teams are responsible for creating components for the GPUs, while the software teams work on developing interfaces like Vulkan and OpenGL to optimize these GPUs for gaming and other graphics tasks.

What appears to be new is that Arm has also begun collaborating with Israeli startups such as Hailo, which produces AI chips for security cameras, and NeuReality, which develops hardware to replace Nvidia servers, according to the sources. This collaboration suggests a potential focus on creating cost-effective alternatives to Nvidia’s AI solutions.

For these reasons, PC Gamer believes that Arm’s new GPU is unlikely to target the discrete gaming graphics card market. Instead, it seems more plausible that Arm is setting its sights on the lucrative AI hardware market, currently dominated by Nvidia’s high-end AI chips.

Arm is actively working to compete with Nvidia and is part of a coalition with other major tech companies like Intel, Google, and Qualcomm that aim to challenge Team Green’s dominance in the AI market. This coalition is working on an open-source software suite to dismantle the proprietary software barriers that have kept developers locked into Nvidia’s ecosystem.

Arm’s involvement in initiatives like the Unified Acceleration Foundation underscores a broader strategy to strengthen its position in the AI market. By supporting open standards and reducing dependency on Nvidia’s CUDA platform, Arm is positioning itself as a more viable competitor in the AI chip space.