Arm Unveils C2 CPU and Mali G2-Ultra NX GPU to Bring AI-Native Ray Tracing to Mobile
Arm's new Compute Subsystem for Mobile 2 integrates a neural engine directly into the GPU, promising a 70 percent reduction in ray tracing overhead for next-generation Android devices.
By Jackson Reed
- Silicon Designers
- Focuses on the architectural breakthrough of moving AI processing directly onto the GPU to eliminate latency.
- Mobile Game Developers
- Values the 70 percent reduction in ray tracing cost as a way to port PC titles without melting smartphone batteries.
- Consumer Hardware Analysts
- Views the announcement as the performance baseline for the 2027 flagship Android release cycle.
Perspectives this story doesn't cover
- Apple Silicon Engineers
- Mobile Esports Competitors
Why this matters
By shifting graphics workloads from brute-force rendering to AI-assisted generation on the GPU itself, this architecture allows mobile developers to port complex PC and console titles to phones without draining the battery in minutes.
On Tuesday, September 8, 2026, Arm Holdings detailed a silicon architecture that cuts the computational cost of mobile ray tracing by 70 percent. The company's new Compute Subsystem (CSS) for Mobile 2 introduces the C2 CPU and the Mali G2-Ultra NX GPU, marking the first time the British chip designer has integrated a dedicated neural engine directly into its graphics hardware.[1][3]
The integration fundamentally changes how Android devices will render complex lighting. Instead of relying on a separate Neural Processing Unit (NPU) across the motherboard, the Mali G2-Ultra NX handles AI upscaling and frame generation internally.[5][6]
Technology outlet 9to5Google reported that Arm's new CPU cores and GPUs for Android chips promise 'desktop-class' games, noting that the localized neural hardware eliminates the latency penalty previously incurred when the GPU and NPU traded data back and forth.[4]
The 70 percent reduction in ray tracing cost specifically targets the thermal limits of modern smartphones. Ray tracing—the physical simulation of light bouncing off virtual surfaces—traditionally forces mobile processors to throttle clock speeds within minutes to prevent overheating.[3]
The 70 percent reduction in ray tracing cost specifically targets the thermal limits of modern smartphones.
By offloading the heaviest calculations to the onboard neural engine, the Mali G2-Ultra NX allows developers to maintain sustained peak performance. Gadgets Now highlighted that building the neural engine into the newest Mali phone GPU directly addresses the battery drain that has historically kept AAA console ports from running smoothly on mobile.[7]
Alongside the graphics overhaul, the CSS for Mobile 2 platform introduces the C2 CPU, which Arm explicitly designed for what the industry terms 'agentic AI.'[1][2]
Agentic AI refers to on-device models that can take autonomous actions across multiple apps without waiting for cloud server validation. The C2 core increases integer and floating-point throughput to process these background language models while the user is actively playing a game at 60 or 120 frames per second.[2][5]
The hardware announcement signals a shift in the mobile silicon market, where Arm provides the foundational blueprints for manufacturers like MediaTek and Samsung. Unite.AI noted that the Mali G2-Ultra NX represents Arm's 'first AI-native mobile GPU,' setting a new baseline for the 2027 flagship Android release cycle.[6]
Silicon vendors typically require 6 to 9 months to integrate new Arm architectures into their own system-on-chip designs. Consequently, the first consumer handsets featuring the C2 and Mali G2-Ultra NX are expected to hit retail shelves in the second half of 2027, dictating the performance ceiling for the next generation of mobile esports.[4][5]
Viewpoints in depth
Silicon Designers
Focuses on the architectural breakthrough of moving AI processing directly onto the GPU to eliminate latency.
For years, mobile chipmakers have relied on a separate Neural Processing Unit (NPU) to handle AI tasks, forcing the GPU to send data across the system bus and wait for a response. Arm's engineers argue that this physical distance creates a latency bottleneck that ruins real-time graphics generation. By embedding the neural engine directly inside the Mali G2-Ultra NX, the architecture eliminates the transit time, allowing the GPU to upscale frames and calculate lighting bounces instantaneously.
Mobile Game Developers
Values the 70 percent reduction in ray tracing cost as a way to port PC titles without melting smartphone batteries.
Studios attempting to bring AAA console games to mobile platforms have consistently hit a thermal wall. Ray tracing requires so much brute-force calculation that smartphones typically overheat and throttle their performance within ten minutes of gameplay. Developers view the 70 percent efficiency gain not just as a visual upgrade, but as a structural fix that will finally allow sustained, high-fidelity gaming sessions on battery power.
Consumer Hardware Analysts
Views the announcement as the performance baseline for the 2027 flagship Android release cycle.
Hardware watchers note that Arm does not build the final chips, but rather licenses these blueprints to manufacturers like MediaTek and Samsung. Because those vendors require six to nine months to integrate the C2 and Mali G2-Ultra NX into their own silicon, analysts are already using this announcement to map out the exact performance capabilities of the premium Android smartphones that will hit the market in late 2027.
Key points
- Arm's CSS for Mobile 2 introduces the C2 CPU and Mali G2-Ultra NX GPU.
- The architecture integrates a neural engine directly into the GPU for the first time.
- The localized AI hardware cuts the computational cost of mobile ray tracing by 70 percent.
- The C2 CPU is specifically optimized to run background 'agentic AI' tasks during active gameplay.
- First consumer devices featuring the chips are expected in the second half of 2027.
Sources
[1]Arm NewsroomSilicon DesignersArm introduces new AI-native compute platform built for agentic AI and mobile graphics
Read on Arm Newsroom →
[2]PocketGamer.bizMobile Game DevelopersARM unveils AI-native mobile platform for agentic AI and graphics
Read on PocketGamer.biz →
[3]shattered.ioMobile Game DevelopersArm C2-Ultra, Mali G2-Ultra NX Cut Ray Tracing Cost 70%
Read on shattered.io →
[4]9to5GoogleConsumer Hardware AnalystsArm's new CPU cores and GPUs for Android chips promise 'desktop-class' games, more AI power
Read on 9to5Google →
[5]FoneArena.comConsumer Hardware AnalystsArm CSS for Mobile 2 with C2 CPU, Mali G2-Ultra NX GPU for agentic AI and AI-native graphics announced
Read on FoneArena.com →
[6]Unite.AIConsumer Hardware AnalystsArm Unveils Mali G2-Ultra NX, Its First AI-Native Mobile GPU
Read on Unite.AI →
[7]Gadgets NowConsumer Hardware AnalystsArm Builds A Neural Engine Into Its Newest Mali Phone GPU
Read on Gadgets Now →
Comments
More in Gaming & Esports
See all →Studio Restructuring
Star Stable Entertainment Cuts 20% of Workforce in Restructuring Move
5 sources
Netcode Architecture
Why Real-Time Multiplayer Games Discard TCP for Custom UDP Netcode
7 sources
Steam Frame
Valve Prepares Randomized Reservation System for Steam Frame VR Headset Launch
2 sources
SteamOS Expansion
Valve Opens SteamOS to Rival Handhelds, Turning OS Into Multi-Vendor Platform
6 sources
Every angle. Every day.
Get Gaming & Esports stories with full source coverage and perspective breakdowns delivered to your inbox.




