Qualcomm and MediaTek are trimming graphics processing power from their flagship chips while preserving artificial intelligence and processor performance, a strategy that allows them to offer cheaper versions of premium phones without redesigning the entire silicon. Both chipmakers are using a technique called "binning," where they disable or remove GPU cores from existing chip designs rather than developing separate silicon for each price tier. This approach keeps CPU and NPU (neural processing unit) capabilities intact, preserving the flagship-tier marketing figures for AI performance while reducing manufacturing costs.

Why Are Chipmakers Cutting GPU Cores Instead of Other Components?

The GPU is becoming the expendable component in flagship cost-cutting. Qualcomm recently released the Snapdragon 8 Elite Gen 5 V Series, a binned variant that trims the Adreno GPU while leaving the CPU and NPU untouched. MediaTek is following a similar playbook with its Dimensity 9500s chip, which ships with two different GPU configurations depending on the device. The Redmi Turbo 5 Max uses the full 12-core Mali Immortalis-G925 MC12 GPU, while the Poco X8 Pro Max reportedly uses a cut-down 11-core MC11 version, despite both phones carrying the same 9500s branding.

This selective approach reflects a deliberate choice by chipmakers and phone manufacturers. By keeping CPU, NPU, and modem capabilities at full strength, they can advertise identical processor speeds and AI capabilities across price tiers. Only the graphics performance takes a hit, a trade-off that matters less to most smartphone users than raw computing power or AI features.

What Does This Mean for On-Device AI Performance?

The preservation of NPU (neural processing unit) cores is the key signal here. As on-device AI becomes increasingly central to smartphone marketing, chipmakers are protecting AI performance at all costs. The NPU is the specialized processor that handles machine learning tasks locally on your phone, without sending data to cloud servers. By keeping these cores intact while cutting GPU cores, Qualcomm and MediaTek ensure that AI features like real-time translation, voice recognition, and image processing remain consistent across their product lineup.

This strategy reflects the industry's recognition that AI performance has become a flagship differentiator. Consumers and reviewers now expect flagship phones to deliver powerful on-device AI capabilities. Cutting GPU cores allows manufacturers to hit lower price points without compromising the AI features that justify the "flagship" label.

How to Understand GPU Binning in Smartphone Chips

• What It Is: GPU binning is the practice of disabling or removing graphics processor cores from a chip design to create lower-cost variants without redesigning the entire silicon from scratch.
• Why It Matters: It allows chipmakers to offer multiple price tiers of the same flagship chip, reducing development costs while maintaining identical CPU and AI performance across all variants.
• The Trade-Off: Graphics performance decreases, but for most smartphone users who prioritize AI features and processing speed over gaming, the impact is minimal.
• Industry Trend: With component costs continuing to climb through 2026, GPU binning is expected to become standard practice for flagship SKUs (stock keeping units) rather than an exception.

The iQOO Neo 11 Extreme Edition provides a concrete example of this trend. A Geekbench listing for the device showed an 11-core Arm Mali-G1-Ultra MC11 GPU instead of the 12-core MC12 found in the standard Dimensity 9500. The device scored 3,445 in single-core and 10,517 in multi-core testing on Geekbench 6.7.1 while running Android 16, with the CPU configuration otherwise matching the standard chip.

Why Are Rising Component Costs Driving This Strategy?

DRAM (dynamic random-access memory) and component costs have been climbing steadily through 2026, squeezing profit margins on mid-range and budget flagship phones. Rather than absorb these costs or reduce features across the board, chipmakers are using GPU binning as a surgical cost-cutting tool. It allows them to preserve the components that matter most for marketing and user experience while trimming the least visible performance metric.

This approach is more efficient than developing entirely separate chip designs for different price tiers. Creating a new silicon variant requires significant engineering investment, testing, and manufacturing setup costs. Binning an existing design is far cheaper; it simply involves disabling certain cores during the manufacturing process or in firmware. As costs continue to rise, expect this strategy to become even more widespread across the industry.

The broader implication is clear: on-device AI has become the performance metric that matters most to consumers and manufacturers alike. GPU power is negotiable; AI capability is not. This shift reflects the growing importance of machine learning features in smartphones and the competitive pressure to deliver flagship-level AI performance across all price tiers.