The newly benchmarked Snapdragon X2 Elite Extreme X2E-96-100 delivers up to 53% higher CPU performance than the Intel Core Ultra X9 388H and up to 83% higher scores than the AMD Ryzen AI 9 465 in Geekbench 7 testing conducted by Signal65, redefining mobile processing benchmarks on battery power.
Silicon Under the Hood: 18 Oryon Cores and 80 TOPS NPU
Qualcomm’s flagship Snapdragon X2 Elite Extreme X2E-96-100 runs on third-generation Qualcomm Oryon architecture. The silicon packs 18 total cores, splitting into 12 prime cores and six performance cores capable of hitting a 5.0GHz clock speed. Graphics processing is handled by the Adreno X2-90 GPU, while the Hexagon NPU delivers up to 80 TOPS for machine learning workloads. The test configuration evaluated by Signal65 paired this silicon with 48GB of LPDDR5X RAM.
According to official specifications from Qualcomm, the X2E-96-100 features 53MB of total cache. It supports LPDDR5X memory running at 9,523 MT/s, yielding a maximum memory bandwidth of 228GB/s on 48GB configurations. By comparison, the benchmarked Intel Core Ultra X9 388H features 16 CPU cores alongside an Intel NPU 5 rated at 50 TOPS. The AMD Ryzen AI 9 465 utilizes 10 cores paired with an XDNA 2 NPU delivering 50 TOPS.
Crushing Multi-Threaded and AI Workloads
Raw processing dominance extends well beyond single-core tasks. In Geekbench 7—a CPU benchmark suite utilizing a newly structured multi-core scoring methodology—the Snapdragon X2 Elite Extreme outpaced the Intel Core Ultra X9 388H by up to 53% and the AMD Ryzen AI 9 465 by up to 83%. When shifting to multi-threaded rendering via Cinebench 2026, the performance gap widened further, with the Qualcomm silicon outperforming both x86 competitors by up to 87%.
AI computer vision workloads show an even starker divergence. Signal65’s Procyon AI Computer Vision benchmarks demonstrate that the Snapdragon X2 Elite Extreme delivers nearly double the performance of the Intel Core Ultra X9 388H and approximately 2.4 times the performance of the AMD Ryzen AI 9 465. This disparity highlights the rising importance of dedicated neural processing units in modern Windows laptops, offloading heavy mathematical inference from the primary CPU cores.
Sustained Battery Performance Defies x86 Conventions
The most consequential operational finding from the Signal65 analysis centers on power states. Testers ran the Snapdragon X2 Elite Extreme system on battery power using its Balanced mode, while pitting it against the Intel system plugged directly into AC power running its Best Performance mode. Even under this battery-constrained operational profile, the Snapdragon system maintained enough efficiency to outscore the plugged-in Intel platform across all evaluated CPU and AI test parameters.
This thermal and power efficiency targets a core limitation of modern x86 mobile architectures, which historically require maximum wall power to hit peak benchmark thresholds. Qualcomm engineered the Snapdragon X2 Elite line to sustain high instruction throughput without steep performance degradation when untethered from a charger. However, Signal65 notes that real-world results will naturally vary based on specific chassis designs, active thermal cooling solutions, memory configurations, and application optimization.
Price-to-Performance Dynamics in the Windows Ecosystem
Economic value is shifting alongside raw throughput. Signal65 evaluated a Snapdragon X2 Elite Extreme system retailing at an estimated starting price of $1,699, comparing it against an Intel setup priced around $2,299. Based on these figures, the Snapdragon configuration delivers roughly double the performance-per-dollar ratio of the Intel counterpart. For dedicated AI processing tasks, that efficiency metric scales to approximately 2.6 times the performance-per-dollar.
These benchmark figures intensify competitive pressure across the Windows PC landscape. As Qualcomm pushes hardware architectures that prioritize power efficiency and integrated neural acceleration, traditional x86 manufacturers face mounting demands to rethink thermal constraints and IPC (instructions per cycle) scaling. As software ecosystems continue adapting to ARM-based Windows environments, hardware choices for enterprise buyers and power users are expanding beyond legacy x86 silicon.
