Dell has started shipping the Pro Max 16 Plus configured with a discrete Qualcomm Artificial Intelligence-100 Ultra neural processing unit. The NPU is rated at 870 INT8 TOPS and operates within a 150 W TDP envelope. Dell pairs the NPU with up to 128 GB of LPDDR5X memory on select models, which the company says provides enough compute and memory capacity to run local Artificial Intelligence models with up to 120 billion parameters on the NPU without requiring additional external memory or compute resources.
The Pro Max 16 Plus combines the Qualcomm NPU with robust CPU options, including up to an Intel Core Ultra 9 285HX vPro with 24 cores. To accommodate the NPU, Dell has removed the traditional discrete GPU from the system. Graphics output is handled by the integrated GPU within the Arrow Lake-HX chip, which Dell positions as suitable for basic display tasks but not for gaming. Dell describes the tradeoff as intentional: the dedicated NPU is intended for Artificial Intelligence developers and consumes the space, power, and cooling headroom that a discrete GPU would normally occupy.
For local development workflows the laptop uses the Qualcomm NPU for model inference and related workloads, paired with the host CPU and configurations commonly shipping with 64 GB of system memory for broader system tasks. Dell emphasizes the NPUu2019s combination of high TOPS and substantial on-board memory as an alternative to a conventional GPU-based design, arguing it offers more computing power and VRAM capacity for on-device Artificial Intelligence workloads. The company positions the Pro Max 16 Plus as a workstation choice focused on AI-centric applications rather than gaming or GPU-accelerated graphics.
