Dell has officially launched its Pro Max Mini AI PC, an ultra-compact powerhouse built around NVIDIA’s cutting-edge GB10 Superchip – and it’s already being hailed as a game-changer for local AI computing. First teased during NVIDIA’s GTC 2025 event, this tiny machine is now making its retail debut with specifications that make even traditional AI workstations look outdated.
At the heart of the Dell Pro Max GB10 lies the NVIDIA Grace Blackwell GB10 Superchip, a marvel of efficiency and brute strength. 
It packs 20 Arm v9.2 cores paired with a Blackwell-based GPU that features 6144 cores – the same count found on an RTX 5070. What truly sets this apart is its ability to deliver 1000 TOPS of FP4 compute and manage AI models up to 200 billion parameters directly on-device. That means developers can fine-tune large language models like Llama 3.3 70B locally, without depending on cloud infrastructure or paying for remote GPU time.
Dell’s integration of 128 GB of LPDDR5x memory clocked at a blistering 9400 MT/s ensures lightning-fast data access, yielding a system bandwidth of up to 273 GB/s
. Storage options are equally impressive: configurations include up to 2 TB PCIe Gen4 and 4 TB of SED Ready M.2 drives. On the connectivity side, this mini beast comes well-equipped – four USB Type-C ports (three 20 Gbps and one with power-in), a 10 GbE LAN port, HDMI 2.1b, and even two QSFP 200 Gbps ports for high-speed networking. 
It’s a dream setup for developers running distributed AI workloads or compact server nodes.
Despite its enormous power, the Pro Max Mini maintains a remarkably small footprint – just 150 x 150 x 51 mm and 1.31 kg in weight. Its sleek design allows it to blend seamlessly into any workspace while drawing energy from a modest 280W AC adapter. Wireless connectivity is equally forward-looking, featuring Wi-Fi 7 and Bluetooth 5.4 for ultra-fast device pairing and data transfer. This isn’t merely a desktop – it’s an intelligent workstation that fits in your hand.
What makes the Pro Max GB10 especially innovative is its ability to stack with another unit using NVIDIA’s ConnectX-7 Smart NIC. Doing so effectively doubles the system’s capacity, enabling users to train or infer models as large as 400 billion parameters. It’s an elegant scaling solution for developers or small teams seeking to expand compute capacity without building a full data center rack.
Running NVIDIA DGX OS and the complete NVIDIA AI software stack, the Dell Pro Max GB10 ensures compatibility across environments – from local desktops to enterprise clusters. Developers can move workloads seamlessly, maintaining consistent results without reconfiguration headaches. This consistent software environment is key for professionals working across hybrid setups where portability and reproducibility matter.
Beyond the hardware, Dell’s message is clear: local AI computing is the future. By cutting dependence on cloud GPUs, the Pro Max Mini offers not only cost efficiency but also enhanced security and low-latency responsiveness – crucial for AI-driven applications like real-time speech generation, robotics, or embedded analytics. Its architecture embodies a growing shift toward personal AI workstations designed for creativity, autonomy, and power efficiency.
The Dell Pro Max GB10 Mini AI PC will hit stores on October 15th, with pricing and bundle configurations to be revealed soon. As AI models grow larger and more sophisticated, Dell’s compact titan might just become the go-to device for developers who demand big performance in small packages.
2 comments
Wish they’d show the price already. Been waiting since GTC!
Imagine stacking two of these on your desk… 400B models at home 😎