Acer unveiled its Veriton GN100 AI mini workstation ahead of IFA 2025, joining a crowded field of PC manufacturers racing to capitalize on NVIDIA’s desktop AI chip strategy. The device starts at $3,999 in North America and €3,999 in Europe, positioning it alongside similar offerings from ASUS, Dell, and HP.
The compact system delivers up to 1 PFLOPS of AI performance through NVIDIA’s GB10 Grace Blackwell Superchip, combined with 128GB of unified memory and 4TB storage. Two units can be linked to handle AI models with up to 405 billion parameters, though this doubles the already steep entry cost to nearly $8,000.
NVIDIA launched Project DIGITS with the GB10 chip in January, promising to bring datacenter-level AI capabilities to desktop environments. The company has since partnered with multiple manufacturers to build systems around the architecture, creating a new category of high-priced AI workstations.
The hefty pricing reflects the specialized nature of these systems, which target developers, researchers, and enterprises willing to pay premium rates for local AI processing. However, at roughly $4,000, the GB10’s AI performance is comparable to NVIDIA’s RTX 5070 graphics card, which retails for around $550, raising questions about the value proposition for cost-conscious buyers.
The success of these systems will largely depend on whether organizations find sufficient benefit in local AI processing to justify the significant expense over cloud-based alternatives.
