MINIX Unveils T4000 and T5000 AI Workstations
The compact systems, built on NVIDIA's Jetson AGX Thor platform, are listed at $6,460 and $9,785.
Updated on Nov. 1, 2026 in Computers
In April 2026, MINIX announced the T4000 and T5000, compact AI workstations based on NVIDIA's Jetson AGX Thor platform, combining Blackwell GPU architecture with Arm Neoverse-V3AE processors. The product is listed as Coming Soon in an article dated November 1, 2026.
Why it matters
The workstations bring Blackwell-class GPU hardware and up to 128GB of LPDDR5X memory into a compact form factor aimed at local AI development. Their listed pricing places them in a range distinct from full-size professional workstations.
The T4000 pairs a 12-core Arm Neoverse-V3AE CPU up to roughly 2.5GHz with 1,536 CUDA cores and 273GB/s memory bandwidth, while the T5000 uses a 14-core CPU up to 2.6GHz and 2,560 CUDA cores at the same bandwidth. Independent verification of the manufacturer-reported inference throughput figures is not yet available.
The players
MINIX
MINIX is a computer hardware company that produces compact mini PCs and workstations.
NVIDIA
NVIDIA is a semiconductor company known for its GPU architectures and its Jetson line of embedded AI computing platforms.
The details
Both systems use NVIDIA's Multi-Instance GPU technology, which lets the GPU be divided into separate instances, with the T4000 supporting six partitions and the T5000 supporting ten. Each includes two 10GbE RJ45 ports, Wi-Fi 6, Bluetooth 5.3, a 1TB PCIe 4.0 NVMe SSD with support up to 4TB, and ships with Ubuntu 24.04.
Timeline
April 2026: MINIX originally announced the workstation.
November 1, 2026: Article published listing the product as Coming Soon.
The Tech Race
The T4000 and T5000 extend NVIDIA's Jetson AGX Thor platform into a packaged desktop workstation, moving Blackwell-generation AI hardware out of embedded and robotics contexts. Their positioning places them against other compact AI workstation offerings competing for local large-model inference.
The T4000 starts at $6,460 and the T5000 at $9,785, making local AI model inference a purchasing decision for developers and small teams. Buyers should weigh the 64GB versus 128GB memory difference against their model sizes before choosing a configuration.
The takeaway
Before buying a compact AI workstation, check whether your target models fit within the available memory, since the 64GB and 128GB configurations differ substantially in capacity. Verify manufacturer-reported inference speeds against independent benchmarks before relying on them.
Further reading
Explore more hardware coverage in the Computers section.
