Local AI model execution is moving from specialized server racks to compact desktops with the ThinkCentre X Ultra. This 1.6-liter machine targets developers and researchers who need high compute density without the noise and footprint of traditional workstations. Buyers requiring quiet, high-performance AI inference capabilities now have a dedicated hardware option that fits on a standard desk.
Compact desktop targets local AI model execution with clustered compute options
Lenovo designed this mini PC around the AMD Ryzen AI Max+ Pro 495 processor. The chip provides 16 cores and 32 threads to handle complex workloads. It delivers 131 TOPS of total AI compute performance, with 55 TOPS coming specifically from the integrated NPU. This architecture allows the system to run large language models locally without relying on cloud infrastructure.
Specifications
- CPU: AMD Ryzen AI Max+ Pro 495 (16C/32T)
- AI Compute: 131 TOPS total (55 TOPS NPU)
- RAM: Up to 128 GB LPDDR5x 8533 MHz
- Storage: Up to two 4 TB M.2 2280 Gen5 SSDs
- Dimensions: 183 x 183 x 51 mm (1.6L)
The system supports up to 128 GB of LPDDR5x RAM running at 8533 MHz. Storage capacity reaches 8 TB total through two M.2 2280 Gen5 SSDs. Connectivity includes two Thunderbolt 4 ports, two USB-C ports, and two USB-A ports. Display output is handled by one DP 2.1 port and one HDMI 2.1 port. The unit weighs 2 kg and measures 183 x 183 x 51 mm.
Lenovo claims up to four units can cluster to provide 524 TOPS of aggregate AI performance. This configuration offers 512 GB of combined RAM for running models like OpenClaw and Meta's Llama 4 Maverick. The system carries a TUV Rheinland Ultra Low Noise certification for quiet operation. Users can order the machine with Microsoft Windows 11 or Linux.
The ThinkCentre X Ultra starts at 3699 USD (around $3,699). Availability begins in November 2026 across global markets. We looked at Lenovo ThinkPad P16v Launches with Core in our earlier Lenovo coverage.
The ThinkCentre X Ultra sets a new standard for compact desktops designed to run local AI models. The combination of high-speed RAM, Gen5 storage, and clustered compute options provides a scalable solution for local AI development.
Source: NotebookCheck
Discussion
0 comments