(PR) GMKtec Unveils 2nd Gen EVO-T2 Desktop AI Supercomputer with 180 TOPS

Key Takeaways

• 180 TOPS AI performance in desktop form factor
• Supports 70B parameter LLM locally
• Intel 18A process with RibbonFET and PowerVia
• Integrated AI SSD reduces memory bottlenecks
• Dual Ethernet, USB4, OCuLink expandability

Summary

GMKtec introduced the second‑generation EVO‑T2 Desktop AI Supercomputer, built on Intel’s Core Ultra X9 388H and X7 358H processors using the 18A (1.8 nm) Panther Lake architecture. The compact system delivers up to 180 TOPS of AI compute and can run local large language models of up to 70 billion parameters without cloud dependence. Integrated Intel Arc B390 GPU, Phison aiDAPTIV+ AI SSD, and extensive I/O—including dual Ethernet, USB4, and OCuLink—provide a full‑stack solution for edge AI workloads. The EVO‑T2 ships with a pre‑configured GMKtec Claw AI environment for immediate deployment.

Pulse Analysis

The EVO‑T2 represents a pivotal moment for edge AI, marrying Intel’s latest 18A (1.8 nm‑class) process technology with a tightly integrated CPU‑GPU‑NPU architecture. RibbonFET gate‑all‑around transistors and PowerVia backside power delivery deliver unprecedented efficiency, allowing the desktop unit to sustain 180 TOPS while maintaining a modest thermal envelope. This level of compute, traditionally reserved for rack‑mounted servers, now fits on a desk, opening new use cases in autonomous robotics, real‑time analytics, and on‑premise AI research.

Beyond raw performance, GMKtec’s partnership with Phison to embed aiDAPTIV+ AI SSD technology addresses a common bottleneck in large‑model inference: memory bandwidth. The SSD dynamically expands usable memory, ensuring that 70‑billion‑parameter language models can be loaded and processed without stalling. Coupled with the Intel Arc B390 GPU’s 12 Xe cores, which support DirectX 12 Ultimate, ray tracing, and XeSS 3 upscaling, the system offers a versatile platform for both AI and high‑end graphics workloads, appealing to developers who need a single device for mixed‑reality and AI development.

From a market perspective, the EVO‑T2’s extensive I/O—dual 10 GbE/2.5 GbE Ethernet, USB4 with 100 W delivery, and OCuLink for external GPU expansion—positions it as a hub for future‑proof edge deployments. Enterprises can integrate the unit into existing networks, leverage high‑speed storage, or attach additional accelerators as workloads evolve. By delivering a ready‑to‑run GMKtec Claw AI environment, the product reduces time‑to‑value, making sophisticated AI accessible to smaller firms and research labs that previously required costly cloud credits. This launch underscores the accelerating convergence of high‑performance computing and AI at the desktop level, a trend likely to reshape how organizations architect their AI infrastructure.