GMKtec EVO-T2 mini PC delivers 180 TOPS for local AI computing with Intel Panther Lake

GMKtec EVO-T2 Brings High-Performance AI Computing to Compact Desktop PC

GMKtec has officially launched the GMKtec EVO-T2, a compact desktop PC designed specifically for local AI workloads rather than conventional mini PC tasks. Unveiled during Intel's third-generation Core Ultra launch event in Shanghai, the system delivers 180 TOPS of combined AI performance by leveraging CPU, GPU, and NPU resources together2

2

. The device represents a shift toward more serious workstation-style use cases, enabling developers and researchers to run large language models with up to 70B parameter models directly on the device without external cloud infrastructure1

1

. This positioning matters for users seeking lower latency, greater control over data, and reduced dependence on cloud services.

Intel Panther Lake Architecture Powers On-Device AI Model Execution

The EVO-T2 is built around Intel Panther Lake platform and comes in two configurations: a Core Ultra X9 388H variant and a Core Ultra X7 358H variant. Both feature up to 16 cores in a 4 P-Core + 8 E-Core + 4 LP-E-Core configuration, clocked at up to 5.1 GHz and 4.8 GHz respectively3

3

. Manufactured using Intel's 18A process technology, the system incorporates RibbonFET gate-all-around transistors and PowerVia backside power delivery, which Intel claims improves efficiency and transistor density1

1

. Graphics capabilities are handled by the integrated Intel Arc B390 GPU with 12 Xe cores, supporting DirectX 12 Ultimate, hardware ray tracing, and AI-assisted upscaling, extending the system's utility beyond AI inference into rendering and content creation workflows2

2

.

aiDAPTIV+ AI SSD Technology Addresses DRAM Limitations

One of the most notable features is GMKtec's partnership with Phison to integrate aiDAPTIV+ AI SSD technology, which dynamically extends available memory by distributing workloads between DRAM and storage1

1

. This "pseudo-memory" mechanism allows large models to be segmented during execution, with active portions processed on the GPU while less active data remains stored across memory and SSD layers. The system ships with up to 64 GB of LPDDR5X-8533 memory soldered to the motherboard, and Phison SSDs expand memory capabilities through caching, offering a boost to memory performance3

3

. The EVO-T2 includes dual M.2 storage slots supporting PCIe 5.0 and PCIe 4.0, with total capacity reaching up to 16 TB. PCIe 5.0 SSDs can achieve sequential speeds exceeding 10GB/s, with some surpassing 15 GB/s, while PCIe 4.0 drives typically reach around 7GB/s under optimal conditions1

1

.

Connectivity and Thermal Design Support Local AI Workloads

Connectivity is robust for the form factor, with USB4 ports offering 40 Gbps transfer speeds and 100W power delivery support, OCuLink for external GPU expansion, and dual Ethernet with 10GbE plus 2.5GbE networking2

2

. The chassis features a triple exhaust airflow design with two fans—one for the SSD compartment and one for the SoC, which uses a vapor chamber. The system sustains a maximum TDP of 60W and can switch to a 45W balanced mode while offering near-quiet operation3

3

. Additional features include Wi-Fi 7 support, Bluetooth 5.4, and the ability to drive four 4K screens simultaneously.

Security Concerns Cloud OpenClaw Integration

The EVO-T2 ships with a pre-configured AI environment called GMKtec Claw, allowing immediate access to AI agents and models without manual setup. However, the device supports OpenClaw despite warnings from the Chinese government about security vulnerabilities1

1

. The software has known weaknesses that can be exploited to steal sensitive data and has previously been used to distribute malware via GitHub repositories. Microsoft has advised against running OpenClaw on standard personal or enterprise devices, raising questions about GMKtec's decision to include it. The X7 variant with 64 GB memory is currently listed for 9999 RMB, approximately $1,400-$1,500 USD3

3

. For organizations and developers prioritizing on-device AI model execution, the EVO-T2 presents an intriguing option, though the security implications of its software stack warrant careful consideration before deployment.