Skip to content

🎉 Celebrate Our 7th Anniversary (May 23 – June 10) – Enjoy Exclusive Deals! Discover More

News

GMKtec × AMD ROCm Lab Build a New Edge AI Ecosystem

by GMKtecOfficial 11 Jun 2026 0 comment

On June 11, 2026, GMKtec, a leader in desktop AI supercomputing, officially announced a deep strategic partnership with AMD ROCm Lab. Together, the two parties will build a new lightweight edge AI ecosystem, leveraging the ROCm computing ecosystem and GMKtec’s hardware portfolio to break down barriers to edge AI deployment. This collaboration aims to deliver cost-effective AI computing terminals for individual users, developers, and small studios, accelerating the adoption and real-world innovation of edge AI technologies.

Today, AI computing is rapidly shifting from cloud-centric infrastructure to localized edge deployment. However, traditional edge AI deployment still faces major challenges, including limited computing power, complex adaptation, and high costs. For individual developers and small teams in particular, access to high-performance AI computing devices remains expensive and difficult. The collaboration between GMKtec and AMD ROCm Lab directly addresses these industry pain points. Centered on hardware-software synergy and ecosystem co-building, the partnership connects the full chain from underlying computing power to end-user applications, making edge AI deployment simpler, more efficient, and more affordable.

As AMD’s core software ecosystem platform for AI workloads, the ROCm open-source software stack provides full-stack computing scheduling capabilities and supports both Windows and Linux systems. It offers significant advantages in local AI model deployment, heterogeneous computing acceleration, and developer tool adaptation, making it highly suitable for diverse edge AI scenarios such as large model inference, multimodal generation, and AI agent development. GMKtec, with seven years of experience in the Mini PC market, is one of the first brands in China to strategically develop desktop AI supercomputing. With its core strengths in extreme performance and portable user experience, GMKtec has launched several benchmark AI computing terminals, including the EVO-X2 and EVO-X3.

Among them, the EVO-X2 is the world’s first desktop AI supercomputer powered by the AMD Ryzen™ AI Max+ 395 processor. Built on the Zen 5 architecture with 16 cores and 32 threads, it integrates an XDNA2 NPU and RDNA 3.5 integrated graphics, delivering up to 126 TOPS of total AI computing power. The EVO-X2 can smoothly run large models with more than 200B parameters, such as Qwen3-235B. In LM Studio, its AI performance is 2.2 times higher than NVIDIA GeForce RTX 4090, while consuming less power. As GMKtec’s major new product of the year, the EVO-X3 continues the flagship processor configuration and natively features a high-speed OCuLink interface, enabling external GPU expansion to push computing boundaries even further. With 128GB LPDDR5X-8000 high-frequency memory, it supports parallel multi-model operation and further raises the ceiling of edge AI computing performance.

Based on this partnership, GMKtec and AMD ROCm Lab will realize deep integration and native adaptation between the ROCm ecosystem and GMKtec hardware. GMKtec’s full range of AI computing terminals will come pre-installed with an optimized ROCm development environment, enabling users to access key development tools such as HelloROCm right out of the box and simplifying the setup process for developers. AMD ROCm Lab will provide GMKtec with underlying technical support, optimizing compute scheduling and accelerating model adaptation for edge AI scenarios, while continuously iterating hardware-software collaborative solutions.

This collaboration brings precise value to different user groups. For individual users, GMKtec AI terminals offer a cost-effective way to run local AI painting, copywriting, video editing, and intelligent assistant applications while ensuring both data privacy and ease of use. For developers, the rich tools of the ROCm open-source ecosystem and the powerful computing performance of GMKtec hardware enable efficient large model fine-tuning, multimodal application development, and AI agent training without the need to invest heavily in computing clusters. For small studios, lightweight edge AI solutions can meet the needs of small-scale AI project development, testing, and deployment, helping teams accelerate technical innovation and commercial transformation.

AMD has always maintained that open source is key to accelerating AI innovation, and remains committed to building an open, powerful, and developer-friendly AI ecosystem. As AMD’s long-term deep strategic partner, GMKtec has already built a strong foundation of collaboration with AMD. From the major debut of the EVO-X2 at CES 2026 and its public recommendation by Dr. Lisa Su, Chair and CEO of AMD, to the EVO-X2 and EVO-X3 receiving signed recognition at AMD AI Developer Day, the two companies have completed a deep transition from hardware collaboration to ecosystem co-building. This joint announcement with ROCm Lab marks an important upgrade in the strategic partnership and signals that edge AI ecosystem development has entered a new stage.

Looking ahead, GMKtec and AMD ROCm Lab will continue to deepen their collaboration, focusing on lightweight edge AI technology innovation. The two parties will keep optimizing hardware-software synergy and expanding edge AI application scenarios. At the same time, they will work closely with developer communities through technical salons, development competitions, training programs, and other initiatives to strengthen the ROCm developer ecosystem and attract more developers to participate in edge AI innovation. Together, GMKtec and AMD ROCm Lab will help drive the high-quality development of China’s edge AI industry and make AI computing power truly accessible to everyone.

Prev Post
Next Post

Leave a comment

Please note, comments need to be approved before they are published.

Thanks for subscribing!

This email has been registered!

Shop the look

Choose Options

Edit Option

Choose Options

this is just a warning
Login
Shopping Cart
0 items