On March 12, 2026, Intel officially unveiled its third-generation Intel Core Ultra processors at a major launch event in Shanghai. As a key ecosystem partner of Intel, GMKtec introduced its next-generation flagship EVO-T2 Desktop AI Supercomputer Series, jointly presented with Intel and Phison.

The EVO-T2 Series features up to 180 TOPS of local AI computing power and was personally endorsed with a signature by Intel CEO Lip-Bu Tan, making it one of the highlighted demonstration systems at the event for local AI workloads and large-model applications.

The new system combines cutting-edge semiconductor technology, high-performance computing, and extensive expandability, redefining what a desktop AI supercomputing workstation can deliver.

The EVO-T2 Series has officially launched on JD.com in China, while global availability via the GMKtec official website is currently being prepared. Early buyers will receive exclusive launch benefits and a pre-configured GMKtec Claw AI environment, allowing users to run local AI tools and models immediately out of the box.

Built on Intel’s 18A Process: A New Era of Performance and Efficiency

The EVO-T2 Series is powered by the Intel Core Ultra X9 388H and X7 358H processors (Panther Lake), representing Intel’s first consumer platform manufactured with the advanced 18A (1.8nm-class) process node.

This breakthrough architecture integrates two key semiconductor innovations:

● RibbonFET gate-all-around transistors

● PowerVia backside power delivery

Together, they deliver significant improvements in performance, efficiency, and transistor density, marking Intel’s transition into the sub-2nm era of consumer computing.

The processors feature a three-tier hybrid architecture consisting of:

● Cougar Cove Performance Cores

● Darkmont Efficient Cores

● Darkmont Low-Power Efficient Cores

The chip includes 16 cores (4P + 8E + 4 LP-E), enabling intelligent workload scheduling and optimized power distribution.

Key performance improvements include:

● 18% IPC improvement on Performance cores

● 18MB L3 cache

● 7.3% single-thread performance gain

● Over 30% multi-thread improvement

In Cinebench 2024, the processor achieved 1207 points in multi-thread performance at 45W, outperforming many systems in the same class and handling demanding workloads such as 3D rendering, content creation, and AI inference with ease.

Intel Arc B390 GPU: Desktop-Class Graphics in a Mini PC

The EVO-T2 also integrates the Intel Arc B390 GPU, based on Intel’s Xe3 architecture.

Key GPU specifications include:

● 12 Xe cores

● Up to 2.5 GHz dynamic frequency

● DirectX 12 Ultimate support

● Real-time ray tracing

● Intel XeSS 3 AI upscaling

This level of graphics performance rivals mid-range discrete GPUs, enabling smooth performance in:

● AAA gaming

● 4K video editing

● 3D modeling and rendering

● AI visual workloads

The EVO-T2 challenges the traditional perception that mini PCs cannot deliver powerful graphics performance.

180 TOPS Local AI Power: Run Large AI Models on Your Desktop

One of the defining features of the EVO-T2 Desktop AI Supercomputer is its up to 180 TOPS AI computing capability.

This is powered by the combined acceleration of:

● CPU

● GPU

● NPU

The system fully meets Microsoft Windows 11 AI+ PC requirements, enabling high-performance local AI computing without cloud dependence.

With this level of compute power, the EVO-T2 can smoothly run large AI models up to 70B parameters, supporting a wide range of applications:

● AI document processing

● Multimodal image generation

● Voice interaction

● Code generation

● Contract analysis

● AI development workflows

In internal tests using LM Studio, AI model execution speed improved multiple times compared to traditional PCs, with tasks such as generating hundreds of lines of code completed in around 40 seconds.

Running AI models locally also ensures maximum data privacy, since sensitive data never needs to be uploaded to the cloud—making the system ideal for developers, creators, and enterprise users.

To simplify the user experience, the EVO-T2 ships with the GMKtec Claw AI environment, pre-installed with popular AI tools and models so users can start immediately without complex setup.

Phison aiDAPTIV+ AI SSD: Solving Memory Bottlenecks for Large Models

The EVO-T2 Series also integrates Phison’s aiDAPTIV+ AI SSD technology, which dynamically expands memory resources to improve AI performance.

The system uses aiDAPTIV Cache, a NAND-based intelligent caching architecture that effectively breaks through traditional DRAM limitations in AI workloads.

The technology allows large models—such as 70B parameter models—to be dynamically segmented during training or inference:

● Active model slices are processed directly by the GPU

● Remaining portions are stored in DRAM and AI SSD

This dramatically reduces memory pressure while maintaining smooth and stable AI performance, enabling powerful AI workloads even without extremely large RAM configurations.

Massive Expandability for Professional Workflows

Despite its compact design, the EVO-T2 offers extensive expandability.

Storage

● Dual M.2 SSD slots

● Support for PCIe 5.0 + PCIe 4.0

● Up to 16TB total storage capacity

● RAID support for improved speed and data security

Connectivity

● USB4 full-function port

● 40Gbps data transfer

● 100W PD power delivery

● OCuLink external GPU support

With OCuLink, users can connect high-end external GPUs, transforming the system into an even more powerful AI or rendering workstation.

Networking

● 10GbE + 2.5GbE dual Ethernet

● Ideal for high-speed data transfer and enterprise networking

Advanced Cooling in a Compact Design

The EVO-T2 features an advanced 3D airflow cooling system, ensuring stable performance even under sustained workloads.

Key thermal design features:

● Up to 60W TDP performance capability

● 45W balanced mode for optimal efficiency

● Quiet operation under heavy workloads

The result is a system that combines high performance, low noise, and a clean desktop footprint.

Redefining Desktop AI Computing

At the launch event, the Founder of GMKtec emphasized the company’s vision:

“GMKtec has always been driven by user needs and technological innovation. By combining flagship mobile processors with desktop-level expandability, the EVO-T2 Series brings powerful AI computing directly to desktops, small businesses, and creators. Our goal is to make supercomputing accessible to everyone.”

Intel representatives also praised the collaboration, stating that the GMKtec EVO-T2 is a benchmark product for the third-generation Intel Core Ultra platform in the mini PC segment, helping accelerate the adoption of AI computing devices worldwide.

About GMKtec

GMKtec is a national high-tech enterprise specializing in the design, manufacturing, and global distribution of intelligent computing devices, including desktop AI supercomputers, mini PCs, portable GPUs, and NAS systems.

The GMKtec brand has been recognized as a “Shenzhen Famous Brand,” alongside leading technology companies such as Huawei and DJI.

The company has received multiple international design honors, including:

● American Design Award – Platinum Winner (2025)

● French Design Award – Gold Winner (2025)

Today, GMKtec products are sold in over 70 countries, supported by distribution and service networks in more than 35 markets.

Through partnerships with global technology leaders such as Intel, Microsoft, and AMD, GMKtec continues to drive innovation in the next generation of desktop AI computing platforms.

About Intel

Intel (Intel Corporation) is a global leader in the semiconductor industry. Founded in 1968 in Silicon Valley by Gordon Moore and Robert Noyce, the company established the technological foundation of the PC era through the x86 processor architecture.

Today, Intel’s core business spans client computing, data center platforms, and AI processors, supported by its IDM 2.0 strategy, which focuses on advancing semiconductor manufacturing and expanding foundry services.

Intel’s latest 18A (1.8nm-class) process node introduces two major innovations:

● RibbonFET gate-all-around transistors

● PowerVia backside power delivery

These technologies significantly improve performance, power efficiency, and transistor density, providing critical infrastructure for AI computing and high-performance computing (HPC).

As a key provider of compute power in the AI era, Intel is actively expanding its AI inference ecosystem and collaborating with industry partners across hardware and software platforms. The company has also maintained a strong presence in the Chinese market for more than 40 years, working with over 15,000 local partners to support innovation and ecosystem growth.

With advanced process technologies and full-stack computing solutions, Intel continues to drive the evolution of the global semiconductor industry and computing ecosystem.

About Intel CEO Lip-Bu Tan

Lip‑Bu Tan is one of the most influential leaders in the global semiconductor industry. Born in 1959 into a Malaysian Chinese family, he holds a Bachelor’s degree in Physics from Nanyang Technological University, a Master’s degree in Nuclear Engineering from MIT, and an MBA from the University of San Francisco.

In 1987, Tan founded Walden International, a venture capital firm that invested in more than 120 semiconductor companies, including leading firms such as SMIC and Advanced Micro-Fabrication Equipment Inc. (AMEC). His investments helped bridge the semiconductor ecosystems of the United States and China.

From 2009 to 2021, Tan served as CEO of Cadence Design Systems, where he doubled company revenue and increased its stock price by more than 3,200% during his tenure.

In March 2025, Tan was appointed CEO of Intel, becoming the first Chinese-heritage CEO in the company’s 57-year history. Under his leadership, Intel is accelerating its strategic transformation with a strong focus on AI computing and semiconductor manufacturing services.

Recognized as a visionary leader in the semiconductor field, Tan has received numerous industry honors, including the prestigious Robert N. Noyce Award. With deep experience across technology, venture investment, and corporate leadership, he continues to play a key role in shaping the global semiconductor landscape.

Keynote Speakers at the Launch Event

Several prominent technology leaders and industry figures delivered keynote presentations during the event, including:

● Intel Vice President and General Manager of Software Engineering and Client Products (China)

● Feng Dawei, Vice President of Intel’s Client Computing Group and General Manager of the Personal Computer Division

● Tim, founder of the popular Chinese technology media channel YingShi JuFeng

● Gao Yu, General Manager of Intel China’s Technical Department

These speakers shared insights on AI computing trends, next-generation PC architecture, and the future of AI-enabled devices.

Event Venue

The launch event was held at The Orbit West Bund Dome, a landmark technology and innovation venue located in Shanghai’s West Bund district.

Industry partners, developers, media representatives, and technology enthusiasts gathered at the venue to witness the debut of Intel’s next-generation AI computing platform and the unveiling of new ecosystem products such as the GMKtec EVO-T2 Desktop AI Supercomputer.