GPD G2 eGPU | Supports RTX 4090 / 5090 | - Shenzhen GPD Technology Co., Ltd.

GPD G2 eGPU | eGFX

WORLD'S FIRST

GPD G2

World's First MCIO GPU Dock

MCIO 8i + USB4® 2.0 dual-port architecture with 256Gbps full-symmetric bandwidth and perfect fit support, instantly transforming your handhelds, laptops, and Mini PCs into performance beasts.

See Specs

GPD G2 Ports

Built for Connectivity

The GPD G2 eGPU dock features dual high-speed MCIO 8i and USB4® Version 2.0 ports, covering all scenarios from exclusive high-speed connections to universal device compatibility. Whether you are a hardcore gamer seeking ultimate bandwidth or a multi-device user requiring broad compatibility, the GPD G2 has you covered.

PORT OVERVIEW

① MCIO 8i Port
② USB4® V2.0 Port
③ USB 3.2 Gen 2 Type-A × 2
④ RJ45 Ethernet Port
⑤ PCIe 5.0 ×16 Slot
⑥ Air Outlet

⑦ IEC 60320 C14 AC Inlet
⑧ 12V-2x6 GPU Power Port
⑨ Power Switch
⑩ Air Intake
⑪ GPU Mounting Screw Hole

Port Details

Each port, designed for peak performance.

ALL PORTS

🔌 MCIO 8i Port

PCIe Gen 4 × 8 | 256 Gbps

Supports all PCIe devices, including graphics cards. Native high-speed direct connection ensures zero-loss computing power.

⚡ USB4® V2.0 Port

80Gbps Symmetric | 120Gbps Asymmetric

PD 3.0 protocol with 100W (PD) output. When connected alongside MCIO, it functions as a Hub and PD power source; when connected independently, it operates as a primary eGPU.

🔌 USB Type A × 2

USB 3.2 Gen 2 | 10 Gbps

Dual high-speed USB-A ports for peripherals and storage, serving as the essential hub in docking station mode.

💾 NVMe M.2 Slot

M.2 2280 | PCIe 3.0 × 2

Supports external boot with a Gunmetal magnetic SSD cover. Achieve a dual upgrade for both "Graphics and Storage" in one go.

🌐 RJ45 Gigabit Ethernet Port

10/100/1000 Mbps Auto-Sensing

A single cable connection grants simultaneous access to high-speed networking and LAN sharing, drastically decluttering your desktop setup.

⚡ Powered By A 12V-2x6 Connector

12V-2x6（ATX 3.1）

Designed specifically for high-performance NVIDIA RTX 40 / 50 series GPUs, ensuring absolute stability under both instantaneous peaks and sustained full-load conditions.

The Ultimate Unleash of Gaming Performance

GPD BOX: Your Instant 4K/120Hz Gaming Console

The combination of the MCIO port and GPD BOX, paired with an RTX 4090, creates the ultimate gaming rig with a performance loss of only about 2%. It eliminates the need for performance compromises in compact devices, allowing players to carry the "core of a console" with them. By simply plugging in the GPU dock anywhere, you gain a peak experience that rivals top-tier gaming desktops. This is more than just an expansion; it’s a revolution in how we play.

MCIO 8i PORT

PCIe Gen 4 × 8

MCIO 8i Port

256 Gbps

Total Aggregate Throughput

31.5 GB/s

Usable Bi-directional Bandwidth

MCIO 8i Port Real-World Gaming Benchmarks

GPD BOX + GPD G2 + RTX 4090

Test Rig: GPD BOX (Ultra 7 356H Edition) + GPD G2 eGPU Dock + NVIDIA RTX 4090

All tests conducted at 4K Resolution + Max Settings + DLSS Performance + Frame Generation (Ray Tracing OFF).

BENCHMARK

League of Legends
VALORANT
Counter-Strike 2
Rocket League
Roblox
Overwatch 2
Warframe
PUBG: BATTLEGROUNDS
Dota 2
Apex legends
War Thunder
Ghostrunner 2
World of Warcraft
Delta Force
Shadow of the Tomb Raider
Returnal
Horizon Forbidden West
God of War: Ragnarok
Painkiller
Call of Duty: Black Ops 7
Grand Theft Auto V Enhanced
Battlefield 6
The Last of Us Part II
Fortnite
Dying Light: The Beast
Hogwarts Legacy
Baldurs Gate 3
F1 25
A Plague Tale: Requiem
Cyberpunk 2077
Resident Evil 4
Avatar: Frontiers of Pandora
Indiana Jones and the Great Circle
Clair Obscur: Expedition 33
Black Myth: Wukong
Assassin's Creed Shadows
Red Dead Redemption 2
S.T.A.L.K.E.R. 2: Heart of Chornobyl
Monster Hunter Wilds
Wuchang: Fallen Feathers
Kingdom Come: Deliverance II
Total War: WARHAMMER III
Sid Meier's Civilization VII

648
587
544
519
477
415
341
337
307
297
254
236
227
226
218
202
188
188
186
183
178
176
174
169
165
158
155
154
154
151
149
148
148
139
138
137
128
128
127
123
122
114
107

Game Name

FPS（AVG）

Note: Data based on real-world testing with GPD BOX (155H Edition) + GPD G2 + RTX 4090 at 4K resolution, Max settings, DLSS Performance, and Frame Generation (Ray Tracing OFF). Actual results may vary depending on game version, driver version, and system environment.

400+ FPS (eSports-level)

226~341 FPS (Smooth-level)

107~218 FPS (AAA-level)

Unleash Full Power with Dual-Link Connectivity

The GPD G2 eGPU dock features dual high-speed MCIO 8i and USB4® Version 2.0 ports, covering every scenario from exclusive high-speed connections to universal device compatibility. Whether you are a hardcore gamer seeking ultimate bandwidth or a multi-device user requiring broad compatibility, the GPD G2 has you covered.

OPTIMAL PERFORMANCE

MCIO 8i

Exclusive High-Speed | PCIe Gen 4 × 8

USB4® V2.0

Universal Compatibility | 80Gbps Bandwidth

CONNECTION MODES

Comparison of Three Connection Modes

The MCIO 8i and USB4® V2.0 ports can be used independently or simultaneously, working in smart synergy to unleash maximum performance.

The MCIO port takes full charge of graphics data transmission, unleashing the entire 256Gbps bandwidth to the GPU for near-lossless, stable transfer. It is ideal for professional computing scenarios where ultimate performance is paramount.

Ultimate Performance Mode

MCIO 8i Connection Only

✅ Full-Speed Graphics Data Transmission

✅ No Hub Functionality

✅ PCIe Gen 4 ×8 Full Bandwidth

Used as an external GPU (eGPU) connected via USB4® V2.0, it strikes the perfect balance between high bandwidth and ultimate compatibility. It is ideal for expanding desktop-class discrete graphics for handhelds and thin-and-light laptops.

Flexible Plug-and-Play Mode

USB4 V2.0 Connection Only

✅ 80Gbps (Sym) / 120Gbps (Asym)

✅ 100W PD Output

✅ Thunderbolt™ 3/4/5 Compatibility

Dual Connection: MCIO 8i + USB4® V2.0

✅ Full-Speed GPU + Hub Expansion

✅ 100W PD Output

✅ USB Port Expansion

Ultimate Expansion Mode

MCIO takes full charge of graphics data, while USB4 v2.0 shifts to become the hub for expansion and PD power. The GPD BOX evolves into a fully-featured desktop expansion center.

Note: The 120Gbps bandwidth in USB4® V2.0 asymmetric mode is achieved by reallocating four 40Gbps lanes, directing three lanes to the transmit direction. This enables a single-monitor asymmetric transmission mode that can effortlessly drive an 8K high-refresh-rate display or dual 6K displays, providing ample bandwidth headroom for future, more demanding display devices. This mode is intelligently triggered by the system, automatically activating when the required display bandwidth exceeds a preset threshold. It ensures that professional users experience ultra-smooth, uncompressed visuals in video editing, 3D rendering, or high-frame-rate gaming.

Bandwidth Comparison: MCIO 8i vs. USB4® V2.0 vs. OCuLink

Whether you are a professional researcher seeking ultimate computing power or a mobile creator in need of flexible expansion, the GPD G2 eGPU dock, with its dual-port architecture, offers two distinct high-speed interconnect options tailored to your needs.

BANDWIDTH COMPARISON

Bi-directional Theoretical Bandwidth
Symmetric Mode
Asymmetric Mode
Uni. Effective Data Bandwidth
Target Scenarios (or Best For)
Hot-swappable

128 Gbps Total Bi-directional Bandwidth
64 Gbps Transmit / 64 Gbps Receive
—
~7.88 GB/s (Unidirectional)
Laptops, Mini PCs, Gaming Handhelds
❌ Not Supported

256Gbps Total Bi-directional Bandwidth
128Gbps Transmit / 128Gbps Receive
—
~15.75 GB/s (Unidirectional)
HPC, Multi-GPU Synergy, AI Training
❌ Not Supported

160 Gbps Total Bi-directional Bandwidth
80 Gbps Transmit / 80 Gbps Receive
120Gbps Transmit / 40 Gbps Receive
~9.6 GB/s / ~14.5 GB/s
eGPU, High-Speed Peripherals, Plug-and-Play
✅ Supported

Specifications

Oculink（PCIe Gen 4 × 4）

MCIO 8i（PCIe Gen 4 × 8）

USB4® Version 2.0

Handheld Doesn't Have an MCIO 8i Port? Don't Worry!

USB4® Version 2.0: The Ultimate Slayer of iGPUs and Mobile dGPUs

Tired of sacrificing performance for heat and frame drops in AAA titles on your handheld? Fed up with stuttering renders and endless waiting on your ultra-portable laptop? Meet the GPD G2 eGPU Dock. Powered by USB4® Version 2.0, it shatters the OCuLink bandwidth bottleneck. With 160Gbps bi-directional ultra-high-speed connectivity, it breaks the performance chains of external graphics, linking you directly to high-end desktop GPUs. Make integrated and mobile graphics a thing of the past.

USB4® VERSION 2.0

The GPD G2 supports Thunderbolt™ 3, 4, 5, USB4, and USB4® V2.0 devices. It is compatible with Intel-based Apple devices (official graphics drivers from Apple may be required) but does not support Apple Silicon (M-series chips) devices.

Supported GPD products include: WIN 3, Pocket 3, WIN Max, WIN Max 2021, WIN Max 2, WIN 4, WIN Mini, Pocket 4, WIN 5, GPD DUO, and third-party products such as Microsoft Surface Pro 11.

MCIO Port · Your Exclusive AI Computing Dock

The MCIO 8i port provides native PCIe Gen 4 × 8 high-speed lanes with a bi-directional bandwidth of up to 256Gbps. This ensures that the massive power of top-tier gaming GPUs, such as the NVIDIA RTX 4090 and 5090, is injected into your host system with zero performance loss.

AI COMPUTING

🧠 AI Researcher & Developer

Leveraging mini-PCs with external RTX 4090/5090 GPUs for high-efficiency local model training and debugging, ensuring data privacy and security without the need for cloud uploads.

🎬 Content Creator

Combine GPU power with cutting-edge AI for smart super-resolution, AI background removal, and real-time effects—empowering creators with a massive productivity boost.

📊 Financial & Data Analyst

Leveraging robust local computing power to process massive datasets and execute complex quantitative trading models or risk assessment algorithms with high efficiency.

GPD G2 with Tesla V100: Build Your AI Training and HPC Clusters

Via PCIe Gen 5 Retimer/Switch expansion cards, the GPD G2 eGPU dock enables parallel expansion, significantly simplifying deployment while massively boosting overall system scalability.

HPC CLUSTER

Qwen3.5-122B
Qwen2.5-72B
Llama 3.1/4 70B
DeepSeek-R1-70B

LLaMA 4 Scout (109B MoE)
Mixtral 8x7B
Qwen3-32B
Llama 3.1/4 32B
DeepSeek-R1-32B

Approx. 600-800 TOPS (Post-Quantization)
Approx. 400-500 TOPS (Post-Quantization)
Approx. 400-500 TOPS (Post-Quantization)
Approx. 400-500 TOPS (Post-Quantization)

Approx. 300-400 TOPS (FP16)
Approx. 250-350 TOPS (FP16)
Approx. 200-250 TOPS (FP16)
Approx. 200-250 TOPS (FP16)
Approx. 200-250 TOPS (FP16)

122B Parameters, Dense
72B Parameters, Dense
70B Parameters, Dense
70B Parameters, Dense

109B Total Params, MoE (17B Active)
46.7B Total Params, MoE (~13B Active per Token)
32B Parameters, Dense
32B Parameters, Dense
32B Parameters, Dense

Yes (NVFP4/FP8 Recommended)
Yes (INT8/INT4 Recommended)
Yes (INT8/INT4 Recommended)
Yes (INT8/INT4 Recommended)

No (FP16 is sufficient)
Optional (FP16 may suffice)
No (FP16 is sufficient)
No (FP16 is sufficient)
No (FP16 is sufficient)

Model Name

TOPS

Parameters / Architecture

Quantization Required?

Via a PCIe Gen 5 Retimer NVMe expansion card, a single PCIe lane can be bifurcated into dual x16 or quad x8 channels, enabling seamless high-speed interconnects and parallel workflows for multiple GPUs. With its built-in power supply and native MCIO port, the GPD G2 directly replaces traditional riser cards and external power setups.

🔄Dual Tesla V100 Expansion: 64GB HBM2 VRAM

Note: The illustration features the Linkreal PCIe 5.0 x16 to Dual-Port Internal MCIO 8i Retimer 5.0 GPU Expansion Card.

Insert a high-performance Switch expansion card into a single PCIe Gen 5 x16 slot on your workstation, and connect directly to four GPD G2 eGPU docks via native MCIO 8i ports. By eliminating complex, multi-layered adapters, it establishes ultra-low latency, full-bandwidth direct channels between the workstation CPU and all four GPU docks.

🔄 Quad Tesla V100 Expansion: 128GB HBM2 VRAM

Note: The illustration features the Linkreal PCIe Gen 5 Switch Expansion Card to 4-Port Internal NVMe MCIO 8i.

A single workstation, utilizing a PCIe Gen 5 Switch expansion card, connects directly to four GPD G2 eGPU docks via MCIO ports. Each dock is equipped with an NVIDIA Tesla V100 32GB HBM2 GPU, featuring 900 GB/s bandwidth and a 4096-bit memory interface. With a combined 128GB of HBM2 VRAM, this setup creates a cost-effective and flexibly deployed inference platform for Large Language Models (LLMs).

🏆 GPD G2 (Quad V100): Mastering 100B+ Parameter Models

4 × GPD G2 (Quad V100): LLM Inference Capability Overview

Inference requirements vary significantly across different model scales. The following table summarizes the estimated TOPS requirements for mainstream models under Quantized / FP16 precision to help you determine the necessary hardware configuration. A 4 × V100 setup provides approximately 560 TOPS (FP16), which, combined with quantization techniques, can cover the majority of mainstream models.

TOPS = Tera Operations Per Second. A higher value indicates a greater requirement for computing power.

Note: The TOPS requirements listed above are estimated values. Actual computing power consumption varies based on quantization schemes, context length, batch size, and inference framework optimization. For MoE architectures, only a fraction of expert parameters are activated, making the actual computing requirements significantly lower than dense models of the same parameter scale.

Uninterrupted Creativity: RJ45 Gigabit Ethernet Port

High-Performance Expansion Meets Stable Networking

Eliminate the need for additional USB ports or wireless cards on your host device. A single Ethernet cable connected directly to the eGPU dock provides high-speed networking while streamlining your desktop cabling. Easily access office wired networks to meet security and management requirements, allowing your workstation to integrate seamlessly into any professional network topology.

GET ONLINE

🔗 All-in-One Connectivity | Maximum Minimalism

A single cable connection to the GPU dock eliminates the need for extra USB ports or wireless cards, delivering the most streamlined connectivity solution.

🖥️ AClean Desktop, Flexible Deployment

Completely streamline your desktop cabling to eliminate clutter, making workstation deployment more flexible and aesthetically pleasing.

🌐 Plug-and-Play: Seamless Integration into Enterprise Networks

Effortlessly connect to standard office or enterprise wired environments with full compatibility. It meets internal security policies and management requirements, ensuring seamless and rapid integration of your workstation into professional network topologies.

Plug and Expand, Hassle-Free

GPD G2 Built-in M.2 Slot

Whether you are a professional researcher seeking ultimate computing power or a mobile creator in need of flexible expansion, the GPD G2 eGPU dock, with its dual-port architecture, offers two distinct high-speed interconnect options tailored to your needs.

STORAGE

▶ M.2 2280 Specification, PCIe 3.0 × 2

▶ Supports External Storage Booting

▶ Gunmetal Magnetic Protective Cover

▶ Tool-Free Design, Shake-Proof Stability

80 Gbps (Bi-Directional)

9.6 GB/s

Full-Featured USB4 V2.0 × 2 (Host)

The Ultimate Device Interconnect

Connect to GPD G2: Single-Cable Integration for Data, Video, and Power.

10 Gbps

1.21 GB/s

6 × USB 3.2 Gen 2 Type-A Ports
4 on Host + 2 on eGPU

High-Speed External Storage & Peripherals

Connect Portable SSDs, High-Speed Card Readers, Capture Cards, and More.

2.5 ×2 Gbps
1 Gbps

312 MB/s
125 MB/s

2.5GbE RJ45 × 2 (Host)
1GbE RJ45 × 1 (eGPU)

Multi-Gigabit Networking Matrix

Direct NAS Connection & Link Aggregation for 4K Editing and High-Bandwidth Needs.
Redundant Networking, Device Management, and Network Isolation.

Ultra-High Speed Internal Storage

15.75 GB/s
3.5 GB/s
1.21 GB/s

3 × M.2 Slots
2 on Host + 1 on eGPU

PCIe Gen5 ×4
PCIe Gen4 ×2
PCIe Gen3 ×1

System Drive / High-Speed Cache Drive.
* The speed depends on the slot specifications and the SSD itself. The expansion dock's built-in M.2 slot is converted via a USB 3.2 Gen 2 interface. *

Functional Orientation

Core Applications / Notes

Bus Speed

Effective Data Rate

Port Types and Quantities

Redefining Your Personal High-Speed Storage Hub

Full-Speed Compact NAS

With an innovative "Storage + Expansion" 2-in-1 design, the GPD G2 eGPU transforms into a portable, high-bandwidth "Personal Full-Speed Storage Hub" via its built-in M.2 slot and dual USB 3.2 Gen 2 ports.

NAS STATION

▶ Core High-Speed Storage Pool: Built-in M.2 NVMe SSD, Up to 1GB/s Sustained Read/Write

▶ Flexible Capacity Expansion: Dual 10Gbps USB 3.2 Gen 2 Ports

▶ Local NAS Mode: SMB/AFP Sharing, Team-Ready

▶ Premium Combo: Dual Ethernet on GPD BOX + Single Port on GPD G2

FLAGSHIP SYNERGY

GPD BOX + GPD G2

CRAFTING THE ULTIMATE PORT COMBO

Unleashing the Full Potential of ITX Systems

Searching for High-Performance GPUs for Your Beloved ITX Build? Struggling Between Size and Power? The GPD G2 eGPU Dock Is Your Ultimate Solution. With Just an MCIO to PCIe Adapter, You Can Bring Desktop-Class Graphics Performance to Your Small Form Factor PC—Shattering Space Constraints Forever.

ITX COMPATIBLE

⚡ Perfect Fit: Zero Performance Loss

Through High-Speed MCIO Ports and Adapter Cards, the GPD G2 Direct-Connects Full PCIe x8 Lane Bandwidth to Your ITX Host. Performance Without Compromise.

🖥️ Minimalist Desktop: Focus on Essentials

Keep Your ITX Build Compact and Refined, While the GPD G2 Handles Heavy Graphics Rendering. Say Goodbye to Cable Clutter Forever.

🌐 One Dock for All: Build for the Future

Beyond Your ITX Build, Inject Desktop-Class Graphics Power Into Laptops, Mini-PCs, or Handhelds via USB4. One Setup, Multiple Empowerments.

🔧 MCIO to PCIe Adapter Card

Low-Profile Graphics Cards for ITX Cases Are Rare, and High-Performance Options Are Nearly Non-Existent. By Installing Our MCIO to PCIe Adapter Card to Connect the GPD G2, You Can Easily Swap Between Full-Height and Low-Profile Brackets—Ensuring Seamless Compatibility With Any ITX Chassis.

Reference Product: ICY DOCK EXLink MB409A5 — MCIO 8i to PCIe Gen 5 x 8 Expansion Card. Bridging PCIe Gen 5 x 8 Slots with MCIO 8i (SFF-TA-1016) Connectivity, Delivering Ultra-High Total Bandwidth of Up to 256 Gbps.

Gold-Rated Heart, Powering Infinite Possibilities

Built-in 800W Gold-rated ATX 3.1 PSU: Sustained full-load power for high-end GPUs like the RTX 4090 / 5090.

POWER SUPPLY

800W

Gold ATX 3.1 PSU

92%+

Conversion Efficiency

100~240V

Global Voltage Support

ATX 3.1

12V-2x6 Power Connector

Dual-Fan Intelligent Cooling System

Innovatively equipped with a dedicated independent dual-fan cooling system for its built-in 800W Gold ATX 3.1 PSU, the GPD G2 lowers temperatures at the source and eliminates heat buildup.

COOLING SYSTEM

Continuously draws in cool ambient air to directly cool the PSU and high-heat components, lowering temperatures at the source.

🌬️ Front Dual-Fan High-Efficiency Intake

💨 High-Airflow Rear Honeycomb Outlet

Working in synergy with the front fans, it creates a high-efficiency directed airflow through the PSU, rapidly and thoroughly exhausting heat from the chassis.

🌡️ Intelligent Fan Speed Control

Fan speeds adjust intelligently based on PSU load and internal temperatures—staying whisper-quiet at low loads and delivering full cooling power at high loads.

PD CHARGING

100W Fast Charging: Leave the Charger Behind

As long as your device features a Type-C port supporting the PD fast charging protocol, it can be charged via GPD G2's USB4 v2.0 port.

100W

PCIe GPU Compatibility

Full compatibility across NVIDIA, AMD, and Intel—from consumer gaming cards to top-tier AI accelerators.

COMPATIBILITY

Gaming GPUs

GeForce GTX 10 Series
GeForce RTX 20 / GTX 16 Series
GeForce RTX 30 Series
GeForce RTX 40 Series
GeForce RTX 50 Series
Radeon RX 400 / 500 Series
Radeon RX Vega Series
Radeon RX 5000 Series
Radeon RX 6000 Series
Radeon RX 7000 Series
Radeon RX 9000 Series
Arc A Series
Arc B Series

Pascal
Turing
Ampere
Ada Lovelace
Blackwell
GCN 4.0
GCN 5.0
RDNA 1.0
RDNA 2.0
RDNA 3.0
RDNA 4.0
Xe HPG
Xe2 HPG

NVIDIA

AMD

Intel

Brand

Series

Architecture
Professional GPUs

NVIDIA

AMD

Intel

Quadro K Series
Quadro M Series
Quadro P Series
RTX A Series
RTX PRO Series
Radeon Pro WX Series / W Series
Radeon Pro W6000 Series
Radeon Pro W7000 Series
Radeon AI PRO R9000 Series
Arc Pro A Series
Arc Pro B Series

Kepler
Maxwell
Pascal
Ampere
Blackwell
GCN
RDNA 2
RDNA 3
RDNA 4
Xe
Xe2

Brand

Series

Architecture
AI Accelerators

NVIDIA

Tesla Series（K/M/P/V/T）
A100 PCIe
H100 / H200 PCIe
B100 PCIe
Instinct MI25 / MI6 / MI8
Instinct MI100
Instinct MI200 Series
Instinct MI300 Series
Instinct MI350 Series
Data Center GPU Flex / Max Series
Gaudi 3 AI Accelerators
Crescent Island

Multi-Architecture
Ampere
Hopper
Blackwell
GCN
CDNA 1
CDNA 2
CDNA 3
CDNA 4
Xe-HPG / Xe-HPC
Gaudi 3
Xe3P

AMD

Intel

Brand

Series

Architecture

Detailed Specifications

SPECIFICATIONS

Basic Specs
Brand
Model
Product Type
Target Devices

OS Support
Connectivity
PCIe Slot
MCIO 8i
USB4 V2.0
USB Type-A
M.2 Slot
LAN Port
Power Specifications
Type
Input
GPU Power Connector
Total Power
Appearance & Materials
Dimensions
Color
Material
Net Weight
Accessories
Power Cord
GPU Power Cable ×2
Others

GPD
G2
eGFX / eGPU
Consumer PCs and high-end gaming consoles, high-end gaming handhelds (supporting Thunderbolt 3/4/5 or USB4/USB4 v2.0), workstations, servers, and HPC scenarios.
Windows 11, Linux (Kernel 6.17 and above required)

× 1, PCIe Gen 5 × 16
× 1, PCIe Gen 4 × 8, bidirectional theoretical speed 256 Gbps, bidirectional effective bandwidth ~31.5 GB/s
× 1, Symmetric 80 Gbps / Asymmetric 120/40 Gbps, PD 3.0 100W
× 2, USB 3.2 Gen 2, 10 Gbps
× 1, M.2 2280, PCIe 3.0 × 2, Supports external storage booting
× 1, RJ45, 10/100/1000 Mbps

ATX 3.1 Gold, 92%+ Efficiency
100-240V 50/60Hz
12V-2x6
800W (12V)

157.3 × 119.8 × 182 mm
Gunmetal
Aluminum-Magnesium Alloy, PC 94V-0
1605 g

IEC 60320 C13 Power Cord, 950mm
12V-2x6 Extension Cable 16 Pin (12+4) 500mm + 16 Pin to Dual 8 Pin 500mm
Gunmetal SSD Magnetic Cover × 1, Warranty Card × 1, User Manual × 1

Frequently Asked Questions

FAQ

Which graphics cards does GPD G2 support?

Compatible with NVIDIA GeForce GTX 10/16/RTX 20/30/40/50 series, AMD Radeon RX 400~9000 series, and Intel Arc A/B series consumer graphics cards; NVIDIA Quadro/RTX PRO series and AMD Radeon Pro W series workstation graphics cards; NVIDIA Tesla/A100/H100/B100, AMD Instinct MI series AI compute cards, and more.

How should I connect my graphics card if it uses dual 6 Pin (or 6 Pin + 8 Pin) connectors?

You can purchase a 12VHPWR (16 Pin) to dual 8 Pin (6 Pin + 2 Pin) GPU power cable.

Can MCIO and USB4 v2.0 ports be used simultaneously?

Yes, it is possible. When connected simultaneously, the MCIO port is dedicated to handling graphics card data transfer, while the USB4 v2.0 port functions as a Hub expansion and PD power delivery center, powering connected devices and providing additional USB interfaces.

Can the GPD G2 charge devices?

Yes, it is possible. As long as the device has a Type-C port and supports the PD fast charging protocol, it can receive a maximum charging output of 100W through the USB4 v2.0 port.

Which operating systems does the GPD G2 support?

Supports Windows 11 64-bit and various Linux distributions (requires kernel 6.17 or higher). Corresponding graphics drivers and USB4 v2.0 controller drivers must be installed before use.

How to Connect an ITX Host to the GPD G2?

Connect via an MCIO-to-PCIe adapter card. The adapter supports interchangeable full-height and low-profile brackets, ensuring compatibility with various ITX cases.

Does the M.2 slot support booting from external storage?

It supports using an SSD as a bootable installation drive, in the M.2 2280 form factor, with a PCIe 3.0 × 2 interface via an internal USB adapter, and includes a purple-grey magnetic protective cover.

Data Specifications:

* Performance & Specifications: All data regarding memory capacity, USB transfer speeds, dimensions, thickness, weight, and screen angles are theoretical design values. Actual measurements may vary within a reasonable margin of error. Unless otherwise specified, all data is provided by processor manufacturers or our internal laboratories.
* Software & Interface: Due to real-time software updates and production batches, pre-installed software versions, UI interfaces, and application names may vary. Please refer to the actual device experience.
* Production Batches: Product specifications, such as weight, may vary by batch. We recommend consulting with your seller for detailed information at the time of purchase.
* Visual Representation: Product images and colors are for reference only. Due to photographic lighting and display settings, the actual product may vary. Please refer to the physical unit.
* Updates & Revisions: All product information, specifications, applications, images, and features mentioned on this website are for reference only. Content is subject to update or revision at any time without prior notice.
* Components & Bundles: PCB layouts and bundled software may differ across production batches and are subject to change without notice.
* Trademarks: All product names mentioned herein are for identification purposes only and may be the registered trademarks or copyrights of their respective companies.
* HDMI Compliance: The terms HDMI, HDMI High-Definition Multimedia Interface, HDMI Trade Dress, and the HDMI Logos are trademarks or registered trademarks of HDMI Licensing Administrator, Inc.
* Pricing & Fees: Prices may not include additional costs such as taxes, shipping, handling, or recycling fees.
* MSRP: For pricing information, GPD only sets the Manufacturer's Suggested Retail Price (MSRP). All distributors and retailers are free to set their own pricing.