Banner Image

GIGABYTE
Accelerated
Deployment
Unit
is

a prefabricated, containerized, modular data center for on-demand token generation. With integrated compute, networking, and storage augmented by innovative power supply, cooling, and cabling, GADU actualizes rapid-deploy AI infrastructure with a minimal footprint in variable environments. GADU provides AI primacy, cost efficiency, regulatory compliance, and on-prem computing, making it ideal for foundational AI or flexible scalability. Backed by our data center architecture services and L12 total solutions, GADU shows the way to AI success.

動畫使用,勿刪除

Banner Image

Accelerated

Prefabricated, plug-and-play modules
expedite time-to-AI by up to 4x faster
Banner Image

Best-fit

Tailor-made configurations scale on
demand with evolving workloads
Banner Image

Controlled

Adaptive, all-environment design
empowers visionary AI conquests

The Pinnacle of AI at Your Service

Content Image

Ubiquitous Supercomputing

GADU can be preconfigured with ultrahigh-density cluster- and rack-level solutions that form the backbone of everything AI. For scalable supercomputing, GIGAPOD offers turnkey AI architecture that combines arrays of topline GPUs into a single cohesive unit. For rackmount systems, our AI Servers add oomph to your growth engine, while our OCP Servers inject versatility into your intelligent transformation.

Content Image

Sustainable Green Computing

GADU can improve power usage effectiveness (PUE) and reduce carbon footprint by incorporating direct liquid cooling (DLC) and immersion cooling. We supply in-row coolant distribution units (CDUs), sidecars, rear door heat exchangers (RDHx), and immersion tanks with built-in CDUs. GIGAPOD supplements DLC with G-REX: our middleware platform for one-stop cooling management.
Content Image

Software Stack for AI & HPC

Unleash hardware's maximum potential with our software stack. Featuring data center infrastructure management (DCIM), real-time monitoring, and workload orchestration, GIGABYTE POD Manager (GPM) gives you complete control over your resource pool. You can further enrich your AIOps environment with software toolkits from NVIDIA, data storage experts, cybersecurity providers, and more.

The Perfect Fit for Your AI Vision

SKU

CPUs

GPUs

Combination

Racks

(Compute + Network/MGMT)

CDU Type / Quantity

In-row Cooler Quantity

IT Power (MW)

NVIDIA Vera Rubin NVL72

144

288

1 x IT container*

1 x power skid**

2 x chillers

9 x 48U

(4 + 5)

In-row / 2 (N+1)

6 (N+1)

1.12

NVIDIA HGX Rubin NVL8

72

288

1 x IT container

1 x power skid

2 x chillers

9 x 48U

(4 + 5)

In-row / 2 (N+1)

6 (N+1)

1.03

NVIDIA GB300 NVL72

144

288

1 x IT container

1 x power skid

2 x chillers

9 x 48U

(4 + 5)

In-row / 2 (N+1)

6 (N+1)

0.70

NVIDIA HGX B300

72

288

1 x IT container

1 x power skid

1 x chiller

7 x 42U

(5 + 2)

In-row / 2 (N+1)

6 (N+1)

0.56

AMD Instinct MI355X

64

256

1 x IT container

1 x power skid

1 x chiller

7 x 42U

(4 + 3)

In-row / 2 (N+1)

6 (N+1)

0.57

* The main module that aggregates compute, networking, and storage equipment in standardized shipping containers.

** Preassembled, self-contained unit integrating critical power components to provide reliable electrical infrastructure.

All materials provided herein are for reference only. GIGABYTE reserves the right to modify or revise the content at any time without prior notice.

What's Next for Your AI Infrastructure?

GIGABYTE Goes Where AI Is Needed

AI Transformation

AI Transformation

GADU was pivotal to a century-old Taiwanese conglomerate's entry into AI data centers—both as a PoC and as an AI compute provider for internal use and external services.
Telecom & Edge

Telecom & Edge

GIGABYTE helped prototype a Japanese telco giant's mobile, containerized immersion cooling data center for network-edge deployment to handle spikes in demand.
Theoretical Science

Theoretical Science

A top research university boosted the capabilities of its data center by adding a cluster made by GIGABYTE, enabling brilliant young minds to pioneer bold discoveries.
Space Exploration

Space Exploration

Breakthroughs beyond the stars can be achieved without burdening the environment, thanks to the debut of GIGABYTE's DLC servers at a renowned European space agency.

Resources

Giga Computing Accelerates Enterprise AI at Scale with 6th Gen AMD EPYC™ Server CPUs

Giga Computing Accelerates Enterprise AI at Scale with 6th Gen AMD EPYC™ Server CPUs

GIGABYTE and Fabric8Labs Deepen Liquid Cooling Collaboration to Advance Next-Generation ECAM Technology for AI Infrastructure

GIGABYTE and Fabric8Labs Deepen Liquid Cooling Collaboration to Advance Next-Generation ECAM Technology for AI Infrastructure

What is Sovereign AI? Why is It Essential to Your AI Strategy?
Article

What is Sovereign AI? Why is It Essential to Your AI Strategy?

Giga Computing Expands GIGAPOD AI & HPC Infrastructure Platforms to Accelerate Deployment of Enterprise AI Factories

Giga Computing Expands GIGAPOD AI & HPC Infrastructure Platforms to Accelerate Deployment of Enterprise AI Factories

Supporting JHPC-quantum Project with a Next-Generation High-Performance Computing System
Article

Supporting JHPC-quantum Project with a Next-Generation High-Performance Computing System

GIGABYTE Connects AI, HPC, and Next-Generation Infrastructure at ISC High Performance 2026

GIGABYTE Connects AI, HPC, and Next-Generation Infrastructure at ISC High Performance 2026

HPC Ignites: How Data-Centric Mindset Revolutionizing Enterprise AI Infrastructure
Article

HPC Ignites: How Data-Centric Mindset Revolutionizing Enterprise AI Infrastructure

The Data Revolution in AI Factories: Driving High-Speed Networking Forward
Article

The Data Revolution in AI Factories: Driving High-Speed Networking Forward

DCIM x AIOps: The Next Big Trend Reshaping AI Software
Article

DCIM x AIOps: The Next Big Trend Reshaping AI Software

GIGABYTE Deep Dive: How We Built Our Industry-leading Liquid Cooling Solution
Article

GIGABYTE Deep Dive: How We Built Our Industry-leading Liquid Cooling Solution

Frequently Asked Questions

What Does "Prefabricated Modular" Mean in a Data Center?

The "prefabricated" quality of GADU means that most of the solution was installed and configured in a factory, therefore minimizing the amount of onsite work necessary and expediting rollout. "Modular" means that all the components of the data center, from core elements like compute, networking, and storage to peripheral support functions like power and cooling, were also pre-engineered to fit bespoke configurations and allow for efficient scalability. These characteristics are why GADU can be up and running within a year—about four times faster than a conventional brick-and-mortar data center.
Similar to traditional data centers, a modular data center like GADU houses servers for compute and storage, as well as servers and switches for networking. The main difference is that the equipment is placed inside standardized shipping containers, which provide a safe, self-contained environment that paves the way for extended operation in remote and rugged areas on the network edge. Additional rack-level equipment such as in-row CDUs or sidecars is required for thermal management. Power supply is also key; depending on the setup, power and cooling may be contained in separate modules. 
The biggest draw of modular data centers is how quickly they can integrate into AI infrastructure. The prefabricated components minimize site-level assembly and deliver reliable rollout up to four times faster than conventional data centers, turning the vision of plug-and-play AI into reality. The modular aspect of solutions like GADU means operators can pick configurations that neatly match actual demand and strike an optimal balance between investment and output. Last but not least, the insulated enclosure inside containers allows high-performance compute to be projected farther afield than ever before, bringing AI compute closer to data sources on the network edge—even across diverse locations and in rugged environments.
Prefabricated modular data centers are not meant to replace conventional data centers and AI factories. Rather, they offer an alternative that has the potential to rapidly scale up or scale out AI capabilities, and to establish AI infrastructure far from the network core. Ingenious operators have used GIGABYTE solutions to efficiently set up mobile data centers where there is a sudden swell in computing demand, only to disassemble and relocate once the surge has passed. Other clients erect GADU in the parking lots or on the rooftops of existing data centers to endow the facilities with new capacity or technology. At the end of the day, the brick-and-mortar data center is still the beating heart of the ongoing AI revolution—but modular data centers can expand its performance or extend its coverage.
Industries that have found success with modular data centers like GADU include cloud (public, private, and "neocloud"), telecom, science and academia, aerospace, and corporations invested in enterprise AI. So long as there is a need to get AI compute online as soon as possible, whether to answer market demand or to maintain a competitive lead, GADU can serve as the client's AI data center in a box, made to order and ready to go.