As a NVIDIA Mobile AI Factory strategic partner, ECOBLOX’s product expansion focuses on the high-performance computing needs of AI applications. All of ECOBLOX products for NVIDIA AI factories are based on the industry leading new NVIDIA DGX Rubin systems and NVIDIA Blackwell Architecture.

ECOBLOX AI Supercomputer with up to 576 NVIDIA Rubin GPUs

Turnkey 40-foot MDC With 576 NVIDIA DGX Vera Rubin NVL72 Direct Liquid Cooled (DLC) Token Factory System

The new ECOBLOX AI Supercomputer integrates NVIDIA’s latest advances in compute, networking, and software to power next-generation AI workloads. Designed to accelerate agentic AI, MoE models, and long-context reasoning, it can deliver up to a 10× reduction in inference token cost compared to ECOBLOX MDC systems using NVIDIA Blackwell GPUs.

The ECOBLOX AI Supercomputer (Tier II is standard, Tier III optional) is equipped with up to 8 racks containing 576 NVIDIA Vera Rubin NVL72 with a DLC system and all UPS and supporting systems. Additional four network and storage racks and two support racks include servers for management and storage. The turnkey solution also contains a fire suppression system, security system, the ECOBLOX DCIM management platform, installation, setup and warranty.

Powered by the NVIDIA Vera Rubin platform, the ECOBLOX AI Supercomputer delivers up to 3.6 EFLOPS FP4 per rack and 28.8 EFLOPS per system—providing 5× the performance of previous NVIDIA GB200-based ECOBLOX AI MDC generations. For AI training, it achieves up to 20.16 EFLOPS, representing a 3.5× performance increase over GB200-based systems.

ECOBLOX AI Modular Data Center (MDC) with up to 720 NVIDIA Blackwell GPUs

Turnkey 40-foot MDC With 360 NVIDIA Blackwell Direct Liquid Cooled (DLC) Servers

This ECOBLOX MDC (Tier II is standard, Tier III optional) is equipped with up to 10 racks containing 72 NVIDIA Blackwell GPU compute servers with a DLC system and all UPS and supporting systems. An additional rack cabinet includes servers for management and storage. The turnkey solution also contains a fire suppression system, security system, the ECOBLOX DCIM management platform, installation, setup and warranty.

This 40-foot MDC architecture natively supports the current GB200 and NVIDIA Blackwell Ultra (GB300) and AMD Mi350/Mi400 generation (available in 2025-2026).

ECOBLOX AI Modular Data Center (MDC) with 512 NVIDIA H200 GPUs

Turnkey 40-foot MDC with NVIDIA H200 Air Cooled Servers and RDHx Racks

The ECOBLOX 40-foot Modular Data Center (MDC) is a turnkey solution with 64 NVIDIA H200 air cooled GPU server nodes (512 GPUs) in rear door heat exchanger (RDHx) racks. This complete solution also contains a high-speed storage network for management and data transfer, an out-of-band management network, a fire suppression system, security system, the ECOBLOX DCIM management platform, installation, setup and 3-year warranty.

ECOBLOX AI Rack Data Center (RDC)

A Complete AI/HPC Data Center in a Single Rack

The ECOBLOX RDC is the ultimate in performance for new AI/HPC data centers or for retrofitting an existing in-building data center. The turnkey solution includes four to eight NVIDIA H200 HGX high density compute racks (32-64 GPUs), additional management and storage servers, and high-speed storage and GPU compute networks. The single wide rack features a rear door heat exchanger (RDHx) requiring an external, free-standing air-cooled chiller or integration with an existing cooling system.

ECOBLOX AI Servers

ECOBLOX offers a variety of configuration options. All configurations come with up to 5 years of warranty. Options that include a liquid cooling system are available with up to 8 years of warranty. All options include a fully featured AI Factory Suite and a 5-year NVIDIA Enterprise Standard suite license with an optional upgrade to Enterprise Premium.

● AI Inference Server - Entry Level
● AI Inference Server - Mid-range
● AI Inference Server - High Performance
● High Performance AI Inference and Training Server - Entry Level
● High Performance AI Inference and Training Server - Mid-range
● High Performance AI Inference and Training Server - High End
● High Performance AI Inference and Training Server - Pinnacle

News and Events

News

ECOBLOX Wins NVIDIA Star Performer Award

The acceleration of AI adoption is redefining global infrastructure and partnerships are playing a critical role in scaling what’s possible.…

News

Stratos Lab, ECOBLOX and Digital Parks Africa Partner to Launch Most Powerful African AI Cloud

- Groundbreaking Collaboration Establishes a New Benchmark for Sovereign AI Compute in Africa Powered by 50+ NVIDIA B300 Servers (400+…

News

ECOBLOX and Hamilton Barnes Partner to Scale Global AI Infrastructure Projects

-- Partnership Provides Integrated End-to-end Framework to Help Companies Execute AI Projects Faster at Every Stage -- Dubai, UAE and…

Events

ACTIVATE South Africa, Cape Town, August 26 - 27, 2026

ECOBLOX is at ACTIVATE South Africa August 26–27, 2026 | CTICC Cape Town, South Africa ACTIVATE South Africa is the…

Get started with ECOBLOX AI/HPC data center solutions