Accelerated & High Density AI

AI FACTORY TOKENS

QUC⁴ – Quattro Ai-HDi

AI FACTORY TOKENS

Trillion Token Inference & Large Language Model Development with Performance Guarantee & Limitless

Scaled Up Triune EXA-ZETA Hyperscale

Ultra Blackwell B300 - Rubin Vera

98.4 Exa = 295.2(Exa)- 2.07(Zeta) (F8+)

Bigger. Better. Faster

Limitless

Future Proofed

AI FACTORY TOKENS

Future Proofed & Limitless

Scaled Up Triune EXA-ZETA Hyperscale

Ultra Blackwell B300 - Rubin Vera

98.4 Exa = 295.2(Exa)- 2.07(Zeta) (F8+)

Bigger. Better. Faster

Future Proofed

Future Proofed & Limitless

Scaled Up Triune EXA-ZETA Hyperscale

Ultra Blackwell B300 - Rubin Vera

98.4 Exa = 295.2(Exa)- 2.07(Zeta) (F8+)

Bigger. Better. Faster

Future Proofed

Future Proofed & Limitless

Scaled Up Triune EXA-ZETA Hyperscale

Ultra Blackwell B300 - Rubin Vera

98.4 Exa = 295.2(Exa)- 2.07(Zeta) (F8+)

Bigger. Better. Faster

Global Zeta AI Factory

NXS-AI Global Factory

Among the Top 5 Largest AI Fabrics Globally

Capacity & Performance

Compute Power: ~171 zettaFLOPS
Throughput: ~3.64B tokens per second
Economic Output: ~$7.55 trillion annual revenue potential

Core Infrastructure Advantages

HVDC³ (High Voltage Direct Current Gen-3): Delivers ultra-efficient long-haul power transmission for

NXS-AI Global Factory

Among the Top 5 Largest AI Fabrics Globally

Capacity & Performance

Compute Power: ~171 zettaFLOPS
Throughput: ~3.64B tokens per second
Economic Output: ~$7.55 trillion annual revenue potential

Core Infrastructure Advantages

HVDC³ (High Voltage Direct Current Gen-3): Delivers ultra-efficient long-haul power transmission for sustainable AI operations.
Dark Fiber Networks: Provides secure, ultra-low latency connectivity across global hubs.
NVIDIA Spectrum-X: Optimized Ethernet fabric for hyperscale AI workloads, enabling unmatched scale and performance.

Global Positioning

Ranked within the Top 5 AI fabrics worldwide, setting benchmarks in compute density, operational efficiency, and economic throughput.
Engineered as a sovereign-scale digital factory, capable of hosting the most demanding AI, blockchain, quantum, and real-time simulation workloads.

Strategic Value

NXS-AI Global Factory integrates energy, compute, and network fabrics into a single optimized ecosystem—delivering the backbone for next-generation AI economies.

NOVEMS+ HVDC3+ QUC4

Global Zeta AI Factory

Powering the future of AI:

Novems + HVDC3 + QUC4,

integrated into the

Global AI Factory for exa- and zettascale outcomes.

2-Building Cluster: 160,000 sq. ft. of hyperscale-ready infrastructure (80,000 sq. ft. per building).

For Hyperscalers, Governments & Institutional Investor

Powering the future of AI:

Novems + HVDC3 + QUC4,

integrated into the

Global AI Factory for exa- and zettascale outcomes.

2-Building Cluster: 160,000 sq. ft. of hyperscale-ready infrastructure (80,000 sq. ft. per building).

For Hyperscalers, Governments & Institutional Investors: Secure, modular, scalable, and investment-ready.

Performance at Scale: Exascale and zettascale compute, advanced AI inference, simulation, and tokenized outputs.

Grid HiQ Assurance: Power certainty, operational continuity, and long-term scalability.

Grid HiQ + NXS-HUB-1 —

Where energy, compute, and intelligence converge.

QUC4 Mesh & Campus

Global Zeta AI Factory

QUC4 Mesh & Campus

Elemental Systems X

AI Fabric for Trillion-Parameter Workloads

Seamless. Scalable. Certain.

Elemental Systems X fuses NVIDIA Spectrum-X fabric with a meshed QUC4 inter-building network, delivering:

Ultra-low latency for real-time inference and training

Terabit-speed connectivity across a two-building cluster

Seven-segment campus integration for

Elemental Systems X

AI Fabric for Trillion-Parameter Workloads

Seamless. Scalable. Certain.

Elemental Systems X fuses NVIDIA Spectrum-X fabric with a meshed QUC4 inter-building network, delivering:

Ultra-low latency for real-time inference and training

Terabit-speed connectivity across a two-building cluster

Seven-segment campus integration for true hyperscale growth

Purpose-built for AI at Scale

Engineered to sustain trillion-parameter workloads with uncompromising speed and resilience

Optimized for seamless data flow, ensuring every GPU, accelerator, and node operates at full potential

Designed with power certainty and resilient architecture, eliminating bottlenecks in compute-intensive AI operations

Why It Matters ?

Elemental Systems X

transforms campus infrastructure into an AI super-fabric,

enabling enterprises,

research, and sovereign clouds to

scale faster, smarter, and

more securely than ever before.

Blackwell Ultra AI Factory Platform

About Elemental

Our Vision

At Elemental, we believe that energy management is the key to a sustainable future. Our vision is to create a world where businesses can thrive while reducing their environmental impact. We aim to be the leading provider of energy management solutions across the globe.

Our Expertise

Our team of energy management experts has decades of combined experience in the industry. We have worked with businesses of all sizes and industries, delivering customized solutions that help them reduce energy consumption and increase efficiency. Our expertise is unmatched.

1 Platform for Develop-2-Deploy NVIDIA "Blackwell Ultra AI Factory Platform"

Powered by Ultra High Voltage (UHV) & Nvidia Future Proofed.
"Blackwell Ultra to QUC⁴ to Ultra Ruben Vera to
Elementals High Density (HDi) Quattro-Ultra Clusters
Elementals Accelerated AI Ultra-Clusters - Ultra-Blackwell.
Delivering 486X's more AI High Compute Dividend than Conventional Computing.
Delivering 72X's more AI High Compute Dividend than H100 & General AI Cloud.
Delivering 27X's more AI High Compute Dividend than general Blackwell Deployment.

"Experience AI w/ Hub-1-One Accelerated AI Compute Infrastructure"

1 Platform for Develop-2-Deploy

Elementals Accelerated AI Systems powered by Grid as a Service Balqnce of Plant )BoP Grid HIQ, Macro-Grid, Grid-IX

Delivering 270X's more AI High Compute Dividend than Conventional Computing.

Delivering 91X's more AI High Compute Dividend than H100 & General AI Cloud.

Delivering 18X's more AI High Compute Dividend than general Blackwell Deployment.

Experience AI w/ Hub-1-One Accelerated AI Compute Infrastructure

NOVEMS

Ruben Vera

100 petaFLOPS of FP4 performance and cram 1 terabyte of even faster HBM4e memory.

NVLink 6 Switch performance of up to 3,600 GB/s, and a touted CX9 SuperNIC component offering up to 1,600 GB/s.

144 of these packages along with an unspecified number of Vera CPUs will be crammed into rack rated for 600 kW of power consumption and thermal output. In total, the chip giant expects the rack-scale system to deliver 15 exaFLOPS of FP4 inference perf and 5 exaFLOPS of FP8 for training.

Service

B300 HGX

Single versatile platform that can easily and efficiently do pretraining, post-training and reasoning AI inference .

NVIDIA Blackwell Ultra boosts training and test-time scaling inference — the art of applying more compute during inference to improve accuracy — to enable organizations everywhere to accelerate applications such as AI reasoning, agentic AI and physical AI.

Built on the groundbreaking Blackwell architecture introduced a year ago, Blackwell Ultra includes the NVIDIA GB300 NVL72 rack-scale solution and the NVIDIA HGX™ B300 NVL16 system. The GB300 NVL72 delivers 1.5x more AI performance than the NVIDIA GB200 NVL72, as well as increases Blackwell’s revenue opportunity by 50x for AI factories, compared with those built with NVIDIA Hopper™.

Service

HGX Platform & Software Stack

HGX Platform - Enabling organizations to leverage the best of NVIDIA AI innovation. With it, every organization can tap the full potential of their DGX infrastructure with a proven platform that includes AI workflow management, enterprise-grade cluster management, libraries that accelerate compute, storage, and network infrastructure, and system software optimized for running AI workloads.

DGX Software Stack - AI workflow management, enterprise-grade cluster management, libraries that accelerate compute, storage, and network infrastructure, and system software optimized for running AI workloads.

AI Nvidia Horsepower

Blackwell

Rubin Vera Ultra Rubin Vera

Blackwell B300 Blackwell Ultra

The Blackwell architecture introduces groundbreaking advancements for generative AI and
accelerated computing. The incorporation of a new second-generation Transformer
Engine, alongside faster and wider NVIDIA® NVLink® interconnects, propel the data center into a new era.

All Blackwell products feature two reticle-limited dies connected b

All Blackwell products feature two reticle-limited dies connected by a 10 terabytes per second (TB/s) chip-to-chip interconnect in a unified single GPU.

Blackwell B300 Blackwell Ultra

Rubin Vera Ultra Rubin Vera

Blackwell B300 Blackwell Ultra

Rubin Vera Ultra Rubin Vera

System is capable of 3.6 exaflops of FP4 inference, and 1.2 exaflops of FP8 training, at some 3.3x the performance of the GB300 NVL72.

The NVL144 has 13TBps HBM4, 75TB 'fast memory,' 260TBps NVLink 6, and 28.8TBps CX9

Grace

Rubin Vera Ultra Rubin Vera

The NVIDIA GB200 Grace™ Blackwell Superchip combines two NVIDIA Blackwell Tensor Core GPUs and a Grace CPU.

It can scale up to the GB200 NVL72—a massive 72-GPU system connected by NVIDIA® NVLink®—to deliver 30X faster real-time inference for large language models (LLMs).

High-performance NVIDIA
Scalable Coherency Fabric with
3.2 terabytes per second (TB/s)
bisection bandwidth

B200

Grace

B200

NVIDIA GB200 Superchip Incl. Two Blackwell GPUs and One
Grace CPU designed for a new type of data center.

These data centers run diverse workloads like AI, data analytics, hyperscale cloud applications, and high-performance computing (HPC). NVIDIA Grace delivers 2X the performance per watt, 2X the packaging density, and the highest memory

NVIDIA GB200 Superchip Incl. Two Blackwell GPUs and One
Grace CPU designed for a new type of data center.