SURPLUS ENERGY-BASED
SMALL MODULAR DATA CENTER (S-SMDC)

Building the Future of Energy-to-Compute Infrastructure: Turning Surplus Renewables into Intelligent Compute Assets.

Proposed Architecture for a Carbon-Neutral Computing Center Powered by Renewable and Surplus Energy

INFRASTRUCTURE OPTIMIZED FOR: NVIDIA H100 (AI Training)BITMAIN Antminer S19 (ASIC)DataBean (Immersion Cooling)Schneider Electric (Power)Ubuntu Linux

A Global Energy Bottleneck

Latest curtailment data from major markets highlighting market urgency

South Korea

The Pilot Market
Wasted Energy:
Expanding to Mainland
Curtailment:
Jeju (18%) → Jeonnam Spread
Details:

Curtailment is no longer limited to Jeju Island. Mandatory output limitations have begun in the Jeonnam region, threatening IPP profitability.

USA (Texas/ERCOT)

Highest Profitability
Wasted Energy:
High Frequency of Negative Prices
Curtailment:
Bottled-up Wind Energy
Details:

West Texas wind farms frequently face negative wholesale prices due to transmission congestion. Ideal environment for Paid-to-Consume business models.

Europe (Germany)

Regulatory Pressure
Wasted Energy:
€3 Billion+ Redispatch Cost
Curtailment:
Ghost Electricity (Geisterstrom)
Details:

Grid operators spend billions annually compensating generators to turn OFF turbines. S-SMDC solves this Redispatch cost burden.

China

Unlimited Scale
Wasted Energy:
100 TWh+ Stranded Power
Curtailment:
West-East Grid Bottleneck
Details:

Massive renewable capacity in Gansu and Xinjiang exceeds UHV transmission limits. Perfect for large-scale, off-grid AI computing clusters.

The S-SMDC Solution

Flexible Load Management: Putting Power In and Taking It Out to Stabilize the Grid

2~3MW Modular Pod Architecture

Component A: Immersion Cooling System

Fire-free immersion cooling technology achieving PUE 1.05. Eliminates battery fire risk completely.

Component B: Modular Scalability

20ft/40ft container design enabling instant add/remove based on grid conditions. Rapid deployment capability.

Component C: Plug-and-Play Interface

Direct connection to 22.9kV high-voltage lines without substations. Immediate grid integration.

Energy-to-Value Flow Process

Step 1

Surplus Energy Ingest

Instantly absorbs excess electricity when renewable energy curtailment occurs.

Step 2

Adaptive AI-EMS Control

EMS increases computing load in real-time to consume power (Load Following).

Step 3

Value Generation

Converts absorbed energy into ASIC mining, AI computing, and edge computing instantly.

Step 4

Grid Stabilization

Reduces grid load to prevent blackouts and optimizes transmission efficiency.

Flexible Load Management: The Core Innovation

Our solution dynamically puts power in when surplus energy is available and takes it out when grid demand rises. This flexible computing load management stabilizes the grid while maximizing value generation.

Dual-Mode Operation: Beyond Passive Waiting

Mode A

Stable Operations (via PPA)

24/7 Carbon-Free Compute

Securing 100% carbon-free power supply through renewable energy PPA contracts. Continuous operation guarantee.

Core Workload

High-value AI model training, LLM inference, and edge computing services. Premium margin operations.

Mode B

Surplus Optimization (via Curtailment)

Instant Energy Absorption

Immediately absorbing wasted electricity during curtailment events at low-cost or zero-cost rates.

Flexible Workload

ASIC-based digital asset mining operations to maximize additional revenue. Profit alpha generation.

Smart Switch Flow: AI-EMS Decision Making

PPA Renewable Energy

24/7 Stable Supply

Surplus/Wasted Energy

During Curtailment

AI-EMS

Smart Switch

AI/Edge Computing

High Margin

ASIC Mining

Profit Alpha

Adaptive Load Balancing

Adjusting computing load by putting power in and taking it out (Flex-load) based on power supply conditions. Maintaining 100% uptime while contributing to grid stabilization.

Economic Edge

Building a stable foundation through PPA contracts, while capturing profit alpha from surplus power. Dual-mode operation maximizes both reliability and profitability.

How It Works: Dual-Mode Operation

24/7 Uninterrupted Operation with Dynamic Profit Optimization

Mode A

Surplus Optimization

Max Profit

Capturing stranded energy for maximum-margin computing. Optimized operations that convert surplus renewable power into high-value digital assets with market-leading efficiency.

Energy cost: Near-zero marginal cost

Mode B

Standard Grid Efficiency

Stable Revenue

Providing stable, high-availability AI services via grid-connected operations. Market-leading immersion cooling technology (DataBean) delivers optimized OPEX efficiency for continuous computing operations.

OPEX reduction: 30%+ optimized savings

Flexible Switching

S-SMDC seamlessly switches between both modes based on grid conditions, ensuring continuous operation while optimizing profitability in real-time. No downtime, maximum efficiency.

Revenue Engine

Four Core Engines: S-SMDC Generates Revenue as a Grid Partner, Not Just a Consumer

Engine 1

AI/GPU Hypercomputing

High-Margin Infrastructure

24/7 high-performance computing for LLM training and AI inference.

Powered by DataBean's immersion cooling for industry-leading OPEX efficiency.

Engine 2

ASIC Digital Asset Mining

Surplus-Driven Arbitrage

Generating high-margin revenue by utilizing surplus renewable energy during curtailment events.

A flexible energy buffer that maximizes profitability with near-zero marginal power costs.

Engine 3

Global DR & Ancillary Services

Paid-to-Consume Model

Earn Rewards for Grid Stabilization (FFR/DR).

Monetizing Grid Instability.

Engine 4

Hyper-Local Edge AI

Ultra-Low Latency

Real-time Data Processing at Source.

Smart City & Autonomous Grid Control.

Grid Partner, Not Just a Consumer

S-SMDC generates revenue as an active partner in grid stabilization. Through Demand Response and Ancillary Services, we receive compensation for consuming power when the grid needs it, creating value for both the energy system and our operations.

S-SMDC vs. Traditional ESS

Why S-SMDC is the superior economic alternative

Profit Model

S-SMDC

Active Computing Revenue

Traditional ESS

Passive Cost Saving

ROI Period

S-SMDC

14-24 Months

Traditional ESS

7-10 Years

Safety

S-SMDC

Immersion Cooling (Fire-free)

Traditional ESS

Battery Fire Risk

Global Roadmap

Building the future of energy-to-compute infrastructure, one phase at a time

1
Phase 12026: Foundation & Seed Funding
Jeju & Jeonnam

Prototype Design, Site Securing (Jeju), Seed Round Fundraising. Building the foundation for the first S-SMDC deployment.

We Are HereCurrent Status
2
Phase 2Global Growth
Texas ERCOT & Germany

Entering the world's largest arbitrage markets. US ERCOT and European power grids expansion.

Planning
3
Phase 3Global Backbone
SE Asia & Middle East

Connecting desert solar farms to the grid. Edge computing supply in infrastructure-deficient regions.

Future