Distributed AI Infrastructure · Calgary + Tulsa

Let's build the
inference edge.

RyuGen turns underused commercial power into distributed AI infrastructure — high-density NVIDIA compute deployed in existing buildings, without waiting years for new grid capacity.

2–4 mo
Deployment, not years
90%+
GPU heat recovered
NVIDIA
Certified systems
01 — The Problem

The bottleneck isn't GPUs. It's power.

AI demand is growing faster than the grid can support it. New data centers wait years for power. Existing commercial buildings already have it.

Traditional
NEW DATA CENTER GRID INTERCONNECTION 5–10 YEARS COMPUTE
RyuGen
EXISTING BUILDING + POWER HELIX 2–4 MONTHS COMPUTE

RyuGen puts that power to work. We deploy where the electricity already exists — behind the meter, ahead of the queue.

02 — What We Do

We turn buildings into AI infrastructure.

RyuGen installs modular AI compute in commercial buildings with available power. We monetize the GPUs for inference — and capture the heat they produce, delivering it back to the building.

EXISTING BUILDINGCommercial real estate with spare electrical capacity
AVAILABLE POWERBehind the meter — no interconnection queue
HELIX COMPUTEModular, liquid-cooled NVIDIA AI infrastructure
AI INFERENCECompute revenue from the network
USEFUL HEATCaptured by HeatSync™ — returned to the building
One building · one power connection · two monetizable outputs

Compute produces heat. We don't throw it away. 90%+ of GPU power becomes recoverable thermal energy — the same purchased electricity creates two value streams.

03 — The Technology

One platform. Three technologies.

Helix

The compute.

Modular, liquid-cooled AI infrastructure designed to deploy inside existing commercial buildings.

32 Blackwell GPUs · 25 kW · one modular node
HeatSync™

The thermal system.

Captures 90%+ of GPU heat and converts a cooling liability into useful thermal energy.

Warm-water loop · thermal storage · dispatchable heat
RyuGen Command

The intelligence.

Coordinates compute, power and thermal resources across the distributed network.

Scheduling · thermal control · power balancing

Helix computes. HeatSync manages heat. Command connects the network.

Explore Helix specifications
4U · 19" COMMERCIAL-AISLE CHASSIS NET · BlueField-3 · 100GbE PWR 480V · COOLANT MANIFOLD SERVER 1 · 8× RTX PRO 6000 SERVER 2 · 8× RTX PRO 6000 SERVER 3 · 8× RTX PRO 6000 SERVER 4 · 8× RTX PRO 6000 40°C SUPPLY 50°C RETURN DIRECT-TO-CHIP LIQUID LOOP
GPUs / node32 × RTX PRO 6000 Blackwell
Form factor4U · 19" commercial-aisle
CoolingDirect liquid-to-chip · ASHRAE W4
Coolant loop40 °C supply / 50 °C return
Heat captured90%+ to liquid loop
Ambient envelope−30 °C to +35 °C
Acoustic< 65 dBA @ 1 m
Power input480 V 3-phase · behind-the-meter
CertificationNVIDIA-Certified server platforms
04 — Why RyuGen Wins

Compute without the data center.

/ 01

Months, not years

Deploy behind existing utility meters — no interconnection queue, no greenfield construction.

/ 02

More compute per building

Liquid cooling handles densities conventional HVAC cannot.

/ 03

Heat becomes useful energy

Recover 90%+ of GPU heat instead of rejecting it to the environment.

/ 04

Repeatable infrastructure

Standardized nodes scale across buildings, markets and cities.

The hard part isn't putting GPUs in a building. It's getting the heat out. Commercial HVAC was never designed for continuous high-density compute, and bolt-on cooling breaks the economics. HeatSync solves the thermal constraint that keeps commercial buildings from becoming AI infrastructure.

90%+ heat captured PUE 1.10–1.15 HeatSync™ patent pending
05 — Why Now

AI has a power problem.

74 GW
Projected U.S. data-center demand by 2028
5–10 yrs
Typical grid interconnection timeline
2–4 mo
Target RyuGen deployment

The grid can't build fast enough. We don't wait for it.

06 — The Network

From one building to a distributed AI network.

Building 01
Building 02
Building 03
Building 04
RYUGEN COMMAND One control plane across every node

Deploy locally. Operate as one network. Each building is a node; together they form dispatchable metro-edge inference capacity.

07 — Team

Built by energy infrastructure people.

Between us: midstream gas operations, energy-trading and risk systems, pipeline and thermal engineering, applied data science, and a decade in capital markets and law — the combination this thesis demands.

Sean Brown

Sean Brown

CEO & Co-Founder

Energy engineer, data scientist and HeatSync™ inventor. 18+ years across energy infrastructure, engineering and applied data science.

Martin S. Hui

Martin S. Hui

Co-Founder

Nearly a decade at Morgan Stanley, Davis Polk and BMO, then founding-team member of a B2B digital-asset startup scaled from $1M to $140M in contract value.

Tristan Walker

Tristan Walker

Operations

Renewable-energy developer and two-time founder with eight years of project-development experience.

Samuel Duperré

Samuel Duperré

Compute

Energy-trading analytics and risk across gas, power, crude and FX — commodity-markets discipline applied to inference capacity.

Hadi Bagherzadeh

Hadi Bagherzadeh

Technology

PhD researcher at the University of Calgary and reservoir engineer with thermal-process expertise.

Josh Harris

Josh Harris

Engineering

Mechanical engineer and pipeline project engineer at Venture Global LNG — field ops, hydraulics and project development.

Backed & advised by
Avatar InnovationsExergy Solutions
The path PILOT · 1 NODE · 25 kW 125 kW COMMERCIAL SITE MULTI-SITE NETWORK
08 — Partner With Us

One network. Value on both sides.

For Property Owners

Turn unused power and space into infrastructure revenue.

  • No data-center development
  • Minimal building disruption
  • Useful thermal energy returned to the property
For AI Customers

Get inference capacity where you need it, faster.

  • Modern NVIDIA GPUs
  • Distributed metro-edge capacity
  • No multi-year construction cycle
For Infrastructure Partners

Deploy a repeatable 125 kW building block.

  • Standardized architecture
  • Behind-the-meter power
  • Scalable across markets

Have power? Let's put it to work.

Property owners · Compute partners · Infrastructure investors