Years waiting
for grid access.
Traditional data centers face interconnection queues, permitting delays and monolithic construction risk.
2–4 years typical grid/permitting friction cited in the plan
GFGRIDFACTORY
PRIVATE INVESTOR MATERIALSEnter the compute
investment room.
Leave your email to access the interactive H100/B200 infrastructure model, investment scenarios and data room.
By continuing, you agree that GridFactory may contact you regarding this investment opportunity. Projections are illustrative and are not investment advice.
GREEN-ENERGY AI INFRASTRUCTURE
Own the machines
that power AI.
GridFactory places high-density GPU clusters directly inside renewable energy parks—turning secured power into productive compute and monthly infrastructure revenue.
2Romanian energy parks secured*
H100 + B200Training-class GPU focus
400GUnified cluster fabric
*Operator-provided project status; documentary evidence available in the data room.
01 / THE INVESTMENT THESISThe bottleneck is not AI.
The bottleneck is not AI.
It is power-connected compute.
Years waiting
Plug into the park.
Not the grid.
Deploy standardized GPU modules where renewable power, land and grid infrastructure already exist.
4–8 weeks target deployment after hardware delivery
POWER → COMPUTE → REVENUE02 / EXPLORE THE MACHINE
One module. A scalable AI factory.
GRIDFACTORY / MODULAR COMPUTE 01
UPSPDUCOOL
LIVE CLUSTER VIEW1 MODULE
SELECTED HARDWARE / RACK 1
8 × H100 SXM GPUs
1 removable eight-GPU SXM server node. Racks 1–4 form the proven H100 training fabric; racks 5–8 add B200 density and premium capacity.
Energy sourceSolar + grid balancing
Shared NVMe storage384 TB
01
400G fabric
NVLink inside each node; low-latency cluster networking between nodes.
02
Integrated power + cooling
UPS, distribution, precision cooling, fire suppression and DCIM monitoring.
03
Repeatable deployment
Start with one module; add capacity without rebuilding the entire campus.
03 / INTERACTIVE INVESTMENT MODELMove the inputs.
Move the inputs.
See the economics move.
Illustrative scenario model—not a promise or guaranteed return. Pricing, utilization and equipment costs require commercial validation before investment.
ESTIMATED MONTHLY EBITDA$299,263B200 reserved-training case
Gross revenue / mo$344,678
Energy / mo$11,341
Indicative CapEx$4,629,500
Simple payback15.5 months
Compute capacity288K TFLOPS
Facility load143 kW
Your modeled infrastructure share21.6%
Pro-rata monthly EBITDA$64,643
Before financing terms, reserves, taxes and investor-structure adjustments.
MODELED INVESTOR VALUE AT YEAR 5$29,465,38429.5× modeled total return
Projected platform ARR$24,443,791
Modeled enterprise value$171,106,539
Your diluted exit proceeds$27,720,035
Cash distributions retained$1,745,349
Y1
Y2
Y3
Y4
Y5
This combines operating distributions with a user-selected revenue multiple for the software-enabled CaaS platform. It is an illustrative exit scenario—not a valuation, liquidity promise or guaranteed return.
04 / TWO COMMERCIAL CASES
Designed for high-value utilization.
CASE A / PROOF + DEMAND
H100 reserved
training cluster
Bankable first deployment logic: mature hardware ecosystem, 12–24 month reserved contracts, 400G fabric and sovereign/enterprise training demand.
- Lower equipment entry cost
- Base rate modeled at $3.90/GPU-hour
- Operator target: first invoices container by container
CASE B / MAXIMUM DENSITY
B200 premium
training fabric
Higher compute density and premium reserved rates can accelerate revenue per rack—subject to procurement price, cooling design and contracted demand.
- 4.5 PFLOPS FP8 dense per GPU reference
- Base rate modeled at $8.50/GPU-hour
- Fewer campuses needed for the long-range ARR ambition
05 / CAPITAL ARCHITECTUREOne physical asset.
One physical asset.
Multiple ways to participate.
Equity
Ownership in GridFactory platform growth and multi-park expansion.
Project debt
Defined coupon, maturity and hardware/project security package.
Revenue share
Contractual share of distributable CaaS revenue from selected capacity.
Infrastructure-backed
Capital tied to defined GPUs, servers or modules with lease economics.
Tokenised GRID units
Digitally represented compute participation, subject to UAE/EU structuring.
Own-to-lease
Investor-funded equipment leased into the operating cluster.
06 / DEPLOYMENT PATHContainer 01.
Container 01.
Campus 01.
Multi-park network.
- NOW
Two Romanian park opportunities secured by the operator; commercial and technical diligence in progress.
- NEXT
Foundation module + first revenue-producing cluster.
- SCALE
Shared campus infrastructure, additional modules and UAE park development.
07 / INVESTOR DATA ROOM
Audit the thesis.
Review the current narrative documents. The interactive model is the working scenario layer; final investment materials will include supplier quotations, energy agreements, customer pipeline evidence and legal structuring.