Coconut, oil palm, sugarcane, cassava and rice grow across Thailand and Southeast Asia, with harvests running all year round.
02 / 06 · Collect
Residue becomes fuel
Husks, shells and crop residue, usually dumped or burned in the open, are collected as feedstock.
03 / 06 · Generate
Firm power on site
A direct-combustion biomass plant with a boiler and steam turbine produces dispatchable 24/7 electricity behind the meter.
04 / 06 · Compute
Power becomes intelligence
Co-located, liquid-cooled GPU modules turn that electricity into AI training and inference capacity.
05 / 06 · Recover
Heat goes back to work
Recovered heat dries incoming fuel and can drive absorption cooling.
06 / 06 · Return
Back to the land
Mineral-rich ash, and in future biochar, returns to farms as a soil amendment.
24/7
Firm, dispatchable on-site generation
~1.2
Design power usage effectiveness (PUE)
12mo
Tropical harvest cycle supplying residue
0km
Transmission distance, generation to compute
Design targets and site characteristics. Not operating results.
03Year-round fuel
In the tropics, the harvest never stops.
Perennial crops such as coconut and oil palm yield continuously, while sugarcane, cassava and rice add seasonal peaks. Drawing on several crops evens out residue supply across the year.
CoconutOil palmSugarcaneRice straw and huskCassava
Residue availability by month (illustrative)
J
F
M
A
M
J
J
A
S
O
N
D
04The opportunity
Southeast Asia's AI build-out has a power problem.
Demand for AI compute across Thailand and Southeast Asia is rising fast, and far more data center capacity is planned than is operating. The constraint is firm power delivered on time. Sites that bring their own power can move first.
Conceptual render of an integrated power-plus-compute site.
A region going digital
Cloud, AI and digital services are drawing data center investment to Thailand and its neighbours.
Power is the bottleneck
Grid connections and new capacity take years, while AI loads are large, steady and needed now.
Our model removes the usual bottlenecks between demand and delivered AI capacity.
Power-first, behind the meter
We secure generation first and connect compute directly, with no wait in the grid interconnection queue.c3
Modular containers
Factory-built compute modules deploy in months, not the years a conventional build takes.c3
Year-round fuel
Tropical agriculture supplies residue in every month, so supply does not depend on one season.c3
Phased PPA model
Start on contracted power, then build our own generation, so capacity grows step by step with demand.c3
What customers get
Fast capacity. Dedicated GPU capacity delivered on a modular timeline.
Green proof. Energy-source reporting through Proof of Green Compute.
Regional latency. In-country infrastructure close to users across Thailand and Southeast Asia.
06Mission
Turn the region's harvest into its AI advantage.
World-class AI infrastructure, built in Southeast Asia on clean power the region grows itself, with value flowing back to the farming communities that supply it.
Spiral AI Energy
