Colocation & Offtakers

Power-Ready AI Capacity

An energy solution for colocation and AI data-centre campuses that need capacity which is actually powered — not a shell waiting years for a grid connection. GRECO campuses build their own power: generation, cooling and storage engineered in from day one, so the limiting factor on your roadmap stops being the utility queue.

Two offtakes on one campus — liquid-cooled compute capacity for AI operators, and recovered district heat for neighbourhood energy utilities. Anchored by our flagship Morrison Park campus in the Vancouver area, with expansion phases across British Columbia and Alberta.

Years
Faster to power
150kW→1MW
Rack density ready
24/7
Baseload + storage
4
Offtake streams

Capacity, Without the Queue

You offtake capacity, not construction risk — and you get powered years sooner.

Loads that cannot wait for a utility connection are exactly the ones BC Hydro's call for demand was created to move.

Time to power

years faster
Grid interconnection queue5–7 years
Greco energy-firstmonths
energized — your own power plant arrives with the campus

Across Canada, utility interconnection for large new loads can stretch five to seven years. Greco generates power on-site, behind the meter — so a campus can energize years before a new substation arrives.

Four Streams You Can Offtake

Every unit of energy on a Greco campus is engineered to work more than once — and the compute itself arrives NVIDIA-accelerated via the Dell AI Factory, or as a powered shell you fit out yourself.

AI Compute Capacity

Lease powered, liquid-cooled rack capacity in a campus that arrives with its own generation, or take a powered shell and fit out the interior yourself. Either way, no interconnection queue between you and go-live.

Recovered Heat

Low-carbon thermal energy captured at 50 °C, suitable for fourth-generation district energy — a revenue stream, not a cost.

Behind-the-Meter Power

Firm, dispatchable power from on-site generation and storage — useful to co-located industrial, fleet, or grid-service offtakers.

Megawatt EV Charging

Co-located Megawatt Charging System capacity for commercial fleets, powered by the hub and enabled for vehicle-to-grid.

Built for the Density Curve

Powered for today’s racks, architected for what’s next — see what each density class means for power, heat and footprint.

Size your capacity

Shipping today
Target IT load40 MW
5 MW · urban edge200 MW · metro campus
Rack density class
todayNVIDIA Kyber / Rubin roadmap →
Racks for this campus
267

40 MW at 150 kW per rack

Power per rack
150 kW

IT load, fully liquid-cooled

Recovered heat
≈ 34 MW

captured at 50 °C, exportable

Cooling regime
Direct-to-chip liquid

engineered into the energy plant

Rack-power and cooling figures track NVIDIA’s published reference designs; campuses are engineered to be ready for them, not claimed as installed today.

Map this to a Greco site

Reliability by Design

On-Site Generation

Fuel-flexible solid-oxide fuel cells and linear generators deliver 24/7 dispatchable baseload behind the meter — independent of grid congestion.

Grid-Scale Storage

Co-located lithium and vanadium-flow batteries ride through grid events and firm the load curve — keeping compute online during disturbances.

Grid Where It Helps

Where clean utility power is available — BC and Quebec exceed 95% hydroelectric — we use it, and generate on-site where it isn’t. Resilience by design, not dependency.

Urban Edge or Metro Campus

Choose proximity to match the workload: urban edge sites inside the city for 1–5 ms inference latency, or larger metro campuses for training and bulk compute. Both can export recovered heat into the surrounding community.

Compare urban & metro

De-Risked Before You Sign

Before you commit to capacity, you can see it. We model each campus — power, cooling, storage and the heat-export loop — in a physics-based digital twin, so power and thermal behaviour are validated at the design stage. Fewer surprises, firmer dates, and a facility you can walk through virtually before it’s built.

From compute to community heat

The heat produced by AI computation, captured with liquid cooling and delivered into municipal district energy — with no combustion at the point of use.

Read the transcript

Every city on Earth faces the same silent crisis. We generate enormous amounts of energy and we waste most of it. The heat escaping from our machines, our buildings, our infrastructure, lost to the atmosphere, doing nothing, helping no one.

But what if waste could become resource? What if every unit of energy could create not one form of value but many? This is the question GRECO Energy has answered, and the answer is changing everything.

The AI data centre. The world sees it as a machine that consumes electricity, water and land — a necessary cost of the digital age. But GRECO Energy sees something different. They see a power plant, a heat source, a piece of public infrastructure that can do far more than compute.

Inside every data centre, AI generates something no one talks about: thermal energy. Enormous quantities of heat produced constantly, relentlessly, as a direct byproduct of intelligence at work. For decades that heat has been thrown away.

GRECO Energy has built a system to catch it. Advanced liquid cooling pulls heat directly from the processors — efficiently, precisely, at scale. Where air cooling wastes energy fighting heat, liquid cooling captures it.

That captured thermal energy enters a recovery loop. It does not disappear; it becomes a resource. Heated fluid moves through insulated pipe networks — the same principle as the district energy systems that have heated European cities for generations, now reimagined for the age of artificial intelligence.

Recovered thermal energy flows into municipal district energy systems: networks of insulated pipes that carry heat directly to homes, hospitals, schools and commercial buildings. No boilers, no burning, no emissions at the point of use. Just intelligent heat, recovered from intelligence itself, warming the places where people live and work and heal.

This is not a pilot project. This is scalable infrastructure, designed to integrate with existing municipal utility systems and grow with the cities it serves.

Capacity questions, answered

Months, not years. Because Greco generates power on-site, behind the meter, a campus can energize before a new substation arrives — sidestepping the five-to-seven-year utility interconnection queue that strands grid-dependent projects in Canada.

Roughly 150 kW racks ship today and our campuses are designed around them. We engineer for the next steps — 600 kW and toward 1 MW racks, with 800 VDC distribution — as that hardware reaches the market on NVIDIA’s published reference roadmap (Kyber / Rubin-class, ~2027).

Yes. Compute heat is captured at 50 °C and can be exported to a neighbourhood energy utility through an Energy Transfer Station — turning what every other operator vents into a revenue stream and a community benefit.

It means generating electricity on-site, on your side of the utility meter, instead of waiting for grid capacity. Fuel-flexible solid-oxide fuel cells and linear generators provide 24/7 dispatchable baseload, firmed by lithium and vanadium-flow battery storage.

Urban edge sites sit inside the city for 1–5 ms inference latency; metro campuses are larger and suited to training and bulk compute. Both can export recovered heat to the surrounding community.

Because the grid can no longer keep up. New utility interconnections for large loads now take five to seven years in much of Canada, while AI capacity is needed now. Building generation on-site — behind the meter — lets a campus energize on the developer's timeline instead of the utility's queue, and lets the load flex to support the grid rather than strain it.

For campuses of this scale the answer is rarely a single utility line. GRECO develops the full stack — behind-the-meter generation, grid-scale storage, liquid cooling and, where clean utility power is available, a grid interconnection — coordinated with the local utility (e.g. BC Hydro) and gas supply. You offtake finished, powered capacity from one counterparty instead of assembling generation, storage, cooling and interconnection yourself.

Talk to us about capacity

Tell us your power, density and latency requirements and we’ll map them to a Greco site.