Cost & Financial Control

Where the $100 million goes

Eighty-five million dollars have been allocated across nine categories, with fifteen million held in reserve.

01 — Allocation

$85M planned, $15M in reserve

The contingency reserve is tied to modular system upgrades on a 36-month cycle. This is because high end computing hardware depreciates fast and will need to be replaced.

$100,000,000

Total federal infrastructure grant, in Canadian dollars.

  • Planned expenditure $85.00M
  • Contingency reserve $15.00M

Planned expenditure by category

Nine categories, ordered by size. Bars are proportional to the largest allocation.
Category Relative size Budget Share
Building construction & site development $22.00M 22.0%
HPC / GPU / computing infrastructure $17.00M 17.0%
Land acquisition: 28310 Huntingdon Rd (151.31 acres) $16.85M 16.85%
Research laboratories & specialized equipment $10.00M 10.0%
Makerspace, robotics & prototyping $6.00M 6.0%
Research & student innovation fund $4.00M 4.0%
Employee & training $3.15M 3.15%
Sustainability, power & cooling infrastructure $3.00M 3.0%
Architecture, engineering & project management $3.00M 3.0%
Planned project expenditure $85.00M 85.0%
Contingency reserve $15.00M 15.0%
Total budget $100.00M 100%

02 — Cost Justification

Where the larger figures come from

Five categories account for $68.85M of the $85M planned, plus the $15m Contingency reserve accounts for 83.85% of project budget.

$22M

Building construction & site development

According to Precedent Developments' 2026 Canadian Construction Costing Guide [4], university teaching buildings in Vancouver are estimated at approximately $750-$1,250 per square foot but no info regarding Abbotsford. Hence this source provides a basis for preliminary planning and the final cost will depend on the facility's size, design, and site requirements.

Source: Altus Group, 2026 Canadian Construction Costing Guide (reference 4).

$17M

HPC / GPU / computing infrastructure

According to Exxact [6] Corporation (2025), NVIDIA H200 NVL GPUs cost over US$25,000 each, while an eight GPU server can cost more than US$200,000. These figures demonstrate the significant investment required to establish AI computing capabilities and support our proposed budget allocation.

Source: Exxact Corporation (2025) (reference 6).

$16.85M

Land acquisition

For land acquisition, Group 8 proposes purchasing the property located at 28310 Huntingdon Road, Abbotsford, BC V4X 1A3.


The property consists of two parcels totaling approximately 151.31 acres with the asking price of $16.85 million CAD (This is more than double the size of current UFV Abbotsford campus). Located approximately a 15 minute drive from UFV, this land offers more than enough space for the proposed AI research and innovation facility.

Source: Fraser Valley MLS listing.

$10M

Research laboratories & specialized equipment

The University of Toronto's Guide to Construction [5] distinguishes building construction from equipment and other project expenses. This supports our approach of budgeting laboratory equipment separately from construction costs.

Source: University of Toronto, A Practical Guide to Addressing Construction & Renovation (2024) (reference 5).

$3M

Sustainability, power & cooling

An eight GPU H200 server can draw approximately 5-7 kW, making substantial electricity and heat management requirements for AI computing. This supports the need to budget for appropriate electrical capacity and efficient cooling systems.

Source: Exxact Corporation (2025) (reference 6).

$15M

Contingency reserve

Held back rather than allocated, and tied directly to the cost of modular systems. It is drawn on for upgrades at a frequency of roughly 36 months, which is the practical response to rapid hardware obsolescence on a project that will not open its doors until 2030.

See risk R-03, rapid hardware obsolescence.