Canada’s AI Compute Landscape🍁

What this page is

This is the methodology backbone for the Zeever AI Compute Index — a public, auditable inventory of GPU providers serving the Canadian market. It documents three things:

1. What’s tracked — the 39 vendors in scope and what data fields are recorded
2. How vendors are ranked — the index formula, the sovereignty taxonomy, and the eligibility rules
3. What’s excluded and why — every vendor that doesn’t appear in the cost ranking, with the specific reason

If you find an error or want to add a vendor, contact us at info [at] zeever [dot] ca. Every claim on this page is sourced to public vendor pages, regulatory filings, or named industry audits — listed at the bottom.

Why this index exists

Canadian organizations choosing where to run AI workloads face a fragmented market. There is no single comparison surface for Canadian GPU pricing. Vendor websites list rates that often don’t reflect Canadian regional availability. Sovereign Canadian providers don’t publish list prices at all. Hyperscalers list flagship GPUs in Canadian catalogs that, in practice, may not be deliverable.

The Zeever AI Compute Index is an attempt to fix that — by maintaining a verified inventory, defining a transparent normalization formula, and publishing the methodology rather than just the conclusions.

The index is opinionated about one thing: published, verified pricing matters. Vendors that publish list rates appear in the cost ranking. Vendors that operate on quote-only pricing are tracked separately and flagged as “Opaque” — not because the vendor is hiding something, but because their pricing model genuinely doesn’t translate to a per-hour comparison.

Scope and inclusion

A vendor is in scope for the inventory if it meets at least one of:

• Sells GPU compute to Canadian organizations, regardless of vendor headquarters
• Operates a Canadian data centre region offering GPU instances
• Maintains a Canadian sovereign AI capacity (defined below), even if not commercially priced

The inventory currently tracks 39 vendors across these categories:

(Some vendors appear in more than one category — Cohere is both sovereign-partial and operates via CoreWeave-managed Cambridge ON capacity, for example.)

The index formula

For vendors with verified, published H100 USD per GPU-hour rates in a Canadian region:

index = vendor_h100_usd_per_hour ÷ ceiling_h100_usd_per_hour

Where ceiling_h100_usd_per_hour = $7.50 — Microsoft Azure Canada Central, the most expensive published rate at time of compilation.

Result: lower index = cheaper. Range 0.18 (Thunder Compute, $1.38) to 1.00 (Azure CA, $7.50).

For vendors that bill per token (Cohere) or only sell quote-based capacity (TELUS, Bell, BUZZ, Hypertec, Consensus Core, CoEvo, ThinkOn), the index is “Opaque” — not zero, not estimated. The index records what’s verifiable rather than fabricating a comparison number.

Why H100 specifically

The H100 is the de facto current-generation training and high-end inference GPU as of Q2 2026. Most serious AI workloads target it directly or compare against it. Normalizing to a single GPU SKU eliminates per-instance configuration variance (vCPU, memory, storage) that would otherwise muddy the comparison.

H200 and B200 rates are recorded in the inventory where published, but the index uses H100 as the anchor.

Currency and conversion

All index values are computed in USD. ISAIC publishes CAD $2.50/hour, which is converted at approximately 1.37 CAD/USD ≈ USD $1.83/hour. The conversion rate is recorded with the row and updated quarterly.

Sovereignty taxonomy

The “Tier” badge in the vendor table classifies each provider on three axes:

• Canadian-owned (Yes / No / Unknown)
• Canada-hosted (Yes / Partial / Marketplace / No / Unknown)
• US jurisdiction risk (Yes / No / Partial — based on whether the operating entity is US-headquartered and therefore subject to the US CLOUD Act)

The combination produces six tiers:

Sovereignty is a procurement criterion, not a value judgment. A US hyperscaler with Canadian data centres is appropriate for many workloads; a sovereign Canadian provider is appropriate for others. The taxonomy makes the trade-off legible.

Excluded vendors — bucket-by-bucket reasoning

The Index ranks 14 vendors with published H100 pricing and surfaces 4 sovereign anchors as “Opaque” examples. The other 21 vendors in the inventory are excluded from the ranking for transparent reasons.

Bucket 1: Other sovereign Canadian providers without published list pricing (6 vendors)

These are tracked as sovereign Canadian capacity but do not publish per-hour rates:

The four sovereign anchors shown in the public ranking (TELUS, Bell AI Fabric, BUZZ HPC, Hypertec Cloud) are representative — not exclusive. The full inventory carries all 11 sovereign Canadian providers.

Bucket 2: Canadian-hosted but H100 not in current catalog (7 vendors)

These vendors operate Canadian data centres but don’t sell H100s. They appear in the full inventory because they’re legitimate Canadian compute options for non-H100 workloads:

Bucket 3: Hyperscalers with H100 listed but Canadian availability unverified (3 vendors)

These are tracked but not indexed because the catalog says yes but the console says no:

If these become verifiable, they’ll move into the ranked tier.

Bucket 4: Marketplace or variable pricing (4 vendors)

Marketplace and aggregator-listed vendors don’t fit a fixed-rate ranking. They’re tracked for completeness.

Bucket 5: Colocation operators (1 vendor not already counted above)

eStruxture (Montreal/Toronto/Vancouver/Calgary) — Canadian-owned colocation operator that anchors the ThinkOn government cloud and hosts compute for several sovereign providers. Not a GPU rental but a foundational Canadian sovereignty enabler. Tracked in the sovereign Canadian count.

Reconciliation

The numbers tie out: 14 indexed + 25 excluded = 39 in scope.

Limitations and known gaps

This methodology is honest about what it can’t do well:

• Canadian H100 availability changes faster than this page updates. Capacity that was constrained last quarter may be available next; quoted rates change with cloud-provider promotions. Verify with the vendor before procurement.
• Microsoft Azure Canada Central H100 pricing ($7.50/GPU·hr) is a midpoint estimate. Azure pricing varies by SKU, region, and reservation tier; pull from the Azure Pricing API for production decisions.
• Oracle Cloud and IBM Cloud Canadian H100 rates are not pulled in this version. They are tracked in the inventory but indexed values are blank.
• Sovereign Canadian providers don’t publish list rates. The four anchors flagged “Opaque” — TELUS, Bell, BUZZ HPC, Hypertec — and the additional sovereign opaques (CoEvo, Consensus Core, ThinkOn, etc.) require direct vendor conversations for pricing and capacity. Request a quote from the vendor for live numbers and capacity.
• Marketplace pricing is real but not directly comparable. Vast.ai hosts may offer H100 at $1.00/hour with no SLA. SaladCloud delivers consumer GPU compute at sub-$0.30/hour rates. These are legitimate options for the right workloads but aren’t comparable to enterprise-grade dedicated capacity.

Update cadence

The inventory is verified quarterly. Each row carries a last_verified_date field — most rows are dated to the most recent compilation pass; a few (notably Canadian Web Hosting) carry older verification dates and are flagged accordingly.

Material vendor announcements (new Canadian regions, sovereignty milestones, major pricing changes) trigger interim updates between quarterly refreshes.

How to read the index responsibly

The Zeever AI Compute Index is a starting point for procurement, not a substitute for one. Three rules:

1. The cheapest published H100 is not always the right H100.ISAIC at CAD $2.50/hour is a research sandbox without SLAs. RunPod Community Cloud at $1.99/hour is shared infrastructure. AWS P5 at $4.10/hour comes with capacity guarantees, network performance, and enterprise support that the cheaper options don’t. Match the price tier to the workload tier.

2. “Opaque” is not “expensive.” The four sovereign anchors marked Opaque on the public card may quote rates competitive with hyperscalers — or higher. The point of the Opaque flag is that list pricing isn’t published, not that the underlying rate is high. The only way to know is to request a quote.

3. Sovereignty has a price, but the price isn’t always what you think. Some sovereign Canadian providers may quote rates higher than US hyperscalers. Some may quote rates lower because their cost structure (Canadian hydroelectric, federal infrastructure subsidies, OEM-direct hardware) is genuinely different. Ask, and compare.

Data sources

The vendor inventory is compiled from:

• Vendor pricing pages verified directly via web request on the compilation date
• Vendor press releases and announcements for capacity, partnership, and regional launch claims
• Regulatory filings including CPPIB infrastructure investment disclosures, Bell Canada/BCE filings, and HIVE Digital Technologies (TSXV: HIVE) public filings for BUZZ HPC capacity claims
• Industry audits including Founder Reality’s Canadian GPU audit for hyperscaler ground-truth pricing and availability
• News reporting from The Globe and Mail, The Logic, Data Center Dynamics, HPCwire, and Connect CRE for Canadian sovereign AI capacity announcements
• Government program documentation including Innovation, Science and Economic Development Canada’s $2-billion Sovereign AI Compute Strategy disclosures

Per-row source URLs are recorded in the v3 CSV under the source_urls column. The full CSV is available for download at zeever.ca/canadas-ai-compute-landscape.

Contact and corrections

Found an error? Vendor pricing changed? New vendor we should track? Email info [at] zeever [dot] ca or file an issue. Every correction is reviewed and the inventory is republished.

The Zeever AI Compute Index is maintained as a public good for the Canadian AI ecosystem. It is not a paid placement service — vendor inclusion is determined by the inclusion criteria above, not by commercial relationships.