TheAICANADIAN DATA CENTER
Bare metal in a Tier-3 facility in Canada. Root access, no virtualization, no oversubscription.
FIRST RACK — 10 NODES · RESERVE AHEAD · Q4 2026
Per-minute B300 cloud — opens Q4 2026.
Reserve ahead: whole bare-metal nodes from $5.30/GPU·hr. Free waitlist — or a 30% prepayment locks your slot & rate.
On-demand $6.90/GPU·hr — below major comparable whole-node offers (Hyperstack $7.40 · Verda $7.50 · Nebius $7.85).
RESERVE NODES ↓Talk to an infrastructure engineer~7M Blackwell GPUs shipped, almost none rentable → read the count
- 1 Reserve — free, no card
- 2 We confirm capacity & talk
- 3 30% invoice locks slot & rate
Reserve capacity.
Whole bare-metal nodes, one config each — pick a SKU, quantity and term. Free waitlist, or a 30% prepayment locks your slot & rate.
PAY AS YOU GO — $6.90/GPU·HR · FROM Q4 2026 · PER-MINUTE · NO COMMITMENTS.
NVIDIA B300
8× B300 SXM6 · 288GB HBM3e · NVLink 5 · 400G RDMA · whole node
= 8× B300 SXM6 · 288GB
Pick a term to see your take-or-pay total
STARTS Q4 2026 · 30% PREPAY LOCKS SLOT & RATE
B300 pricing vs the field.
| Provider · B300 $/GPU·HR | On-demand | 1 month | 3 months | 6 months | 1 year |
|---|---|---|---|---|---|
| TheAI Cloud reserved · opens Q4 2026 | $6.90 at launch | $6.75 | $6.55 | $5.70 | $5.30 |
| Hyperstack | $7.40 | — | — | — | — |
| Verda | $7.50 | $7.35 | $7.28 | $7.20 | $6.90 |
| Nebius | $7.85 | — | — | — | — |
PUBLIC PRICING & DIRECT QUOTES, JUL 23, 2026 · B300 PER GPU·HOUR
COREWEAVE & LAMBDA QUOTE B300 RESERVED VIA SALES ONLY · AWS CAPACITY BLOCKS LIST AT $14.04/GPU·HR
On-demand at $6.90 runs below the major comparable whole-node offers — Hyperstack $7.40, Verda $7.50, Nebius $7.85. Six-month reserved at $5.70 runs ~21% below Verda's; one-year at $5.30 is ~23% under Verda's annual and still beats a two-year full-prepay elsewhere ($5.625).
Built for training runs.
NVLINK 5 + 400G RDMA
Non-blocking fabric between nodes — multi-node training works day one.
PREPAID · PER-MINUTE
At launch you pay per minute from a prepaid balance. No egress fees. SLA 99.5% with service credits.
SCREENED CUSTOMERS
Every organization is verified before workloads run — U.S. export rules enforced.
Node spec.
One SKU — a whole 8× B300 SXM6 node. Exact configuration, nothing shared.
| Node | Aivres KR9288-X3 10U · 8× NVIDIA B300 SXM6 288GB (HGX B300 8-GPU) · 2.3TB HBM3e per node |
|---|---|
| CPU | 2× AMD EPYC 9555 — 64C / 128T each (128 cores, 256 threads per node) |
| Memory | 2.3TB DDR5 (24× 96GB) |
| Storage | 4× 3.84TB NVMe (15.4TB local) + 2× 960GB M.2 NVMe boot · 20TB WEKA shared storage included |
| Networking | ConnectX-7 400GbE OSFP fabric (RoCE v2) · NVLink 5 (NV18) in-node · 24 public IPv4 · no egress fees |
| Location | Canada (Tier-3 facility) |
| Every contract | NCCL acceptance benchmarks before billing · 99.5% SLA with credits · per-minute billing |
Before you reserve.
When exactly does capacity open?
Q4 2026. Reservation holders get the exact opening date by email at least two weeks in advance, plus a 48-hour confirmation window before their slot activates.
Is the reservation paid or binding?
No. It's a free waitlist entry — no card required. Rates are subject to change until confirmed.
How do I lock my slot and rate?
Submit a reservation → we get in touch → you receive a 30% prepayment invoice (wire or USDC) → your slot and rate are locked. Locked terms are take-or-pay. Payment schedule for the remaining 70% is agreed individually — monthly invoicing is standard.
Can I cancel a locked reservation?
A locked reservation is a rental contract, not a hold. The 30% prepayment secures your slot and rate on a take-or-pay basis and is not returned if you cancel — that's what makes the lock real. Until you lock, a waitlist entry is free and carries no obligation.
What happens at zero balance?
PAYG instances are terminated immediately and drives are wiped — keep your checkpoints synced out. We email you at ~24h and ~4h of remaining runway, and the console shows a live estimate.
Can I train across multiple nodes?
Yes. B300 nodes interconnect over a non-blocking 400G RDMA fabric (RoCE v2) — reserve two or more and they're delivered as one cluster.
How do I pay at launch?
Card (Stripe), wire transfer, or USDC — all top up a prepaid balance. PAYG bills per minute; there are no egress fees. Locked reservations follow your capacity agreement: 30% already paid, the remainder typically via monthly invoicing.
What SLA do you offer?
99.5% monthly uptime per node at launch, with service credits for shortfalls — see Terms for the credit schedule.
Who can use the service?
Our nodes are hosted in Canada; the hardware is U.S.-origin, so U.S. export controls apply and we follow them: every organization passes screening before workloads run. Card verification is automatic; wire/USDC onboarding is manual.
From the blog.
How we price whole-node B300 — and what we learned scanning the market.
How Much Blackwell Actually Exists?
Blackwell shipments set records every quarter — and the median public B300 on-demand price rose 57% since November. We reconstruct how much has actually shipped (~7M packages), how little of it is rentable (two sub-12-month listings), and why prices rise while supply booms.
The Blackwell Fact Table
Every figure from “How Much Blackwell Actually Exists?” with its source: shipments, the B300 slice, capex and anchor deals, manufacturing ceilings, Rubin, the H100 precedent — plus methodology notes and unit conventions.
Kimi K3 (2.8T) and six other open models on a single 8×B300 node — full vLLM numbers
We ran seven current open models — from Gemma 4 31B to Kimi K3's 2.78T parameters — through out-of-box vLLM on one 8× B300 node. TTFT, throughput, what didn't launch, and what surprised us.
Team.
Oleksiy Sitnichenko
CEO
Over 20 years of global executive leadership: business management, entrepreneurship, and strategic development across diverse international markets.
Viacheslav Zhvart
CTO, GPU infrastructure supply
Leads a unit that puts datacenter GPU capacity into production. Talks to DCs, engineers, and buyers in one thread — from rack constraints to contract language.
Mykola Baichak
Head of ML Infrastructure
Senior engineer with 10+ years building high-performance systems across autonomous driving, quantitative trading, and large-scale ML inference.
Pavel Solovskyi
ML Infrastructure & Performance Engineering
Infrastructure engineer specializing in confidential AI compute, GPU performance engineering, and large-scale inference operations.