Rent GB300 NVL72 in Australia: Availability and Wholesale Rates

GB300

Australia

{"@context":"https://schema.org","@graph":[{"@type":"Service","@id":"https://gpuaas.com/gpu/gb300-australia#service","name":"GB300 NVL72 GPU Cloud in Australia","provider":{"@type":"Organization","name":"GPUaaS.com","url":"https://gpuaas.com"},"areaServed":{"@type":"Country","name":"Australia"},"serviceType":"GPU cloud infrastructure","description":"Rent NVIDIA GB300 NVL72 rack capacity in Australia from vetted partners. SharonAI's 40,000-GPU commitment anchors regional supply. Quotes within 24 hours."},{"@type":"WebPage","@id":"https://gpuaas.com/gpu/gb300-australia#webpage","url":"https://gpuaas.com/gpu/gb300-australia","name":"GB300 NVL72 GPU Cloud in Australia","isPartOf":{"@type":"WebSite","name":"GPUaaS.com","url":"https://gpuaas.com"}},{"@type":"BreadcrumbList","@id":"https://gpuaas.com/gpu/gb300-australia#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://gpuaas.com"},{"@type":"ListItem","position":2,"name":"GPU Cloud in Australia","item":"https://gpuaas.com/gpu-cloud/australia"},{"@type":"ListItem","position":3,"name":"GB300 in Australia","item":"https://gpuaas.com/gpu/gb300-australia"}]},{"@type":"FAQPage","@id":"https://gpuaas.com/gpu/gb300-australia#faq","mainEntity":[{"@type":"Question","name":"Can I rent GB300 NVL72 capacity in Australia?","acceptedAnswer":{"@type":"Answer","text":"Yes. SharonAI's agreement with NVIDIA, signed 8 June 2026, commits to a 72MW AI factory scaling to as many as 40,000 GB300 GPUs."}},{"@type":"Question","name":"How much does GB300 cost per hour in Australia?","acceptedAnswer":{"@type":"Answer","text":"Tracked on-demand rates span roughly $4.00 to $18.00 per GPU as of September 2026, across a thin set of providers globally."}}]}]}

◆ THE SHORT ANSWER

Yes. You can rent GB300 NVL72 capacity in Australia now through vetted partners, contracted directly with the operator running your rack. SharonAI signed an agreement with NVIDIA on 8 June 2026 for a 72MW AI factory in Australia, scaling to as many as 40,000 GB300 GPUs, one of the largest confirmed Blackwell Ultra commitments in the region. On GB300 price per hour, tracked on-demand rates span roughly $4.00 to $18.00 per GPU across a handful of providers as of September 2026, a wider band than established generations since fewer platforms publish a rate at all. GB300 NVL72 pairs 72 Blackwell Ultra GPUs with 36 Grace CPUs in a single 130TB/s NVLink domain, each GPU carrying 288GB of HBM3e for 20TB of aggregate rack memory. Australian-incorporated operators sit outside the US CLOUD Act. For smaller workloads, B300 and B200 in Australia remain the cheaper, more established options.

◆ PRICING

What GB300 capacity costs in Australia

GB300 pricing here spans a wide band since few providers publish a rate at all. Confirming an operator's actual power headroom for this specific generation matters more than comparing quoted rates, since B300 or B200 capacity doesn't guarantee GB300 headroom.

Market reference ranges as of September 2026, quoted in USD, not Australia-specific quotes. Figures are global and directional given the thin, fast-moving GB300 provider set.

Wholesale rates through GPUaaS.com are quoted per enquiry and vary by commitment term, configuration and placement.

$0$2.50$5$7.50$10$12.50$15/GPU-HR

Market low, tracked providers

Cheapest tracked GB300 on-demand

$4.00

Approx. median, tracked providers

Directional given thin provider count

$9.50

Market high, specialist providers

Specialist providers, Australia placement

$16.00

Hyperscaler on-demand

What you pay without a broker

$18.00

◆ GPUaaS.com wholesale

Vetted partners · direct operator contract

quoted per enquiry

◆ GB300 RATES VARY WIDELY ACROSS A THIN PROVIDER SET

◆

Where Australian GB300 capacity earns its keep

The workloads Australian GB300 capacity is built for.

SharonAI's 8 June 2026 agreement with NVIDIA commits to a 72MW AI factory in Australia, scaling to as many as 40,000 GB300 GPUs, one of the largest confirmed Blackwell Ultra buildouts anywhere in the region. Australian-incorporated operators sit outside the US CLOUD Act, which government and regulated buyers increasingly treat as the deciding factor.

/01

Rack-scale coherent training

Train frontier models that need coherent memory across dozens of GPUs at once, backed by one of the largest confirmed Blackwell Ultra commitments in the region, where the 130TB/s NVLink domain avoids the interconnect bottleneck of individual-card clusters.

GB300 NVL72 · 72-GPU domain · 20TB HBM3e

/02

Trillion-parameter inference

Serve trillion-parameter reasoning models in production, where NVIDIA rates GB300 NVL72 at roughly 10x the tokens-per-second-per-user of Hopper-based platforms for test-time scaling workloads.

GB300 rental · NVFP4 · test-time scaling

/03

Long-context at rack scale

Run the longest-context reasoning chains without offloading KV cache anywhere, holding more context in-rack than any generation covered on this site, from Australian-incorporated operators outside the US CLOUD Act.

20TB aggregate · 288GB per GPU · CLOUD Act

/04

Agentic workloads at scale

Serve high-concurrency, multi-agent workloads where token throughput per rack, not just per GPU, decides your cost per request, at NVIDIA's rated 5x improvement in throughput per watt over Hopper.

GB300 cluster · Dynamo · multi-agent

◆ LIVE NETWORK · 12 LOCATIONS

Vetted GPU partners worldwide, including Australian GB300 capacity.

Australia hosts one of the largest confirmed Blackwell Ultra commitments in the region. Tell us node count, interconnect and placement, and you talk straight to the operator running your rack.

Read the full guide to GPU cloud in this location →

GPU GENERATIONS

20+

VETTED PARTNERS

PLACEMENT OPTIONS

24h

QUOTE TURNAROUND

◆ USA◆ CAN◆ UK◆ DEU◆ FRA◆ NLD◆ UAE◆ SAU◆ IND◆ SGP◆ JPN◆ AUS

◆ COST COMPARISON

See how much you save at scale

Wholesale rates against cloud list price for a 64-GPU cluster.

CLUSTER SIZE

8 HGX nodes

64 × GPUS · 730 HRS/MO

ASSUMPTIONS · BLENDED $6.00/GPU-HR · INDICATIVE ONLY

SOURCEEST. MONTHLYVS GPUAAS

Retail cloud

On-demand list price · reserved discounts require lock-in

~$280k

+$84k

Direct datacentre negotiation

Long-term commitment · slow procurement cycle

~$230k

+$34k

◆ BEST VALUE

GPUaaS.com wholesale

Vetted partners · direct operator contract · quotes in 24 hours

~$196k

SAVE ~$84k/MO

Need single-GPU compute? packet.ai has you covered.

◆ HOW IT WORKS

A matchmaker, not a marketplace.

We connect you to our vetted partners. You contract directly with the operator running your nodes.

STEP 01/4

Tell us the requirement

GPU model, count, placement and timeline. Add workload detail if you have it.

STEP 02/4

We match capacity

We find vetted partners with capacity that fits, in the jurisdiction you need.

STEP 03/4

Quotes in 24 hours

Real quotes from partners who hold the capacity, not listings that may not exist.

STEP 04/4

Contract and provision

You contract directly with the operator. We smooth the provisioning process.

Ready to spec your cluster?

Get H100 cluster quotes
in under 24 hours.

Tell us the essentials. We'll line up real quotes from vetted wholesale providers . direct, no platform fee.

◆Quotes in under 24 hours

◆Direct contact with operators

◆No middle-man markup

◆20+ vetted providers · 10 regions

ESSENTIALS

OPTIONAL

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.

◆ FAQ

Frequently Asked Questions

Can I rent GB300 NVL72 capacity in Australia?

Yes. SharonAI's agreement with NVIDIA, signed 8 June 2026, commits to a 72MW AI factory scaling to as many as 40,000 GB300 GPUs, one of the largest confirmed Blackwell Ultra commitments in the region. Tell us your requirement and we will confirm availability within 24 hours.

How much does GB300 cost per hour in Australia?

Tracked on-demand rates span roughly $4.00 to $18.00 per GPU as of September 2026, across a thin set of providers globally. Several major platforms quote GB300 only on request rather than publishing an hourly rate. Wholesale rates through GPUaaS.com are quoted per enquiry and vary by commitment term, configuration and rack allocation.

GB300 vs B300: what's different?

GB300 NVL72 is a rack-scale system, not a single GPU: 72 Blackwell Ultra GPUs and 36 Grace CPUs share a single 130TB/s NVLink domain with 20TB of aggregate HBM3e. Full GB300 specs put each GPU at 288GB of HBM3e; B300 is the same silicon sold as an individual GPU or GB300 server node. GB300 suits jobs that need coherent memory across many GPUs at once; B300 suits everything that fits on one card.

Does having B300 capacity mean an operator has GB300 too?

Not automatically. Grid power secured for earlier generations at their lower draw does not guarantee headroom for GB300's roughly 135-140kW-per-rack draw, so this is worth confirming with an operator directly for this specific generation.

Does Australian GB300 capacity keep me outside the US CLOUD Act?

Yes, if the operator is Australian-incorporated. US-incorporated providers remain reachable under the CLOUD Act even when their servers sit in Sydney, since the Act attaches to incorporation, not server location.

What commitment terms are available for Australian GB300 capacity?

Both on-demand and reserved terms exist, though reserved capacity is where most volume sits given how concentrated GB300 deployment remains among operators who control their own power and cooling. Tell us your timeline and we will confirm which route fits.

GB300 in Australia, at wholesale price.