Why choose the Gaudi2?

Comes with 96GB HBM2e and 24x 100GbE RoCEv2 ports integrated on-chip, so clusters scale out over standard Ethernet without separate NICs. FP8 support and competitive pricing make it a cost-efficient option for training and inference on the Intel Gaudi software stack.

When is the Gaudi2 not a good fit?

The Gaudi software stack supports fewer frameworks and models than CUDA, so workloads relying on niche libraries may need porting. For broad ecosystem compatibility, H100 or A100 remain easier to adopt.

Are Gaudi2 prices going up or down?

The median on-demand price has held steady around $1.13/hr per GPU across providers.

What size AI models can the Gaudi2 run?

With 96GB of VRAM, the Gaudi2 is well suited to 30B-class models in FP16, and 70B-class models in 4-bit or 8-bit quantized form.

How much VRAM does the Gaudi2 have?

The Gaudi2 has 96GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.

What is the Gaudi2's memory bandwidth?

The Gaudi2 has 2,450 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.

What data types does the Gaudi2 support?

The Gaudi2 supports 6 precision formats. Training: BF16, FP16, TF32, FP32. Inference: FP8, INT8.

Does the Gaudi2 support NVLink?

No. The Gaudi2 is a PCIe-only GPU with no NVLink, so it is better suited to single-GPU inference and smaller-scale workloads than large distributed training jobs.

How much does the Gaudi2 cost per hour?

Gaudi2 pricing currently ranges from $0.58/hr to $1.21/hr per GPU, depending on the provider, instance type, and billing model.

How much does the Gaudi2 cost per month?

At 720 hours per month, one Gaudi2 can cost between $420.30 to $871.20 per month, depending on the provider. Reserved and spot pricing can lower that further.

Which cloud providers offer the Gaudi2?

The Gaudi2 is available from 2 cloud providers: Sesterce, Cyfuture AI.

Can I rent the Gaudi2 in the cloud?

Yes. We currently track 17 Gaudi2 listings across 2 cloud providers: Billing type Listings Avg $/GPU/hr On-demand 5 $1.08/hr Reserved 12 $0.71/hr

Intel Gaudi2

Intel's data center AI accelerator with on-chip Ethernet for cost-efficient scale-out.

Last update 7 minutes ago

Compare vs other GPUs →

Aggregating historical prices...

Key Specifications

Architecture

Gaudi (2nd Gen)

Memory

96GB HBM2e

Memory Bandwidth

2,450 GB/s

Release date

Q2 2022

Gaudi2 Pricing and Availability

We don't see much volatility for the Gaudi2 right now. Most providers are clustered between $1.02 and $1.21/hr per GPU, so availability is likely the deciding factor.

2 providers

How this list works

The two views

By provider shows one card per company, placed where its best offer ranked. By configuration lists every offer.

Order

What you can rent comes first: in stock, then waitlist, then not reported, then out of stock. Priced offers rank ahead of quote-only, and on-demand ahead of other billing types.

Within a group, five factors set the order, largest first:

Location: datacenter proximity, blended with provider HQ.
Price: hourly price, per GPU and in total.
Billing type: reserved ahead of spot, spot ahead of quote-only.
Specs: more VRAM, vCPUs and RAM.
Provider diversity: in By configuration, a provider's repeat rows rank slightly lower, so one company can't take all the top spots.

Search

We match the provider, GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.

Transparency and funding

Ads and sponsors: Paid placements sit at the top of the list and are always labeled as sponsored content. Sponsorship never influences the organic ranking itself.
Affiliates: Affiliate links are marked. We may earn a commission if you click them, but commissions never affect the order.
Prices: Shown in USD, converted at daily reference rates where a provider publishes in another currency. A month means 720 hours. How we estimate costs.

Sesterce In stock

France 1 config 8x

On-Demand from $1.21

From $1.21 / GPU / hr On-Demand Visit website


8x Gaudi2	768GB	152	940GB	On-Demand	$1.21	$9.68	In stock

Cyfuture AI

India 16 configs 1x-8x

On-Demand from $1.02 Reserved from $0.58

From $1.02 / GPU / hr On-Demand Visit website

Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.

Frequently Asked Questions

Why choose the Gaudi2?: Comes with 96GB HBM2e and 24x 100GbE RoCEv2 ports integrated on-chip, so clusters scale out over standard Ethernet without separate NICs. FP8 support and competitive pricing make it a cost-efficient option for training and inference on the Intel Gaudi software stack.
When is the Gaudi2 not a good fit?: The Gaudi software stack supports fewer frameworks and models than CUDA, so workloads relying on niche libraries may need porting. For broad ecosystem compatibility, H100 or A100 remain easier to adopt.
Are Gaudi2 prices going up or down?: The median on-demand price has held steady around $1.13/hr per GPU across providers.
What size AI models can the Gaudi2 run?: With 96GB of VRAM, the Gaudi2 is well suited to 30B-class models in FP16, and 70B-class models in 4-bit or 8-bit quantized form.
How much VRAM does the Gaudi2 have?: The Gaudi2 has 96GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.
What is the Gaudi2's memory bandwidth?: The Gaudi2 has 2,450 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.
What data types does the Gaudi2 support?: The Gaudi2 supports 6 precision formats. Training: BF16, FP16, TF32, FP32. Inference: FP8, INT8.
Does the Gaudi2 support NVLink?: No. The Gaudi2 is a PCIe-only GPU with no NVLink, so it is better suited to single-GPU inference and smaller-scale workloads than large distributed training jobs.
How much does the Gaudi2 cost per hour?: Gaudi2 pricing currently ranges from $0.58/hr to $1.21/hr per GPU, depending on the provider, instance type, and billing model.
How much does the Gaudi2 cost per month?: At 720 hours per month, one Gaudi2 can cost between $420.30 to $871.20 per month, depending on the provider. Reserved and spot pricing can lower that further.
Which cloud providers offer the Gaudi2?: The Gaudi2 is available from 2 cloud providers: Sesterce, Cyfuture AI.
Can I rent the Gaudi2 in the cloud?: Yes. We currently track 17 Gaudi2 listings across 2 cloud providers:

Billing type Listings Avg $/GPU/hr

On-demand 5 $1.08/hr

Reserved 12 $0.71/hr

Billing type	Listings	Avg $/GPU/hr
On-demand	5	$1.08/hr
Reserved	12	$0.71/hr

Technical Specifications


Architecture	Gaudi (2nd Gen)
Process	7nm
Memory	96 GB HBM2E
Memory Bandwidth	2.45 TB/s
On-chip SRAM	48 MB
Compute Engines	2x Matrix Multiplication Engine (MME), 24x Tensor Processor Core (TPC)
Networking	24x 100 GbE RoCEv2 (2,400 Gb/s)
Host Interface	PCIe Gen4 x16
Media	Integrated decoders (HEVC, H.264, VP9, JPEG)
Supported Data Types	FP32, TF32, BF16, FP16, FP8, INT8
Form Factor	OAM (HL-225H)
Max TDP	600W