Standard H200: 15 available nowfrom $7.70/GPU-hourNo enclave, for work that isn't sensitive.Rent a standard H200 →

Hardware check on a real rented standard H200, 16 October 2026: matches its listing

NVIDIA H200 141GB Confidential GPU Compute

Rent an NVIDIA H200 141GB HBM3e inside an Intel TDX Confidential VM for $8.08 per hour, billed per second, or a standard H200 (no enclave) at the live price shown above. Hopper-refresh architecture, 4.8 TB/s memory bandwidth, 700W TDP. Fine-tune 70B+ parameter models, run multi-million token inference, and train on proprietary data with Intel TDX memory isolation. On the single-GPU H200 Confidential VM, the Intel TDX quote and the NVIDIA GPU attestation were both verified from inside the VM on 4 September 2026, with the GPU in CC mode. GDPR Article 28 processor obligations, DPA available. Deploy in minutes.

Hourly prices of 8x H200 at AWS and Azure, dated and sourced: H200 price, AWS vs Azure vs a Confidential VM. Which cloud offers a confidential H200 at all: confidential GPU providers compared. Choosing between GPUs? See the NVIDIA B200 180GB for native FP4 and frontier-model training, the NVIDIA H100 80GB for the lowest-cost confidential entry point, our live prices, and the Intel TDX security architecture.

Confidential GPU

NVIDIA H200

141 GB HBM3e, Intel TDX

The high-memory AI GPU, available inside Intel TDX confidential VMs. 141 GB HBM3e, 4.8 TB/s bandwidth, Hopper architecture. Fine-tune 70B+ models, run million-token inference, and train on proprietary data.

$8.08/GPU/hour

Deploy H200 Now How It Works

Specifications

Hardware Details

GPU	NVIDIA H200
VRAM	141 GB HBM3e
Architecture	Hopper
Memory Bandwidth	4.8 TB/s
TDP	700W
Confidential Computing	Intel TDX ✓
Per-second billing	Yes ✓

Use Cases

Built For

Large Model Training

Train 70B+ parameter models on proprietary data. 141 GB of HBM3e means no model sharding needed.

High-Throughput Inference

Run multi-million token inference workloads with 4.8 TB/s bandwidth. Ideal for long-context confidential document processing.

Confidential RAG

Build retrieval-augmented generation pipelines on sensitive corporate data. Embeddings and context stay inside the Intel TDX confidential VM.

Distributed Training

Scale to multi-GPU training with hardware-enforced isolation on every node. Data stays encrypted across the cluster.

Pricing

Simple, Transparent Pricing

1x H200

$8.08

per hour

Monthly (730h)

$5,898

per GPU/month

8x H200

—

per hour (all 8 GPUs)

Security

Why Confidential Matters

On the confidential tier, every H200 is attached to an Intel TDX confidential VM: VM memory is encrypted by the CPU and isolated from the host operator.

Memory encrypted by the CPU, with a key per trust domain that the host cannot read

GPU inside the trust domain (NVIDIA confidential-computing mode on for single-GPU H200 VMs; the 8x H200 node runs NVIDIA's multi-GPU Protected PCIe mode, not attested yet)

Publicly verifiable attestation, cryptographic proof the enclave is genuine

GDPR Article 28 processor obligations, DPA available

Operator-excluded trust boundary: we have no technical access to your trust-domain memory or disk

Ready to Deploy?

Deploy an H200 in minutes. Per-second billing, no commitment. Welcome credit of $3.80, one hour on an RTX PRO 6000 Blackwell VM, with your first top-up of $10 or more (verified email).

Get started · $3.80 welcome credit Talk to Sales

Other GPUs you can rent

B300 · B200 · H100 · RTX 5090 · RTX 4090 · RTX 3090 · A100 · A6000 · L40S · RTX A4000 · RTX A5000 · Live stock and prices