Confidential NVIDIA RTX PRO 6000 Blackwell in an Intel TDX VM
96 GB, one GPU, Intel TDXAn NVIDIA RTX PRO 6000 Blackwell with 96 GB of GPU memory costs $3.80 per GPU-hour in a VoltageGPU Confidential VM (Intel TDX), billed per second with no commitment; the VM comes with 32 vCPUs and 128 GB of RAM. Its Intel TDX quote and NVIDIA GPU attestation were generated inside the VM and verified on 17 September 2026, and again on 9 October 2026. Stock on 22 November 2026 at 12:35 UTC: 11 VMs available. Check a quote or a token yourself, field by field, in your browser.
The VM you get
GPU memory, vCPUs and RAM as our provider lists the size on 22 November 2026 at 12:35 UTC; the GPU name and its confidential-computing mode as nvidia-smi read them inside the VM.
| GPU | NVIDIA RTX PRO 6000 Blackwell Server Edition |
| GPUs per VM | 1 |
| GPU memory | 96 GB |
| vCPUs | 32 |
| RAM | 128 GB |
| Size (API resource_name) | rtx6000b-small |
| Confidential computing | Intel TDX ✓ |
| GPU confidential-computing mode | ON (nvidia-smi, 17 Sep 2026) ✓ |
| NVIDIA GPU attestation | Verified 17 Sep and 9 Oct 2026 ✓ |
| Per-second billing | Yes ✓ |
RTX PRO 6000 Blackwell price per hour
At this reading it is the cheapest single-GPU Confidential VM we list. Every size and price
What the attestation shows
Read inside an rtx6000b-small VM on 17 September 2026, on a challenge the tenant chose. Every file is public.
- 17 September 2026: the capture inside the VM, RESULT: VERIFIED
- 17 September 2026: deploy, attest and verify through the public API alone
- 9 October 2026: re-run, verified offline, a wrong challenge rejected
- Raw outputs: nvidia-smi conf-compute output · voltage-verify output
What a proof does not establish, with its sources: the limits on the proofs page.
Check it yourself
The proof is generated inside your VM, on your challenge, and verified on your own machine.
# Inside the VM (the image ships without pip)
curl -sS https://bootstrap.pypa.io/get-pip.py | python3 - --user --break-system-packages
python3 -m pip install --user --break-system-packages "voltage-verify[attest]>=0.3.0"
export PATH=$HOME/.local/bin:$PATH
voltage-verify manifest --challenge auto -o manifest.json
sudo env PATH=$PATH PYTHONPATH=$(python3 -c 'import site;print(site.getusersitepackages())') \
voltage-verify attest --manifest manifest.json --mode single-gpu -o bundle.json
# On your own machine, with bundle.json copied out of the VM
pip install "voltage-verify>=0.3.0"
voltage-verify verify bundle.json --challenge <the challenge printed above> --hwmodel GB20X --gpus 1
# RESULT: VERIFIED; a wrong --challenge answers NOT VERIFIEDvoltage-verify is MIT-licensed, on PyPI, source on GitHub. To read a quote or a token field by field without installing anything: the free decoders.
Confidential RTX PRO 6000, in short
How much does a confidential RTX PRO 6000 cost per hour?
$3.80 per hour for the whole VM, which has one RTX PRO 6000 Blackwell GPU (96 GB), billed per second with no commitment. Price read from our live inventory on 22 November 2026 at 12:35 UTC.
How much GPU memory does the RTX PRO 6000 VM have?
The listing gives 96 GB. Inside the VM on 17 September 2026, nvidia-smi reported 97,887 MiB in total, with 99,463,296 KiB protected by confidential computing and 0 KiB unprotected.
What does the attestation prove?
That the VM is an Intel TDX trust domain whose quote verifies up to Intel's root, and that its GPU ran in confidential-computing mode, with NVIDIA-signed attestation tokens bound to a challenge the tenant chose. It does not tell you which software runs in the VM, and it does not rule out an attacker with physical access to the server.
How do I verify it myself?
Inside the VM, run voltage-verify attest on a challenge you choose, copy bundle.json out, then run voltage-verify verify bundle.json --challenge <your challenge> on your own machine. A wrong challenge answers NOT VERIFIED. The bundles published on this page can be checked the same way.
How many GPUs does the VM have?
One. The listed size, rtx6000b-small, is a single-GPU VM, and nvidia-smi inside it reads Multi-GPU Mode None.
Run your workload on an attested RTX PRO 6000
Prepaid balance, billed per second, the unused part of the first hour refunded when you stop.
NVIDIA RTX PRO 6000 Blackwell facts: prices, billing and limits
Prices read from our live inventory on 22 November 2026 at 12:35 UTC, in US dollars.
Confidential VM (Intel TDX)
- RTX PRO 6000 Blackwell, 1 GPU: $3.80 per hour for the whole VM, 11 VMs available.
Standard GPUs (no enclave)
This page covers the Confidential VM only. Standard GPUs, without an enclave, are listed with their live stock on /standard-gpus.
Billing
- A Confidential VM price covers the whole VM, every GPU in it; a standard GPU price is per GPU-hour.
- Prices are shown per hour and billed per second. The first hour is charged when the machine starts; after that the running time is prorated to the second, and unused prepaid time is refunded when you stop it.
- Storage volumes on standard machines are billed separately, on the data stored.
Limits
- Spot capacity can be reclaimed by the host, and it is not always cheaper than on demand.
- We have run no attestation on B200 and B300 GPUs, so we claim none for them.
- Bare-metal servers, standard GPUs and sandboxes have no enclave and are never attested: use them only for data that is not sensitive. The confidential container tier, which gave the tenant no attestation, was retired on 1 October 2026.
What a Confidential VM proof establishes
- An Intel TDX trust domain, not in debug mode
- A platform Intel rates up to date
- Both proofs are bound to one manifest
- NVIDIA GPU attestation, on the GPUs where we ran it
- Fresh, when you drew the challenge
What it does not establish
- Attacks by someone with physical access to the server
- Which software runs in the VM
- That the trust domain and the GPU are the same machine
- That your SSH session ends inside the attested trust domain
- The NVLink fabric of 8-GPU nodes