#1
Voltage Park
US neocloud selling on-demand and reserved bare-metal NVIDIA HGX H100 clusters from $1.99/hr; acquired TensorDock (2025) and merged with Lightning AI (Jan 2026).
from $1.99 per GPU/hr
Petals Review (archive)
An influential open-source distributed-inference research project that is now effectively abandoned: no commits since September 2024, no releases since 2023, and the public network health endpoint and chat demo no longer work reliably.
dead no KYC
Visit Petals
Read full review
site reachable · checked 11 h ago · daily check
// Review
by the bestcloudmining editorial desk · review version 1 · prose updated 2026-09-24 · facts as of 2026-09-24 · how we verify
Overview
Petals is a GPU compute marketplace in our compute category, operating since 2022. Facts are unverified as of 2026-09-24 - treat claims accordingly.
What you can rent
BitTorrent-style collaborative LLM inference: run large models by pooling consumer GPUs across the internet; not a rental marketplace, a swarm-inference framework You do not own the underlying hardware unless the provider explicitly says otherwise - with a contract you hold a claim on hashrate or compute for a term, not an asset you can resell.
Contract & terms
Minimum purchase: not documented. Contract term: not documented. Maintenance/fee: not documented. Payouts: not documented. N/A Where a term is not documented, assume it can change and get it in writing before paying.
Proof & verifiability
Proof of hashrate/compute is not documented. None (trust-based peer swarm) A vendor claim is not proof: check for a payout address, a pool worker you can watch, or a hardware attestation before trusting a large contract.
Reliability & history
No verified reputation data yet. Noted concerns: . Facts not fully verified (unverified) as of 2026-09-24 - start with the smallest possible purchase and test a payout before scaling up.
Pricing
Documented pricing: not documented - see the provider site. Payment methods not documented. KYC policy not documented. Compare prices only in the same unit - a per-TH/s-day figure and a per-GPU-hour figure are not the same product.
Verdict
See the facts table above and compare with alternatives. Listed pending verification - start small and confirm a payout before scaling.
++ What works
- MIT-licensed, 10.5k stars, a genuinely novel BitTorrent-style approach to distributed LLM inference
- Historically demonstrated very large models (BLOOM 176B, Llama up to 405B) over consumer GPUs
- Free to use and can still be run as a private swarm on your own hardware
-- What to watch
- Effectively abandoned: last commit 2024-09-07, last release v2.2.0 on 2023-09-06
- Public network health endpoint (health.petals.dev) times out; the chat demo reports 'out of capacity'
- No maintenance response: the open issue 'Is this project dead?' (2026-09-04) has zero replies
- No verification, no SLA and no economic model - unsuitable for any production workload
Pricing reality check
| Documented price | see provider site |
|---|---|
| Notes | No pricing: Petals is a free volunteer/community GPU swarm, not a paid service. The project's own health endpoint (health.petals.dev) that lists available models timed out on 2026-09-24, and the chat demo reports the network is out of capacity. |
Run the long math before paying: cloud mining calculator →
// Is Petals legit?
Short answer: Based on our checks as of 2026-09-24: Petals scores 4.5/10 in our weighted review. We logged no red flags. User reports and operator background are documented with dated sources in the meta-review below.
| Status | dead |
|---|---|
| Operator | Petals (BigScience workshop research project) (medium) |
| Founded | 2022 |
| Proof | none published |
| KYC | none |
Red flags
No red flags logged as of 2026-09-24.
What "legit" does not mean
A real, operating provider can still lose you money: with a contract you own no hardware, revenue follows coin price and difficulty, and most published contracts are priced so the seller comes out ahead. "Legit" means the operator exists and pays what its terms promise - not that the deal is profitable. 6 dated sources back this page; they are listed under sources.
// Alternatives to Petals
Same category, scored on the same criteria - so the comparison stays honest. Compare only in the same unit, and read the minimum, KYC and proof columns before switching.
#2
Salad / SaladCloud
Distributed consumer-GPU cloud with published per-second prices and measured host earn rates; no token, fiat only.
no KYC
from $0.015 per GPU/hr
#3
Lambda (Lambda Labs)
Established US GPU cloud: on-demand H100/B200 instances, 1-Click Clusters and Superclusters, billed per minute with no ingress/egress charges.
from $0.690 per GPU/hr
#4
Akash Network
Open permissionless marketplace for containerised GPU/CPU compute; providers bid, tenants pay per block.
from $0.090 per GPU/hr min $0.500
// Petals FAQ
What do I actually get from Petals?
A rental of hashrate or compute. You do not own the hardware unless stated - the contract is the asset.
What is the total cost including fees?
Documented pricing: not documented - see the provider site. Fees: not documented - factor maintenance into any comparison.
Is there a minimum or a trial?
No minimum is documented - confirm with the provider.
How do payouts work?
Payout schedule is not documented.
Is it legit and verifiable?
Proof of hashrate/compute is not documented.
What happens if it shuts down?
In a contract you own no hardware, so if the operator disappears the contract is likely worthless - the pattern behind several documented collapses in our Scam Watch.
Should I use Petals or just buy the coin or hardware?
If you cannot verify a payout and a clear term, buying the coin or owning a miner leaves you in control - run the numbers with our mining calculator first.
No comments yet - be the first. Posting runs a short proof-of-work in your browser (anti-spam), no account needed.