io.net
US company (IO.NET Inc., Delaware) running a decentralized GPU network (DePIN, Solana-based IO Coin) that resells partner datacenter GPUs and its own worker network as pay-as-you-go VMs and containers, plus managed Ray, Kubernetes, bare metal and an inference API.
Trust Score v0.1- Company & track record73
- Infrastructure control60
- Pricing & billing transparency65
- Reliability signals0
- Security & compliance20
- Customer signal—
Scored Sep 29, 2026 from the provider's own pages. Dimensions not yet measured are excluded from the denominator. Criteria ↓
Trust Score · 5 of 6 dimensions measuredThe score, criterion by criterion
Each point comes from something the provider publishes, with the page it was read on and the date. Anything not disclosed earns zero.
| Founding year published | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 4 |
| Headquarters published | IO.NET Inc.cloud.io.net · Sep 29, 2026 | 3 / 3 |
| Legal entity named | IO.NET Inc.cloud.io.net · Sep 29, 2026 | 4 / 4 |
| Funding or ownership disclosed | Private company with a listed token (IO Coin). Team page names angel investors (Anatoly Yakovenko, Mo Shaikh, Avery Ching, Yat Siu, Sebastien Borget); institutional rounds and amounts not published on the pages read. Docs date the project to the Solana Hackathon, Feb 2023; before June 2022 the team built quantitative trading systems.cloud.io.net · Sep 29, 2026 | 4 / 4 |
| Capacity model published | third-party hostscloud.io.net · Sep 29, 2026 | 4 / 10 |
| Regions published | 8 regions or countriescloud.io.net · Sep 29, 2026 | 5 / 5 |
| Public prices without an account | yes (read on the pricing page)cloud.io.net · Sep 29, 2026 | 5 / 5 |
| Billing granularity published | mixedcloud.io.net · Sep 29, 2026 | 4 / 4 |
| Egress policy published | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 4 |
| Stopped-instance billing published | No stop/resume state documented: stopping a PAYG booking ends billing and removes the booking; prepaid time is forfeited. Termination may lose data.cloud.io.net · Sep 29, 2026 | 4 / 4 |
| Public IPv4 policy published | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 3 |
| Status page | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 6 |
| SLA published | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 6 |
| SLA with a numeric uptime commitment | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 3 |
| Uptime history published | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 5 |
| Observed availability (30 days) | not measured yet | — |
| SOC 2 | not published (footer shows an "AICPA SOC2" badge; report type and scope not stated)cloud.io.net · Sep 29, 2026 | 0 / 8 |
| ISO 27001 | not disclosedcloud.io.net · Sep 29, 2026 | 0 / 4 |
| HIPAA or GDPR DPA | GDPR badge in footer; no DPA foundcloud.io.net · Sep 29, 2026 | 3 / 3 |
| Aggregated public reviews | not collected yet | — |
What the market report notes about io.net
Our internal report lists 248 actors with their zone, segment, status and the prices it could read, dated Oct 25, 2026. Its notes are reproduced as written, in French, with their sources.
- Listed as
- io.net
What io.net charges, against the market
Cheapest current on-demand true cost here, next to the market median for the same GPU. Negative means cheaper than the market.
| GPU | Cheapest here | Market median | vs market | Current offers | Modes | |
|---|---|---|---|---|---|---|
| H100 | $1.00 | $3.17 | −68% | 9 | on-demand | H100 prices |
| A100 80GB | $1.29 | $1.99 | −35% | 10 | on-demand | A100 80GB prices |
| H200 | $2.00 | $4.76 | −58% | 1 | on-demand | H200 prices |
| RTX PRO 6000 | $2.06 | $2.14 | −4% | 6 | on-demand | RTX PRO 6000 prices |
| L40S | $0.84 | $1.26 | −34% | 6 | on-demand | L40S prices |
| B200 | $4.89 | $7.50 | −35% | 1 | on-demand | B200 prices |
| B300 | $4.50 | $8.96 | −50% | 2 | on-demand | B300 prices |
| L4 | $0.90 | $0.80 | +13% | 4 | on-demand | L4 prices |
| RTX A6000 | $0.49 | $0.54 | −9% | 4 | on-demand | RTX A6000 prices |
| RTX 5090 | $0.89 | $0.69 | +29% | 1 | on-demand | RTX 5090 prices |
| RTX 4090 | $0.30 | $0.44 | −32% | 2 | on-demand | RTX 4090 prices |
| T4 | $0.47 | $0.51 | −7% | 2 | on-demand | T4 prices |
| A100 40GB | $0.79 | $1.98 | −60% | 2 | on-demand | A100 40GB prices |
| RTX 6000 Ada | $0.68 | $0.79 | −13% | 7 | on-demand | RTX 6000 Ada prices |
| L40 | $0.66 | $0.89 | −25% | 3 | on-demand | L40 prices |
| Tesla V100 32GB | $1.26 | $2.48 | −49% | 5 | on-demand | Tesla V100 32GB prices |
| RTX 3090 | $0.25 | $0.19 | +32% | 1 | on-demand | RTX 3090 prices |
| A30 | $1.42 | $0.77 | +83% | 3 | on-demand | A30 prices |
| RTX A4000 | $0.15 | $0.19 | −19% | 1 | on-demand | RTX A4000 prices |
| RTX A5000 | $0.19 | $0.27 | −29% | 1 | on-demand | RTX A5000 prices |
| A40 | $1.51 | $0.85 | +78% | 1 | on-demand | A40 prices |
| RTX 4090 D | $0.30 | — | — | 1 | on-demand | RTX 4090 D prices |
Every io.net offer, with its receipt
Current offers (last 45 days) ranked by true cost, all rental modes, each with its read date. Older readings stay listed below, in grey.
| GPU | Configuration | Mode | Region | Stock | List | True cost | Read | |
|---|---|---|---|---|---|---|---|---|
| RTX A4000 | 1× RTX A4000 | On-demand | io.net network (location not stated) | Unavailable | $0.15 | $0.15 | 21:10 UTCfresh | Receipt |
| RTX A5000 | 1× RTX A5000 | On-demand | io.net network (location not stated) | Unavailable | $0.19 | $0.19 | 21:10 UTCfresh | Receipt |
| RTX 3090 | 1× RTX 3090 | On-demand | io.net network (location not stated) | Unavailable | $0.25 | $0.25 | 21:10 UTCfresh | Receipt |
| RTX 4090 D | 1× RTX 4090 D | On-demand | io.net network (location not stated) | Unavailable | $0.30 | $0.30 | 21:10 UTCfresh | Receipt |
| RTX 4090 | 1× RTX 4090 | On-demand | io.net network (location not stated) | Unavailable | $0.30 | $0.30 | 21:10 UTCfresh | Receipt |
| RTX 4090 | 1× RTX 4090 | On-demand | io.net network (location not stated) | Unavailable | $0.30 | $0.30 | 21:10 UTCfresh | Receipt |
| T4marketplace floor | 2× T424 vCPU · 100 GB RAM · 1,800 GB disk | On-demand | India | In stock | $0.47 | $0.47 | 21:10 UTCfresh | Receipt |
| T4 | 1× T412 vCPU · 50 GB RAM · 900 GB disk | On-demand | India | In stock | $0.47 | $0.47 | 21:10 UTCfresh | Receipt |
| RTX A6000marketplace floor | 8× RTX A600062 vCPU · 384 GB RAM · 2,560 GB disk | On-demand | US | In stock | $0.49 | $0.49 | 21:10 UTCfresh | Receipt |
| RTX A6000 | 2× RTX A600014 vCPU · 96 GB RAM · 512 GB disk | On-demand | US | In stock | $0.49 | $0.49 | 21:10 UTCfresh | Receipt |
| RTX A6000 | 1× RTX A60006 vCPU · 48 GB RAM · 256 GB disk | On-demand | US | In stock | $0.49 | $0.49 | 21:10 UTCfresh | Receipt |
| L40marketplace floor | 4× L4050 vCPU · 288 GB RAM · 2,500 GB disk | On-demand | US | In stock | $0.66 | $0.66 | 21:10 UTCfresh | Receipt |
| L40 | 2× L4026 vCPU · 144 GB RAM · 1,250 GB disk | On-demand | US | In stock | $0.66 | $0.66 | 21:10 UTCfresh | Receipt |
| L40 | 1× L4014 vCPU · 72 GB RAM · 625 GB disk | On-demand | US | In stock | $0.66 | $0.66 | 21:10 UTCfresh | Receipt |
| RTX 6000 Adamarketplace floor | 8× RTX 6000 Ada104 vCPU · 640 GB RAM · 2,800 GB disk | On-demand | US | In stock | $0.68 | $0.68 | 21:10 UTCfresh | Receipt |
| RTX 6000 Ada | 4× RTX 6000 Ada52 vCPU · 288 GB RAM · 1,400 GB disk | On-demand | US | In stock | $0.68 | $0.68 | 21:10 UTCfresh | Receipt |
| RTX 6000 Ada | 2× RTX 6000 Ada26 vCPU · 144 GB RAM · 700 GB disk | On-demand | US | In stock | $0.68 | $0.68 | 21:10 UTCfresh | Receipt |
| RTX 6000 Ada | 1× RTX 6000 Ada12 vCPU · 72 GB RAM · 350 GB disk | On-demand | US | In stock | $0.68 | $0.68 | 21:10 UTCfresh | Receipt |
| A100 40GB | 1× A100 40GB | On-demand | io.net network (location not stated) | Unavailable | $0.79 | $0.79 | 21:10 UTCfresh | Receipt |
| L40Smarketplace floor | 8× L40S94 vCPU · 576 GB RAM · 5,000 GB disk | On-demand | US | In stock | $0.84 | $0.84 | 21:10 UTCfresh | Receipt |
| L40S | 4× L40S46 vCPU · 288 GB RAM · 2,500 GB disk | On-demand | US | In stock | $0.84 | $0.84 | 21:10 UTCfresh | Receipt |
| L40S | 2× L40S24 vCPU · 144 GB RAM · 1,250 GB disk | On-demand | US | In stock | $0.84 | $0.84 | 21:10 UTCfresh | Receipt |
| L40S | 1× L40S | On-demand | io.net network (location not stated) | Unavailable | $0.89 | $0.89 | 21:10 UTCfresh | Receipt |
| RTX 5090 | 1× RTX 5090 | On-demand | io.net network (location not stated) | Unavailable | $0.89 | $0.89 | 21:10 UTCfresh | Receipt |
| L40S | 1× L40S | On-demand | io.net network (location not stated) | Unavailable | $0.89 | $0.89 | 21:10 UTCfresh | Receipt |
What the hourly price does not say
Billing granularity, minimums, what a stopped machine costs, traffic, addresses, deposits and identity checks, as published by the provider.
Billing
- Billing granularity
- mixed
- Minimum charge
- 1 hour charged upfront at deployment, not refunded if you stop the booking (Terms 4.2, PAYG docs). After that, billed in advance in 5-minute blocks, including idle time.
- Stopped instance
- No stop/resume state documented: stopping a PAYG booking ends billing and removes the booking; prepaid time is forfeited. Termination may lose data.
- Egress
- not published
- Public IPv4
- not published
- Billing model
- prepaid credits
- Minimum deposit
- 10 USD minimum IO Credits purchase. Purchased credits expire 12 months after purchase.
- Free credits
- not published
Account
- Payment
- card · crypto (yes: USDC and IO Coin via SpherePay (Solana or Aptos wallet), used to buy IO Credits; no fee on crypto top-ups, 5% fee on card top-ups via Stripe)
- Sign-up
- email (one-time passcode) or Google, Apple ID, GitHub, Hugging Face, X or Worldcoin login, age 18+ and not on US/EU/UK sanctions lists (Terms)
- Identity check
- none
- Support
- email ([email protected]), contact form, chat (contact page says 24/7)
- Enterprise support
- not published
- SLA
- not published
- Status page
- not published
- Published uptime
- not published
Infrastructure
- Regions
8 named regions in 4 countries
Partner listings · United States (United States) · Partner listings · Finland (Finland) · Partner listings · India (hardware ids reference Delhi) (India) · Partner listings · Poland (Poland)- Countries
- United States, Finland, India, Poland
- Platform
- api, jupyter, ssh, custom_docker_image, bare_metal · SDK : — · Terraform : — · Kubernetes : managed (on request via sales · self-service withdrawn) · interconnexion : — · clusters : Ray, Kubernetes, Bare Metal and Confidential Compute (Intel TDX / AMD SEV-SNP on H100, H200, B200) as managed services on request via [email protected] · self-service Bare Metal withdrawn 1 Oct 2025.
- Storage
- ephemeral · included in the hourly price · Each listing shows local storage (250 GB to 5.7 TB). No separately priced persistent volume, network or object storage found in the IO Cloud docs.
- Migration out
- No volumes or snapshots to export: data lives on the booking's local storage and must be copied out (SSH/scp, object storage of the customer's choice) before stopping, because stopping or credit exhaustion removes the booking. Container workloads use standard Docker images and move as-is; purchased IO Credits can be withdrawn after manual finance review (3 to 7 business days).
Security & compliance
- SOC 2
- not published (footer shows an "AICPA SOC2" badge; report type and scope not stated)
- ISO 27001
- not published
- HIPAA
- not published
- GDPR / DPA
- GDPR badge in footer; no DPA found
- Other
- Confidential Compute docs claim workloads comply with HIPAA, GDPR and PCI DSS, without an audit report. FAQ says io.net prioritises SOC2-compliant suppliers.
- Legal entity
- IO.NET Inc.
- Funding
- Private company with a listed token (IO Coin). Team page names angel investors (Anatoly Yakovenko, Mo Shaikh, Avery Ching, Yat Siu, Sebastien Borget); institutional rounds and amounts not published on the pages read. Docs date the project to the Solana Hackathon, Feb 2023; before June 2022 the team built quantitative trading systems.
io.net, in short
What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.
Strengths
- Large self-service catalogue visible without an account: 45 partner VM configurations plus 17 io.net-network devices, from RTX A4000 at 0.15 USD/h to 8x A100 and 8x H100 nodes.
- Pays in crypto (USDC, IO Coin via SpherePay) with no top-up fee; cards via Stripe in local currency.
- Sign-up with email one-time passcode or social login; no KYC documented and the FAQ presents the absence of KYC as a feature.
- Public REST API for VMs (VMaaS) and containers (CaaS), including hardware list, price check and burn-rate endpoints.
- Confidential Compute on H100/H200/B200 (Intel TDX, AMD SEV-SNP) with attestation guides, as a managed service.
Weaknesses
- Dynamic pricing: the rate of a running booking can change when the supplier changes its price (Terms 4.3); prices moved between two reads on the same day.
- One hour charged upfront at each deployment and not refunded; all charges final (Terms 4.4). Purchased credits expire after 12 months.
- Card top-ups carry a 5% fee.
- Cheapest data-centre GPUs (io.net network H100 at 1.99, H200 at 2.00, B200 at 4.89 USD/h) were all sold out on the check date; available H100 started at 2.60 USD/h (1 GPU, US partner).
- No stop/resume, no persistent volumes and no snapshots documented: terminating a booking removes it and may lose data.
- No SLA, no public status page, no published SOC 2 report; partner-supplied VMs use a preloaded image that cannot be changed.
- Documentation contradicts itself on payments (older pages describe direct USDC/IO Coin payment per cluster, newer pages say compute is paid only from IO Credits).
Ideal for
- Crypto-native teams wanting to pay in USDC or IO Coin without KYC.
- Short experiments on older or mid-range GPUs (A6000, A100, V100, L40) where a broad partner catalogue helps.
- Buyers needing confidential computing on H100/H200/B200 through a managed engagement.
Not published
- vCPU and RAM per listing: shown only behind "More details" in the console
- the VMaaS hardware API that returns them requires an API key (HTTP 401 without one).
- Egress and IPv4 pricing: not published.
- SLA and uptime: not published
- no status page found.
- SOC 2: only a footer badge
- report type, scope and date not published.
- Form factor (SXM or PCIe) of the partner "H100" and several "A100" 80GB listings is not stated
- recorded as model "other".
- Whether the IO Explorer "Ray" and "BAREMETAL" rates are still bookable: docs say both are now managed services on request.
- The "$0.74 / hr" H100 figure on io.net/cloud: the widget does not say which provider it refers to.
- Company founding year, headquarters city and funding totals: not stated on the pages read.
- catalogue_count counts 27 NVIDIA models (B300, B200, H200, H100 SXM, H100 PCIe, A100 SXM 80GB, A100 PCIe 80GB, A100 40GB, L40S, L40, L4, A40, A30, T4, V100, RTX PRO 6000, RTX 6000 Ada, RTX A6000, RTX A5000, RTX A4000, RTX 5090, RTX 4090, RTX 4090 D, RTX 3090, RTX 4070 Ti, RTX 4060 Ti, RTX 3060)
- the six Apple M3/M4 chips in the IO Explorer are excluded.
Observations
io.net/pricing returns a 404; prices live in the cloud console (cloud.io.net), the home page carousel and the IO Explorer, all visible without login.
Home hero claims "H100s from $2.19/hr vs AWS at $12.29/hr"; no current listing at 2.19 was found (live H100 listings: 2.60 partner, 1.99 io.net network sold out).
Home page also carries a price table labelled "Prices correct as of 30th July 2025" (RTX 4090, H100 PCIe, H100 SXM5, H200 by VM / CaaS / Bare Metal), older than the live console prices.
Console cards show a large discount against a struck-through list price on some items, e.g. H200 "-94%" from $63.30/hr to $3.79/hr, B200 "-44%" from $14.24/hr.
The console's default view hides sold-out items; with "Show unavailable" all 17 io.net-network devices appear as "Sold Out" with a "Request" button.
Prices are per node (per VM) with GPU count shown as x1 to x8; vCPU/RAM appear only under "More details".
io.net/cloud has a "Compare compute prices" widget showing "Cheapest: H100 $0.74 / hr" without naming the providers compared.
Footer shows AICPA SOC2 and GDPR badges and "All services are online" without a link to a status page.
A cookie/consent overlay from NextRoll with 22 advertising partners covers the console.
Pages read
- cloud.io.net/cloud/virtual-machines/create-cluster — Live VM prices, GPU counts, storage, supplier, location
- cloud.io.net/cloud/container/create-cluster — Container prices, recipe templates
- explorer.io.net/explorer/home — Ray, Container, Baremetal, VM rates for io.net network devices; network size
- io.net/ — Carousel prices, 30 July 2025 price table, hero claims, SOC2/GDPR badges
- io.net/cloud — Site observations
- io.net/pricing — Confirmed no pricing page
- io.net/docs/guides/payment/io-cloud-payments — Credits-only compute, 5% card fee, no crypto fee, Solana/Aptos wallets
- io.net/docs/guides/payment/io-credits — 10 USD minimum, USDC via Spherepay, card via Stripe, 12-month expiry, withdrawals
- io.net/docs/guides/payment/pay-as-you-go — First hour upfront, incremental charging, termination on low balance, burn-rate API
- io.net/docs/guides/payment/card-payments-faqs — 5% Stripe fee, local currency
- io.net/docs/guides/payment/deploy-cluster-and-pay — IO Coin and USDC via Sphere Pay (legacy per-cluster flow
- io.net/terms — Legal entity, 5-minute block billing, dynamic pricing, no refunds, sanctions/age eligibility
- io.net/docs/guides/loginsignup — Email OTP and social sign-up
- io.net/docs/guides/clouds/start-using-io-cloud — Sign-up methods, cluster types, Jupyter/VS Code tools
- io.net/docs/guides/faq — No KYC positioning, supply sources, SOC2 suppliers
- io.net/docs/guides/clouds/deploy-vm-on-demand — Images, SSH, partner image restriction, ports
- io.net/docs/guides/clouds/deploy-containers — Custom Docker images, 1-hour minimum
- io.net/docs/guides/clouds/managed-services — Bare Metal, Kubernetes, Ray, Confidential Compute on request; bare metal self-service withdrawn 1 Oct 2025
- io.net/docs/guides/clouds/confidential-compute-overview — TDX/SEV-SNP, H100/H200/B200, compliance claims
- io.net/docs/reference/vmaas/get-started-with-vmaas-api — API availability and authentication
- io.net/docs/reference/vmaas/get-hardware-list — Hardware fields (price per hour USD, vCPU, memory, nvlink
- io.net/team — Angel investors, leadership
- io.net/contact-us — Support channels
Questions about io.net
What GPUs does io.net offer?
io.net has 32 GPU types with a fresh offer today, for example H100 from $1.00, A100 80GB from $1.29, RTX PRO 6000 from $2.06, H200 from $2.00, L40S from $0.84, B200 from $4.89 per GPU-hour on-demand. The full list with prices and market medians is on this page.
How does io.net bill?
Billing granularity: mixed. Minimum charge: 1 hour charged upfront at deployment, not refunded if you stop the booking (Terms 4.2, PAYG docs). After that, billed in advance in 5-minute blocks, including idle time. Stopped instance: No stop/resume state documented: stopping a PAYG booking ends billing and removes the booking; prepaid time is forfeited. Termination may lose data.
How trustworthy is io.net?
Its Trust Score is 42/100 (v0.1), read from its own documents. Weaker on infrastructure control, pricing & billing transparency, reliability signals, security & compliance. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.
Where are io.net's data centres?
8 named regions in 4 countries: United States, Finland, India, Poland.
How do you pay io.net, and are there credits?
Payment by card, crypto. No sign-up credit is published. Identity check: none.
