Operator · West Palm Beach, FL · founded 2014
Vultr
US developer cloud with 33 regions selling fractional and full-GPU Cloud GPU VMs (A16, A40, L40S) and 8-GPU bare-metal servers (H100, B200, MI300X, MI325X, MI355X, A100, L40S, GH200) billed hourly.
GPU types10with a current offer
Current offers8489 readings on file · last 45 days
H100 from—per GPU-hour, on-demand
Entry price$0.47cheapest GPU-hour listed
Countries2033 named regions
Trust Score v0.175/100
- Company & track record73
- Infrastructure control100
- Pricing & billing transparency85
- Reliability signals30
- Security & compliance100
- 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.
Company & track record
weight 15%
73 / 100
| Founding year published | 2014api.vultr.com · Sep 29, 2026 | 4 / 4 |
| Headquarters published | The Constant Company, LLCapi.vultr.com · Sep 29, 2026 | 3 / 3 |
| Legal entity named | The Constant Company, LLCapi.vultr.com · Sep 29, 2026 | 4 / 4 |
| Funding or ownership disclosed | not disclosedapi.vultr.com · Sep 29, 2026 | 0 / 4 |
Infrastructure control
weight 15%
100 / 100
| Capacity model published | operates or rents dedicated capacityapi.vultr.com · Sep 29, 2026 | 10 / 10 |
| Regions published | 33 regions or countriesapi.vultr.com · Sep 29, 2026 | 5 / 5 |
Pricing & billing transparency
weight 20%
85 / 100
| Public prices without an account | yes (read on the pricing page)api.vultr.com · Sep 29, 2026 | 5 / 5 |
| Billing granularity published | per hourapi.vultr.com · Sep 29, 2026 | 4 / 4 |
| Egress policy published | metered: plan bandwidth allowance (1024 to 15360 GB/month per GPU plan) plus 2 TB free per account per month, overage 0.01 USD/GBapi.vultr.com · Sep 29, 2026 | 4 / 4 |
| Stopped-instance billing published | Billed in full: stopped Cloud GPU and Bare Metal servers keep incurring normal hourly charges (compute, storage, IP) until destroyedapi.vultr.com · Sep 29, 2026 | 4 / 4 |
| Public IPv4 policy published | not disclosedapi.vultr.com · Sep 29, 2026 | 0 / 3 |
Reliability signals
weight 20%
30 / 100
| Status page | https://status.vultr.comapi.vultr.com · Sep 29, 2026 | 6 / 6 |
| SLA published | not disclosedapi.vultr.com · Sep 29, 2026 | 0 / 6 |
| SLA with a numeric uptime commitment | not disclosedapi.vultr.com · Sep 29, 2026 | 0 / 3 |
| Uptime history published | not disclosedapi.vultr.com · Sep 29, 2026 | 0 / 5 |
| Observed availability (30 days) | not measured yet | — |
Security & compliance
weight 15%
100 / 100
| SOC 2 | Type IIapi.vultr.com · Sep 29, 2026 | 8 / 8 |
| ISO 27001 | ISO/IEC 27001:2022api.vultr.com · Sep 29, 2026 | 4 / 4 |
| HIPAA or GDPR DPA | SOC 2 Type II with HIPAA Security Ruleapi.vultr.com · Sep 29, 2026 | 3 / 3 |
Customer signal
weight 15%
not measured
| Aggregated public reviews | not collected yet | — |
provisional v0.1: evidence from the provider's own pages; dimensions not yet measured are excluded from the denominatorFull method →
Prices by GPU · 10 GPU types with a current offerWhat Vultr 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 | $2.99 | $3.17 | −6% | 4 | on-demandspot | H100 prices |
| A100 80GB | $2.40 | $1.99 | +20% | 5 | on-demandspot | A100 80GB prices |
| L40S | $1.67 | $1.26 | +32% | 5 | on-demandspot | L40S prices |
| B200 | $8.50 | $7.50 | +13% | 4 | on-demandspot | B200 prices |
| MI300X | $3.99 | $2.71 | +47% | 2 | on-demandspot | MI300X prices |
| A40 | $1.71 | $0.85 | +102% | 2 | on-demand | A40 prices |
| MI325X | $4.62 | $2.65 | +74% | 4 | on-demandspot | MI325X prices |
| GH200 | $1.99 | $3.87 | −49% | 1 | on-demand | GH200 prices |
| MI355X | $2.59 | $5.99 | −57% | 2 | on-demand | MI355X prices |
| A16 | $0.47 | — | — | 3 | on-demand | A16 prices |
All offers · 89 readings on fileEvery Vultr 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 | |
|---|---|---|---|---|---|---|---|---|
| A16 | 16× A1696 vCPU · 960 GB RAM · 1,700 GB disk | On-demand | no location can deploy it right now | Unavailable | $0.47 | $0.47 | 21:08 UTCfresh | Receipt |
| A16 | 1× A166 vCPU · 60 GB RAM · 350 GB disk | On-demand | Germany +5 | In stock | $0.47 | $0.47 | 21:08 UTCfresh | Receipt |
| A16 | 2× A1612 vCPU · 120 GB RAM · 700 GB disk | On-demand | Germany +4 | In stock | $0.47 | $0.47 | 21:08 UTCfresh | Receipt |
| A100 80GB | 8× A100 80GB224 vCPU · 2,048 GB RAM · 960 GB disk | Spot | no location can deploy it right now | Unavailable | $1.49 | $1.49 | 21:08 UTCfresh | Receipt |
| L40S | 8× L40S128 vCPU · 2,048 GB RAM · 480 GB disk | Spot | no location can deploy it right now | Unavailable | $1.49 | $1.49 | 21:08 UTCfresh | Receipt |
| L40S | 1× L40S16 vCPU · 180 GB RAM · 1,200 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.67 | $1.67 | 21:08 UTCfresh | Receipt |
| L40S | 2× L40S32 vCPU · 375 GB RAM · 2,200 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.67 | $1.67 | 21:08 UTCfresh | Receipt |
| L40S | 4× L40S64 vCPU · 750 GB RAM · 2,600 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.67 | $1.67 | 21:08 UTCfresh | Receipt |
| L40S | 8× L40S128 vCPU · 2,048 GB RAM · 480 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.67 | $1.67 | 21:08 UTCfresh | Receipt |
| A40 | 1× A4024 vCPU · 120 GB RAM · 1,400 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.71 | $1.71 | 21:08 UTCfresh | Receipt |
| A40 | 4× A4096 vCPU · 480 GB RAM · 1,400 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.71 | $1.71 | 21:08 UTCfresh | Receipt |
| MI300X | 8× MI300X256 vCPU · 2,048 GB RAM · 3,576 GB disk | Spot | no location can deploy it right now | Unavailable | $1.85 | $1.85 | 21:08 UTCfresh | Receipt |
| GH200 | 1× GH20072 vCPU · 480 GB RAM · 960 GB disk | On-demand | no location can deploy it right now | Unavailable | $1.99 | $1.99 | 21:08 UTCfresh | Receipt |
| MI325X | 8× MI325X252 vCPU · 2,872 GB RAM · 14,336 GB disk | Spot | no location can deploy it right now | Unavailable | $2.00 | $2.00 | 21:08 UTCfresh | Receipt |
| MI325X | 8× MI325X256 vCPU · 3,072 GB RAM · 3,576 GB disk | Spot | no location can deploy it right now | Unavailable | $2.00 | $2.00 | 21:08 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM216 vCPU · 1,883.9 GB RAM · 13,312 GB disk | Spot | no location can deploy it right now | Unavailable | $2.30 | $2.30 | 21:08 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM224 vCPU · 2,015.4 GB RAM · 960 GB disk | Spot | no location can deploy it right now | Unavailable | $2.30 | $2.30 | 21:08 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB12 vCPU · 120 GB RAM · 1,400 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.40 | $2.40 | 21:08 UTCfresh | Receipt |
| A100 80GB | 8× A100 80GB96 vCPU · 960 GB RAM · 2,200 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.40 | $2.40 | 21:08 UTCfresh | Receipt |
| A100 80GB | 4× A100 80GB96 vCPU · 1,024 GB RAM · 450 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.40 | $2.40 | 21:08 UTCfresh | Receipt |
| MI355X | 8× MI355X252 vCPU · 2,872 GB RAM · 14,336 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.59 | $2.59 | 21:08 UTCfresh | Receipt |
| MI355X | 8× MI355X256 vCPU · 3,072 GB RAM · 3,576 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.59 | $2.59 | 21:08 UTCfresh | Receipt |
| A100 80GB | 8× A100 80GB224 vCPU · 2,048 GB RAM · 960 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.80 | $2.80 | 21:08 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM216 vCPU · 1,883.9 GB RAM · 13,312 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.99 | $2.99 | 21:08 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM224 vCPU · 2,015.4 GB RAM · 960 GB disk | On-demand | no location can deploy it right now | Unavailable | $2.99 | $2.99 | 21:08 UTCfresh | Receipt |
32 fresh offers · the 25 cheapest shown here, all of them in the comparison tool · prices per GPU-hour · true cost includes the add-ons the provider publishesOpen in the comparison tool →
True cost & billing · read Sep 29, 2026What 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
- per hour
- Minimum charge
- Minimum billing unit is one hour; charges start at deployment
- Stopped instance
- Billed in full: stopped Cloud GPU and Bare Metal servers keep incurring normal hourly charges (compute, storage, IP) until destroyed
- Egress
- metered: plan bandwidth allowance (1024 to 15360 GB/month per GPU plan) plus 2 TB free per account per month, overage 0.01 USD/GB
- Public IPv4
- not published
- Billing model
- both
- Minimum deposit
- None with a linked credit card (0.00 USD deposit option); "a small deposit" for other payment methods, amount not stated
- Free credits
- Free credits for up to 30 days advertised in the sign-up guide; amount not stated in docs
Account
- Payment
- card · invoice · PayPal · crypto (yes: BTC, BCH, ETH, DOGE, PAX, BUSD, LTC, USDC, GUSD via BitPay)
- Sign-up
- email, password, payment method (card authorization hold or deposit)
- Identity check
- card verification
- Support
- tickets
- Enterprise support
- not published
- SLA
- not published
- Status page
- status.vultr.com ↗
- Published uptime
- not published
Infrastructure
- Regions
33 named regions in 20 countries
ams (Netherlands) · atl (United States) · blr (India) · bom (India) · cdg (France) · del (India) · dfw (United States) · ewr (United States) · fra (Germany) · hnl (United States) · icn (South Korea) · itm (Japan) · jnb (South Africa) · lax (United States) · lhr (United Kingdom) · mad (Spain) · man (United Kingdom) · mel (Australia) · mex (Mexico) · mia (United States) · mxp (Italy) · nrt (Japan) · ord (United States) · sao (Brazil) · scl (Chile) · sea (United States) · sgp (Singapore) · sjc (United States) · sto (Sweden) · syd (Australia) · tlv (Israel) · waw (Poland) · yto (Canada)- Countries
- Australia, Brazil, Canada, Chile, France, Germany, India, Israel, Italy, Japan, Mexico, Netherlands, Poland, Singapore, South Africa, South Korea, Spain, Sweden, United Kingdom, United States
- Platform
- API, CLI, serverless, SSH, persistent volumes, snapshots, spot or interruptible, bare metalSDK: Go (govultr, used by the CLI and Terraform provider); others not verifiedTerraform: officialKubernetes: managedClusters: Clusters product in docs (instance templates, cluster availability by region); interconnect specs not read
- Storage
- ephemeral · included in plan · Local disk included in each GPU plan (see includes)object · 18.00 USD/month Standard tier incl. 1000 GB, then 0.018 USD/GB; Premium 36.00 USD/month then 0.036 USD/GB; Performance 50.00 USD/month then 0.050 USD/GB; Accelerated 100.00 USD/month then 0.100 USD/GB · Each tier includes 1000 GB outbound; extra outbound 0.01 USD/GBvolume · snapshots 0.05 USD/GB-month · Block Storage per-GB price not readable (vultr.com/pricing blocked)
- Migration out
- Snapshots (0.05 USD/GB-month) and backups can be created; custom ISO upload and iPXE boot allow reproducible images; data leaves via metered egress (0.01 USD/GB beyond allowance). Instance export format for use elsewhere not documented in pages read.
Security & compliance
- SOC 2
- Type II
- ISO 27001
- ISO/IEC 27001:2022
- HIPAA
- SOC 2 Type II with HIPAA Security Rule
- GDPR / DPA
- not published
- Other
- ISO/IEC 20000-1:2018, ISO/IEC 27017:2015, ISO/IEC 27018:2019; PCI-DSS reports for select data center locations; reports via console after authentication
- Legal entity
- The Constant Company, LLC
- Funding
- not published
Read on the provider's own pages · 32 pages read · facts marked "not published" were looked for and not foundPages read ↓
Assessment · documents read Sep 29, 2026Vultr, in short
What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.
Strengths
- Widest bare-metal accelerator mix in this set: B200, H100, MI300X, MI325X, MI355X, A100 SXM/PCIe, L40S and GH200 plans in the public API
- Fractional vGPU plans from USD 0.059/hour (1/8 A16, 2 GB VRAM) for small inference or VDI workloads
- 33 regions in 20 countries from the public regions API
- Crypto accepted via BitPay (BTC, ETH, USDC and six other coins) alongside cards, PayPal, Alipay, UnionPay and wire
- SOC 2 Type II with HIPAA Security Rule, ISO/IEC 27001/27017/27018/20000-1 reports available to customers
- Official API, CLI and Terraform provider; managed Kubernetes (VKE)
- Plan catalogue with prices is readable without an API key
Weaknesses
- Stopped servers are billed at the full hourly rate until destroyed (Cloud GPU and Bare Metal)
- Hourly minimum billing unit, not per second
- Most GPU plans list no available location in the API on the check date; only A16/A40 slices and the 8x A100 SXM bare-metal plan show a region
- H100, B200, AMD Instinct, L40S and GH200 bare metal have deploy_ondemand=false in the API, so the on-demand price is published but not orderable on demand
- Only 8-GPU bare-metal configurations for H100, B200 and AMD Instinct; no single H100/H200 VM in the current API catalogue
- Egress metered beyond plan allowance at 0.01 USD/GB; Cloud GPU instances cannot be upgraded or change GPU type after deployment
Ideal for
- Teams needing fractional GPUs (A16/A40 vGPU) close to users in many regions
- Buyers who need compliance reports (SOC 2 Type II, ISO 27001, HIPAA) with a self-service account
- Full 8-GPU bare-metal nodes, including AMD Instinct MI300X/MI325X/MI355X, where stock is available
Not published
- Meaning of the bare-metal "preemptible" price fields (interruption rules, notice, or commitment): not documented on any docs.vultr.com page found
- vultr.com pricing pages blocked
- H100 form factor: API/docs say NVIDIA H100 only
- SXM appears only on blocked vultr.com product pages
- Reserved or committed-term GPU pricing: not readable (vultr.com blocked)
- Block Storage per-GB price and IPv4 pricing: not in docs pages read
- vultr.com/pricing blocked
- SLA terms: www.vultr.com/legal/sla returned HTTP 403, so no uptime percentage is recorded
- Company HQ, founding year and funding: not read on an official page
- Support tiers and response times: not documented in docs read
- Serverless Inference pricing (token-based product): not captured
- H200 and other GPUs possibly listed on vultr.com product pages but absent from the public API: not verifiable
Observations
www.vultr.com (product and pricing pages, SLA, trust center) returned a Cloudflare challenge (HTTP 403) to automated readers on the check date
The public API /v2/plans and /v2/plans-metal return hourly and monthly prices for every plan without authentication
Cloud GPU docs still describe an A100 vGPU variant that no longer appears in the vcg plan list
Pages read
- api.vultr.com/v2/plans?type=vcg&per_page=500 — Cloud GPU prices, specs, locations
- api.vultr.com/v2/plans?type=all&per_page=500 — confirms 15 GPU VM plans among 151
- api.vultr.com/v2/plans-metal?per_page=500 — bare-metal GPU prices (on-demand and preemptible fields), specs, locations
- api.vultr.com/v2/regions?per_page=500 — regions and countries
- docs.vultr.com/products/compute/instances/cloud-gpu/explore-gpu-variants — vGPU vs passthrough, GPU memory
- docs.vultr.com/products/compute/instances/bare-metal/explore-gpu-variants — bare-metal GPU list incl. A100 PCIe/SXM, B200, MI355X
- docs.vultr.com/support/platform/billing/how-am-i-billed-for-my-servers — hourly billing, 730/672 h caps, stopped servers billed
- docs.vultr.com/products/compute/instances/cloud-gpu/faq — stopped charges, no upgrade, SSH access
- docs.vultr.com/products/compute/instances/bare-metal/faq — stopped charges, custom OS via iPXE
- docs.vultr.com/support/platform/billing/do-stopped-bare-metal-servers-incur-charges — stopped billing
- docs.vultr.com/support/platform/billing/what-payment-methods-do-you-accept — cards, BitPay coins, PayPal, Alipay, UnionPay
- docs.vultr.com/platform/billing/make-account-payments — wire transfer, crypto, gift code
- docs.vultr.com/platform/billing/faq — card authorization hold, taxes, USD invoices
- docs.vultr.com/support/platform/billing/why-do-i-need-to-enter-a-payment-method — payment method used for identity verification
- docs.vultr.com/platform/create-an-account — sign-up requirements, 0 USD card link, free credits
- docs.vultr.com/platform/billing/manage-account-limits — instance and cost limits, increase requests
- docs.vultr.com/support/platform/billing/how-are-bandwidth-caps-calculated — 2 TB free per account, pooled bandwidth
- docs.vultr.com/support/platform/billing/what-is-the-bandwidth-overage-rate — 0.01 USD/GB overage
- docs.vultr.com/support/platform/billing/how-is-object-storage-billed — object storage tiers
- docs.vultr.com/support/platform/billing/does-vultr-charge-for-stored-snapshots — snapshot price
- docs.vultr.com/support/platform/compliance/how-can-i-access-vultrs-compliance-reports — SOC 2, ISO certifications
- docs.vultr.com/support/platform/compliance/where-to-learn-more-about-vultr-datacenter-compliance — PCI-DSS for select locations
- docs.vultr.com/reference/terraform — official Terraform provider, API rate limit
- docs.vultr.com/reference — CLI coverage
- docs.vultr.com/products/kubernetes/provisioning — managed Kubernetes
- docs.vultr.com/products/compute/instances/cloud-gpu/provisioning — console/API/CLI/Terraform provisioning, ISO, snapshots
- docs.vultr.com/products/compute/instances/bare-metal/provisioning — iPXE, RAID options
- docs.vultr.com/products/compute/instances/cloud-gpu/gpu-enabled-images — preinstalled drivers, CUDA, ROCm, Docker
- docs.vultr.com/products/compute/instances/cloud-gpu — legal name (footer), status link
- docs.vultr.com/platform/other/tickets — ticket support channel
- docs.vultr.com/products/serverless/inference — serverless product exists
- s24.q4cdn.com/538403808/files/doc_news/Vultr-Secures-329-Million-in-Credit-Financing-to-Expand-Global-AI-Infrastructure-and-Cloud-Computing-Platform-2025.pdf — Headquarters
AlternativesProviders people compare with it
ScalewayOperator · 5 GPU types · H100 $3.20Trust 85
Cyfuture AIOperator · 10 GPU types · H100 $3.36not scored
AceCloudOperator · 12 GPU typesTrust 73
VerdaOperator · 15 GPU types · H100 $4.79Trust 81
RuncrateReseller · 20 GPU types · H100 $2.83not scoredQuestions about Vultr
What GPUs does Vultr offer?
Vultr has 10 GPU types with a fresh offer today, for example H100 from $2.99, A100 80GB from $2.40, L40S from $1.67, B200 from $8.50, MI300X from $3.99, A40 from $1.71 per GPU-hour on-demand. The full list with prices and market medians is on this page.
How does Vultr bill?
Billing granularity: per hour. Minimum charge: Minimum billing unit is one hour; charges start at deployment Stopped instance: Billed in full: stopped Cloud GPU and Bare Metal servers keep incurring normal hourly charges (compute, storage, IP) until destroyed
How trustworthy is Vultr?
Its Trust Score is 75/100 (v0.1), read from its own documents. Strong on infrastructure control, security & compliance. Weaker on reliability signals. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.
Where are Vultr's data centres?
33 named regions in 20 countries: Australia, Brazil, Canada, Chile, France, Germany, India, Israel, Italy, Japan, Mexico, Netherlands, Poland, Singapore, South Africa, South Korea, Spain, Sweden, United Kingdom, United States.
How do you pay Vultr, and are there credits?
Payment by card, invoice, crypto. Free credits: Free credits for up to 30 days advertised in the sign-up guide; amount not stated in docs. Identity check: card verification.