Operator · London, UK · founded 2020
Hyperstack
Hyperstack provides GPU cloud computing with high-performance NVIDIA GPUs for AI, rendering, and compute workloads.
GPU types10with a current offer
Current offers173184 readings on file · last 45 days
H100 from$2.50per GPU-hour, on-demand
Entry price$1.00cheapest GPU-hour listed
Countries33 named regions
Trust Score v0.181/100
- Company & track record73
- Infrastructure control100
- Pricing & billing transparency100
- Reliability signals75
- Security & compliance53
- 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 | 2020hyperstack.cloud · Sep 29, 2026 | 4 / 4 |
| Headquarters published | NexGen Cloud Limitedhyperstack.cloud · Sep 29, 2026 | 3 / 3 |
| Legal entity named | NexGen Cloud Limitedhyperstack.cloud · Sep 29, 2026 | 4 / 4 |
| Funding or ownership disclosed | not disclosedhyperstack.cloud · Sep 29, 2026 | 0 / 4 |
Infrastructure control
weight 15%
100 / 100
| Capacity model published | operates or rents dedicated capacityhyperstack.cloud · Sep 29, 2026 | 10 / 10 |
| Regions published | 3 regions or countrieshyperstack.cloud · Sep 29, 2026 | 5 / 5 |
Pricing & billing transparency
weight 20%
100 / 100
| Public prices without an account | yes (read on the pricing page)hyperstack.cloud · Sep 29, 2026 | 5 / 5 |
| Billing granularity published | per minutehyperstack.cloud · Sep 29, 2026 | 4 / 4 |
| Egress policy published | included (pricing page: Egress/Ingress traffic Free; FAQ: no bandwidth or data transfer charges)hyperstack.cloud · Sep 29, 2026 | 4 / 4 |
| Stopped-instance billing published | SHUTOFF (stopped) VMs are billed in full, hardware stays reserved. HIBERNATED VMs: GPU/CPU/RAM not billed; root disk copied to Cloud-SSD and billed at 0.000096774 USD/GB-hour (about 0.07 USD/GB-month); public IP billed only if retained; attached volumes keep billing. Ephemeral disk is wiped on hibernation.hyperstack.cloud · Sep 29, 2026 | 4 / 4 |
| Public IPv4 policy published | extra: 0.00672043 USD/hour per public IP (docs: about 5 USD/month); spot discounts do not apply to IPhyperstack.cloud · Sep 29, 2026 | 3 / 3 |
Reliability signals
weight 20%
75 / 100
| Status page | https://status.hyperstack.cloud/hyperstack.cloud · Sep 29, 2026 | 6 / 6 |
| SLA published | 99.5% monthly uptime (Service Level Addendum v1, 15 Aug 2023) on GPU, CPU, memory and storage services; pro-rata account credit on claim within 5 days; spot VMs excludedhyperstack.cloud · Sep 29, 2026 | 6 / 6 |
| SLA with a numeric uptime commitment | 99.5% monthly uptime (Service Level Addendum v1, 15 Aug 2023) on GPU, CPU, memory and storage services; pro-rata account credit on claim within 5 days; spot VMs excludedhyperstack.cloud · Sep 29, 2026 | 3 / 3 |
| Uptime history published | not disclosedhyperstack.cloud · Sep 29, 2026 | 0 / 5 |
| Observed availability (30 days) | not measured yet | — |
Security & compliance
weight 15%
53 / 100
| SOC 2 | Type IIhyperstack.cloud · Sep 29, 2026 | 8 / 8 |
| ISO 27001 | in progress (security page lists ISO 27001 as upcoming)hyperstack.cloud · Sep 29, 2026 | 0 / 4 |
| HIPAA or GDPR DPA | planned (listed as upcoming)hyperstack.cloud · Sep 29, 2026 | 0 / 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 →
Market report · Oct 25, 2026 · Specialised GPU clouds, hourly rental, pods and inference · #43 of the 200 to audit · prices read on the provider's pageWhat the market report notes about Hyperstack
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
- Hyperstack
- Cross-checkfr
- B200 6,00 $ à la demande (catalogue 192 Go) ; on-demand, réservé et spot.
Prices by GPU · 9 GPU types with a current offerWhat Hyperstack 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 | $2.50 | $3.17 | −21% | 7 | on-demandreservedspot | H100 prices |
| A100 80GB | $1.35 | $1.99 | −32% | 7 | on-demandreservedspot | A100 80GB prices |
| RTX PRO 6000 | $1.85 | $2.14 | −14% | 3 | on-demandreservedspot | RTX PRO 6000 prices |
| H200 | $3.99 | $4.76 | −16% | 2 | on-demandreserved | H200 prices |
| B200 | $6.00 | $7.50 | −20% | 2 | on-demandreserved | B200 prices |
| B300 | $7.40 | $8.96 | −17% | 1 | on-demand | B300 prices |
| RTX A6000 | $0.50 | $0.54 | −7% | 3 | on-demandreservedspot | RTX A6000 prices |
| L40 | $1.00 | $0.89 | +13% | 3 | on-demandreservedspot | L40 prices |
| RTX A4000 | $0.15 | $0.19 | −19% | 2 | on-demandreserved | RTX A4000 prices |
All offers · 184 readings on fileEvery Hyperstack 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 | Reserved | NORWAY-1: 0 | Unavailable | $0.11 | $0.11 | 21:10 UTCfresh | Receipt |
| RTX A4000 | 1× RTX A4000 | On-demand | NORWAY-1: 0 | Unavailable | $0.15 | $0.15 | 21:10 UTCfresh | Receipt |
| RTX A6000 | 1× RTX A6000 | Reserved | CANADA-1: 0 | Unavailable | $0.35 | $0.35 | 21:10 UTCfresh | Receipt |
| RTX A6000 | 1× RTX A6000 | Spot | CANADA-1: 9 | In stock | $0.40 | $0.40 | 21:10 UTCfresh | Receipt |
| RTX A6000 | 1× RTX A6000 | On-demand | CANADA-1: 0 | Unavailable | $0.50 | $0.50 | 21:10 UTCfresh | Receipt |
| L40 | 1× L40 | Reserved | CANADA-1: 25+ | In stock | $0.70 | $0.70 | 21:10 UTCfresh | Receipt |
| L40 | 1× L40 | Spot | CANADA-1: 1 | In stock | $0.80 | $0.80 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | Reserved | CANADA-1: 0 | Unavailable | $0.95 | $0.95 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | Reserved | CANADA-1: 0 | Unavailable | $0.98 | $0.98 | 21:10 UTCfresh | Receipt |
| L40 | 1× L40 | On-demand | CANADA-1: 25+ | In stock | $1.00 | $1.00 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | Spot | CANADA-1: 5 | In stock | $1.08 | $1.08 | 21:10 UTCfresh | Receipt |
| RTX PRO 6000 | 1× RTX PRO 6000 | Reserved | — | Stock unknown | $1.30 | $1.30 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | On-demand | CANADA-1: 0 | Unavailable | $1.35 | $1.35 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | Reserved | US-1: 0 | Unavailable | $1.36 | $1.36 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | On-demand | CANADA-1: 0 | Unavailable | $1.40 | $1.40 | 21:10 UTCfresh | Receipt |
| RTX PRO 6000 | 1× RTX PRO 6000 | Spot | — | Stock unknown | $1.48 | $1.48 | 21:10 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | On-demand | US-1: 0 | Unavailable | $1.60 | $1.60 | 21:10 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | Reserved | CANADA-1: 3 | In stock | $1.75 | $1.75 | 21:10 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | Reserved | CANADA-1: 0 | Unavailable | $1.82 | $1.82 | 21:10 UTCfresh | Receipt |
| RTX PRO 6000 | 1× RTX PRO 6000 | On-demand | — | Stock unknown | $1.85 | $1.85 | 21:10 UTCfresh | Receipt |
| H100PCIE | 1× H100 PCIE | Spot | CANADA-1: 2 | In stock | $2.00 | $2.00 | 21:10 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | On-demand | CANADA-1: 3 | In stock | $2.50 | $2.50 | 21:10 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | On-demand | CANADA-1: 0 | Unavailable | $2.60 | $2.60 | 21:10 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | Reserved | CANADA-1: 0 | Unavailable | $2.72 | $2.72 | 21:10 UTCfresh | Receipt |
| H200SXM | 1× H200 SXM | Reserved | CANADA-1: 0 | Unavailable | $2.79 | $2.79 | 21:10 UTCfresh | Receipt |
30 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 minute
- Minimum charge
- No per-launch minimum published. To deploy a VM, the balance must cover all running resources plus the new VM for at least one hour.
- Stopped instance
- SHUTOFF (stopped) VMs are billed in full, hardware stays reserved. HIBERNATED VMs: GPU/CPU/RAM not billed; root disk copied to Cloud-SSD and billed at 0.000096774 USD/GB-hour (about 0.07 USD/GB-month); public IP billed only if retained; attached volumes keep billing. Ephemeral disk is wiped on hibernation.
- Egress
- included (pricing page: Egress/Ingress traffic Free; FAQ: no bandwidth or data transfer charges)
- Public IPv4
- extra: 0.00672043 USD/hour per public IP (docs: about 5 USD/month); spot discounts do not apply to IP
- Billing model
- both
- Minimum deposit
- not published
- Free credits
- Not published as a standing offer; vouchers can be redeemed and the T&Cs say test credits are offered occasionally
Account
- Payment
- card · invoice
- Sign-up
- email and password, or SSO (Google, Microsoft, GitHub), billing information to activate the account (organisation name, address, tax details), card via Stripe with 3D Secure, know-your-customer information per T&Cs clause 7.1, age 18+
- Identity check
- card verification
- Support
- email, ticket, chat, phone (sales)
- Enterprise support
- not published
- SLA
- 99.5% monthly uptime (Service Level Addendum v1, 15 Aug 2023) on GPU, CPU, memory and storage services; pro-rata account credit on claim within 5 days; spot VMs excluded
- Status page
- status.hyperstack.cloud ↗
- Published uptime
- not published
Infrastructure
- Regions
3 named regions in 3 countries
CANADA-1 (Canada) · NORWAY-1 (Norway) · US-1 (United States)- Countries
- Canada, Norway, United States
- Platform
- api, ssh, persistent_volumes, snapshots, spot_or_interruptible · SDK : python, go, javascript (all alpha) · Terraform : official · Kubernetes : managed · interconnexion : NVLink on SXM and PCIe-NVLink flavors · SR-IOV inter-VM networking up to 350 Gbps in network-optimized environments (CANADA-1, US-1), 125 Gbps otherwise on those flavors, 16 Gbps default vNIC · clusters : Dedicated GPU clusters on request from the console (minimum 2 GPUs, priced by sales) · GB200 NVL72 and GB300 NVL72 via reservation / Secure Private Cloud
- Storage
- ephemeral · included · Local NVMe root disk and ephemeral disk included in GPU flavors (sizes fixed per flavor, e.g. 100 GB root + 32,000 GB ephemeral on 8x H100 SXM CANADA-1)volume · 0.000096774 USD/GB-hour (docs: about 0.07 USD/GB-month; pricing page: about 0.10 USD per TB per hour) · Shared Storage Volumes (Cloud-SSD, network block storage); billed while AVAILABLE, IN-USE or RESERVED; snapshots billed at the same rate on consumed storageobject · 0.000019397 USD/GB-hour (about 14.29 USD/TB-month) · S3-compatible, CANADA-1 only, no geographic replication
- Migration out
- VMs use standard Linux images; data can be moved over SSH/rsync (egress free). Shared Storage Volumes and snapshots persist independently of VMs, custom images can be created, and object storage is S3-compatible, so data can be pulled with standard S3 tools. No documented image export format. Ephemeral disks are wiped on hibernation/deletion, so data must be copied before leaving.
Security & compliance
- SOC 2
- Type II
- ISO 27001
- in progress (security page lists ISO 27001 as upcoming)
- HIPAA
- planned (listed as upcoming)
- GDPR / DPA
- Data Processing Agreement referenced in T&Cs; UK/EU data residency options stated
- Other
- SOC 2 held by parent NexGen Cloud; Tier 3 data centers; bug bounty programme
- Legal entity
- NexGen Cloud Limited
- Funding
- not published
GPUBank market study, read on the provider's own pages · 30 pages read · facts marked "not published" were looked for and not foundPages read ↓
From the study · assessmentHyperstack, in short
What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.
Strengths
- Per-GPU list prices published for 13 GPU SKUs from RTX A4000 (0.15 USD/h) to B300 (7.40 USD/h), with separate reservation and spot tables
- Egress and ingress traffic free; CPU, RAM, root and ephemeral NVMe included in GPU flavor price
- Hibernation stops GPU billing while keeping the root disk (about 0.07 USD/GB-month)
- SR-IOV networking up to 350 Gbps between VMs in CANADA-1 and US-1; 8x SXM and NVLink flavors
- Published SLA (99.5%) with account-credit rebates and a public status page
- SOC 2 Type II attestation (parent NexGen Cloud)
- Managed on-demand Kubernetes with free master node, Terraform provider, Python/Go/JS SDKs, MCP servers
- Pricebook API returns the account's billed rate per resource
Weaknesses
- On-demand H100 SXM is 3.20 USD/GPU-h and H200 3.99 USD/GPU-h, both only as 8x flavors
- Stopped (SHUTOFF) VMs are billed in full; only hibernation reduces cost
- Public IPv4 billed separately (0.00672043 USD/h)
- Card payment through Stripe only for self-serve; no crypto; T&Cs require KYC information
- Terraform provider and SDKs are labelled alpha
- Three regions, none in the EU (Norway region only serves RTX A4000 flavors); most GPUs only in CANADA-1
- Live stock page on 2026-09-29 showed B300, H200, H100 SXM, H100 PCIe and H100 NVLink on-demand out of stock in CANADA-1
- Reservation prices are 'starting from' with no published term; SLA uptime level 99.5% is lower than many cloud SLAs
- Spot VMs have no termination notice and are not covered by the SLA
Ideal for
- Teams wanting per-minute VMs with free egress and published per-GPU list prices in North America and Norway
- Users who pause work often and can use hibernation to stop GPU billing
- Multi-VM training needing SR-IOV high-speed networking without a bare-metal contract
- Buyers requiring a SOC 2 Type II supplier and card or invoice billing
Not published
- Spot prices for H100 SXM, H100 NVLink and A100 SXM: spot flavors exist (flavors/stock pages) but the pricing page spot table does not list them
- the Spot VMs page figures conflict with current on-demand prices, so they were not recorded
- Reservation term lengths and minimum quantities behind 'Starting from' prices not published
- B300 reservation price and GB200/GB300 NVL72 prices not published (sales only)
- Minimum top-up amount not published
- No official CLI documented (API, SDKs, Terraform, MCP server only)
- Whether identity documents are requested beyond the T&Cs 'know-your-customer information' clause is not documented
- Exact data-center cities not published (only Quebec, Vestland and Texas)
- Funding and ownership details of NexGen Cloud not read on an official page
- Support gating (minimum spend for tickets/chat) not published
Observations
Pricing page shows per-GPU hourly prices without an account in three tables: On-Demand (with VRAM, max pCPUs per GPU, max RAM per GPU), Reservation ('Starting from', with 'Reserve here' links) and Spot VM
Header claims 'Up to 75% more cost-effective than Hyperscalers' and states billing per minute
B200 and B300 rows appear at the bottom of the on-demand table after A4000; a site banner says 'NVIDIA B300s are coming to Hyperstack - On-Demand in August, reserved private clusters in Q4'
B300 has no reservation price; the spot table lists only 5 models (H100 PCIe, RTX Pro 6000 SE, A100, L40, A6000)
Storage row priced as '$0.000096774 per GB per hour (approximately $0.10 per TB per hour)'; public IP as '0.00672043 per hour' without a currency symbol; egress/ingress 'Free'; Kubernetes 'Free master node'
Payment method only visible by expanding the FAQ ('We process payments using Stripe ... add your credit card')
Same page has an AI Studio tab with per-million-token prices (e.g. Llama 3.3 70B 0.80 USD in/out) and fine-tuning at 0.063 USD per minute
The separate Spot VMs page shows older, conflicting prices (H100 SXM on-demand 2.4 / spot 1.92, H200 3.5 / 2.80) that do not match the pricing page (H100 SXM 3.20, H200 3.99)
Home page repeats the on-demand and reservation table but omits B200 and B300
GB200 NVL72 and GB300 NVL72 product pages show no price, only 'Reserve now' (Secure Private Cloud)
Pages read
- hyperstack.cloud/gpu-pricing — on-demand, reservation and spot GPU prices, storage, IP, egress, CPU flavors, AI Studio prices, FAQ (Stripe payments, per-minute billing, balance drain
- hyperstack.cloud/ — site observations, regions claim, product list
- docs.hyperstack.cloud/docs/hardware/flavors — flavor catalogue: GPU counts, CPU, RAM, disks, regions, spot flavors, networking
- docs.hyperstack.cloud/docs/billing/pricebook — per-GPU billing model, included CPU/RAM/disk, hibernation billing, Pricebook API
- docs.hyperstack.cloud/docs/billing/billing-policies — sign-up methods, prepaid/postpaid, per-minute billing, Stripe 3DS, minimum balance rule, low-balance actions
- docs.hyperstack.cloud/docs/billing/states-and-billing — stopped vs hibernated billing, root disk and IP rates
- docs.hyperstack.cloud/docs/billing/snapshots — snapshot storage billing
- docs.hyperstack.cloud/docs/billing/contracts — reservation contracts
- docs.hyperstack.cloud/docs/resource-management/regions — regions, locations, Tier 3, SOC 2 Type 2, SR-IOV availability
- docs.hyperstack.cloud/docs/virtual-machines/spot — spot terms and limitations
- hyperstack.cloud/spot-vms — conflicting spot price table (site observation only
- docs.hyperstack.cloud/docs/storage/object-storage — object storage price and region
- docs.hyperstack.cloud/docs/network/public-ip — public IP billing and behaviour
- docs.hyperstack.cloud/docs/network/high-speed-networking — interconnect
- docs.hyperstack.cloud/docs/hardware/gpu-stock-information — live stock on 2026-09-29, spot flavors in stock
- docs.hyperstack.cloud/docs/hardware/gpu-cluster-request — clusters
- docs.hyperstack.cloud/docs/libraries/ — SDKs and Terraform (alpha
- docs.hyperstack.cloud/docs/libraries/terraform — Terraform
- docs.hyperstack.cloud/docs/kubernetes/ — managed Kubernetes
- docs.hyperstack.cloud/docs/getting-started — billing starts when VM is ACTIVE, SSH, images
- docs.hyperstack.cloud/docs/storage/local-nvme — local NVMe storage
- docs.hyperstack.cloud/docs/support/ — support channels
- hyperstack.cloud/slas — SLA 99.5%
- status.hyperstack.cloud/ — status page
- hyperstack.cloud/secure-by-design — SOC 2 Type II, ISO 27001/HIPAA plans
- hyperstack.cloud/terms-and-conditions — KYC clause, card prepayment in USD, 18+, sanctions, negative balance, spot exclusions, governing law
- nexgencloud.com/about — company founding year and ownership
- hyperstack.cloud/nvidia-hgx-b300 — B300 on-demand 7.40 USD/h confirmation
- hyperstack.cloud/nvidia-blackwell-gb200 — GB200 availability (reserve only
- hyperstack.cloud/nvidia-gb300-nvl72 — GB300 availability (Secure Private Cloud
AlternativesProviders people compare with it
Impossible CloudOperator · 7 GPU typesnot scored
RuncrateReseller · 20 GPU types · H100 $2.83not scored
Thunder ComputeOperator · 4 GPU typesTrust 75
DaytonaOperator · 8 GPU types · H100 $3.95not scored
VerdaOperator · 15 GPU types · H100 $4.79Trust 81
EmpirioLabs AIOperator · 26 GPU types · H100 $3.80not scoredQuestions about Hyperstack
What GPUs does Hyperstack offer?
Hyperstack has 10 GPU types with a fresh offer today, for example H100 from $2.50, A100 80GB from $1.35, H200 from $3.99, RTX PRO 6000 from $1.85, B200 from $6.00, B300 from $7.40 per GPU-hour on-demand. The full list with prices and market medians is on this page.
How does Hyperstack bill?
Billing granularity: per minute. Minimum charge: No per-launch minimum published. To deploy a VM, the balance must cover all running resources plus the new VM for at least one hour. Stopped instance: SHUTOFF (stopped) VMs are billed in full, hardware stays reserved. HIBERNATED VMs: GPU/CPU/RAM not billed; root disk copied to Cloud-SSD and billed at 0.000096774 USD/GB-hour (about 0.07 USD/GB-month); public IP billed only if retained; attached volumes keep billing. Ephemeral disk is wiped on hibernation.
How trustworthy is Hyperstack?
Its Trust Score is 81/100 (v0.1), read from its own documents. Strong on infrastructure control, pricing & billing transparency. Weaker on security & compliance. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.
Where are Hyperstack's data centres?
3 named regions in 3 countries: Canada, Norway, United States.
How do you pay Hyperstack, and are there credits?
Payment by card, invoice. Free credits: Not published as a standing offer; vouchers can be redeemed and the T&Cs say test credits are offered occasionally. Identity check: card verification.