CoreWeave vs Hyperstack
OCT 2026 auditA comparison of CoreWeave and Hyperstack built from values read directly from each provider's own documentation, with the source recorded against every figure.
The short answer
Choose CoreWeave if…
Training runs that want a whole HGX node and move large datasets — free egress and free I/O operations remove the charge that usually dominates a data-heavy job.
Editorial · Palash Bagchi · approved
Choose Hyperstack if…
Steady European workloads that can commit — reserved H200 at $2.79 against $3.99 on demand is the clearest published discount in this set, and egress is free.
Editorial · Palash Bagchi · approved
Consider something else if…
- CoreWeave: You want one H100 on demand: it is sold as an eight-GPU instance at $49.24 an hour, and the $6.16 per-GPU figure CoreWeave prints applies only to its inference-platform customers.
- Hyperstack: Billing is per minute rather than per second, so very short jobs round up in a way they do not at RunPod.
Neither quite fits? Every other option, ranked by iScore: CoreWeave alternatives · Hyperstack alternatives
16 sourced criteria separate them — see where, with sources, below.
Pick the criteria you care about. The chart counts how many of them lean toward each provider — the same read as scanning the bars below, just totalled for the ones you chose.
CoreWeave 0
Hyperstack 0
At a glance.
| Criterion | CoreWeave | Hyperstack |
|---|---|---|
| Pricing | ||
| Pricing model | Four CKS capacity plans - Flex Reservations, Reserved Instances, Spot Instances and On-Demand; On-Demand instances are billed on actual use with no long-term commitment | Pay-as-you-go at hourly rates billed per minute; prepaid credit by default, with postpaid accounts billed at a fixed interval, typically monthly. |
| Egress over allowance | Free. CoreWeave charges nothing for egress or for I/O operations, and runs a zero-egress migration programme for moving data in. | Free. Egress and ingress traffic are not charged; a public IP address costs $0.00672043 per hour. |
| Infrastructure | ||
| Regions | Availability Zones listed across seven Super Regions (CA-EAST, EU-NORTH, EU-SOUTH, EU-WEST, US-CENTRAL, US-EAST, US-WEST), sited in the US, Canada, UK, Spain, Sweden, Norway and Denmark | Three documented regions: NORWAY-1 (Vestland, Norway), CANADA-1 (Quebec, Canada) and US-1 (Texas, United States); more may appear per account. |
| GPU | ||
| H100 SXM, per GPU-hour | $6.16 per GPU-hour for NVIDIA HGX H100 (80GB) is CoreWeave's inference-platform price, printed in its "Inference Single GPU Price" column, which it says applies only to inference platform customers. The on-demand instance price is $49.24/hour for the 8-GPU HGX H100 instance, spot $19.71/hour (North America). | $3.20 per hour for H100 SXM 80GB on demand. The H100 NVLink is $2.60, and the row the page lists as plain "NVIDIA H100" is $2.50. |
| Cheapest GPU, per hour | $1.25 per GPU-hour for NVIDIA L40 (48GB) is CoreWeave's inference-platform price, from its "Inference Single GPU Price" column, which applies only to inference platform customers. The on-demand instance price is $10.00/hour for the 8-GPU L40 instance, spot $6.27/hour (North America). | $0.15 per hour for an A4000 with 16GB; the A100 SXM 80GB is $1.60. |
| Largest GPU offered | 279GB per GPU on GB300 NVL72, which is Contact sales only. HGX B300 is next at 270GB, with no on-demand price and a $35.84/hour spot price for the 8-GPU instance. The most VRAM with a published on-demand price is GB200 NVL72 at 186GB per GPU: $42.00/hour instance price for 4 GPUs, or $10.50 per GPU-hour on the inference platform. | 288GB per GPU on B300 at $7.40 an hour; B200 at 192GB is $6.00. |
| Billing granularity | Invoice line items are per SKU with quantity in hours; usage attribution across Reserved, Flex and On-Demand capacity runs every 30 seconds | Per minute, with enterprise-scale contracts negotiated separately. |
| Reserved pricing | Up to 60% below on-demand for reserved compute capacity. | Published reserved rates rather than a percentage: H200 SXM falls from $3.99 on demand to $2.79 reserved, about 30% below. |
| Compliance | ||
| SOC 2 | SOC 2 Type II is listed as a badge in the CoreWeave Trust Center; detailed documentation is available by requesting access. | SOC 2 Type 2 attested, achieved via parent company NexGen Cloud; the same page shows SOC 2 Type 1 as certified. |
| ISO 27001 | ISO 27001 (with ISO 27017 and ISO 27018) is among the frameworks CoreWeave's compliance evidence maps to, available via its Compliance Documentation Portal. | Not certified: ISO 27001 sits on the upcoming-certifications list, shown as "ISO 20071" (sic) in progress. |
| HIPAA | Security docs say multi-tenant isolation and encryption controls help customers address regulatory requirements such as GDPR, HIPAA and CCPA. Stated as helping customers meet requirements, not as a HIPAA attestation or an offer of a BAA. | Not yet: HIPAA is listed under "Coming Next", the security page's roster of upcoming certifications. |
| GDPR / data residency | Regions are placed to meet data residency requirements, and Geos are used to ensure compliance with regulatory requirements. | Regional data residency options across the UK, EU and other supported regions, with data centre regions in the EU, Canada and the US. |
| Encryption at rest | AI Object Storage encrypts object data at rest by default with AES-256 at no additional cost; object metadata is not encrypted by default. Read for CoreWeave AI Object Storage; other storage products document encryption separately. | Not server-side for Object Storage: docs tell users who need encryption at rest to encrypt client-side. The Secure By Design page separately claims data is encrypted at rest and in transit. |
| Role-based access control | IAM Access Policies assign roles to users or groups across CoreWeave services; CKS clusters use Kubernetes RBAC and Object Storage its own bucket policies. | Role-based access control: custom user roles built from policies and individual permissions; invited members start with none until a role is assigned. |
| Private networking / VPC | VPCs define network isolation for attached products such as CKS clusters; VPC prefixes reach other VPCs, remote clouds or on-premises networks over VPN or Direct Connect. | Each environment is allocated its own private /24 subnet; every VM in it gets a private IP from that subnet and they share one network space. |
| Dedicated infrastructure | Dedicated Access AZs are single-tenant: no other customer's workloads run in the AZ and no other tenant shares the underlying network fabric; reserved for select customers. | Secure Private Cloud: dedicated, single-tenant infrastructure with zero shared resources, running on dedicated hardware with no shared GPUs. |
Where they differ.
Pricing model
- CoreWeave
- Four CKS capacity plans - Flex Reservations, Reserved Instances, Spot Instances and On-Demand; On-Demand instances are billed on actual use with no long-term commitment
- Hyperstack
- Pay-as-you-go at hourly rates billed per minute; prepaid credit by default, with postpaid accounts billed at a fixed interval, typically monthly.
Sources (2) →Sources ↓
- About capacity plans - CoreWeave Docs ↗
“CoreWeave Kubernetes Service (CKS) offers four capacity plans: Flex Reservations Reserved Instances Spot Instances On-Demand [...] These instances are billed based on actual use without a long-term commitment.”
Read 2026-10-17 · official docs · Report a correction
- Accounts, Billing Policies, and Usage Monitoring - Hyperstack Docs ↗
“Hyperstack supports three account types, each with distinct billing behavior [...] Postpaid Account : Instead of pre-funding, postpaid accounts are billed periodically based on accumulated usage. Billing occurs at the end of a fixed interval, typically monthly. [...] Our billing system operates on a per-minute structure”
Read 2026-10-17 · official docs · Report a correction
Egress over allowance
- CoreWeave
- Free. CoreWeave charges nothing for egress or for I/O operations, and runs a zero-egress migration programme for moving data in.
- Hyperstack
- Free. Egress and ingress traffic are not charged; a public IP address costs $0.00672043 per hour.
Sources (2) →Sources ↓
- Pricing — CoreWeave ↗
“Egress and Input Output Operations per Second - Free [...] Data transfer between CoreWeave and internet Free [...] Zero Egress Migration (0EM) Move your data to CoreWeave with zero egress fees”
Read 2026-10-17 · official pricing · Report a correction
- GPU Pricing — Hyperstack ↗
“Public IP address 0.00672043 per hour Egress/Ingress traffic Free On-Demand Kubernetes Free master node”
Read 2026-10-17 · official pricing · Report a correction
Regions
- CoreWeave
- Availability Zones listed across seven Super Regions (CA-EAST, EU-NORTH, EU-SOUTH, EU-WEST, US-CENTRAL, US-EAST, US-WEST), sited in the US, Canada, UK, Spain, Sweden, Norway and Denmark
- Hyperstack
- Three documented regions: NORWAY-1 (Vestland, Norway), CANADA-1 (Quebec, Canada) and US-1 (Texas, United States); more may appear per account.
Sources (2) →Sources ↓
- All Availability Zones - CoreWeave Docs ↗
“View all CoreWeave Availability Zones, grouped by Super Region [...] CA-EAST EU-NORTH EU-SOUTH EU-WEST US-CENTRAL US-EAST US-WEST [...] CA-EAST-01A Ontario, Canada General Access”
Read 2026-10-17 · official docs · Report a correction
- Regions & Data Centers - Hyperstack Docs ↗
“Hyperstack provides multiple regions (e.g., CANADA-1 , NORWAY-1 , US-1 ) [...] NORWAY-1 Vestland, Norway Sustainably Powered Region Not Available CANADA-1 Quebec, Canada Sustainably Powered Region Available* US-1 Texas, United States Standard Energy Region Available*”
Read 2026-10-17 · official docs · Report a correction
H100 SXM, per GPU-hour
- CoreWeave
- $6.16 per GPU-hour for NVIDIA HGX H100 (80GB) is CoreWeave's inference-platform price, printed in its "Inference Single GPU Price" column, which it says applies only to inference platform customers. The on-demand instance price is $49.24/hour for the 8-GPU HGX H100 instance, spot $19.71/hour (North America).
- Hyperstack
- $3.20 per hour for H100 SXM 80GB on demand. The H100 NVLink is $2.60, and the row the page lists as plain "NVIDIA H100" is $2.50.
Sources (2) →Sources ↓
- Pricing — CoreWeave ↗
“On-Demand Price (Per Hour) Spot Price (Per Hour) Inference Single GPU Price(Per Hour)² [...] NVIDIA HGX H100 8 80 128 2,048 61.44 $49.24 NVIDIA HGX H100 On-Demand Price: $49.24 / Hour Spot Price: $19.71 / Hour Inference Single CPU Price: $6.16 / Hour [...] Pricing based on GPUs is exclusively applicable to the CoreWeave inference platform customers.”
Read 2026-10-17 · official pricing · Report a correction
- GPU Pricing — Hyperstack ↗
“NVIDIA H100 SXM 80 24 240 $3.20 NVIDIA H100 NVLink 80 31 180 $2.60 NVIDIA H100 80 28 180 $2.50”
Read 2026-10-17 · official pricing · Report a correction
Cheapest GPU, per hour
- CoreWeave
- $1.25 per GPU-hour for NVIDIA L40 (48GB) is CoreWeave's inference-platform price, from its "Inference Single GPU Price" column, which applies only to inference platform customers. The on-demand instance price is $10.00/hour for the 8-GPU L40 instance, spot $6.27/hour (North America).
- Hyperstack
- $0.15 per hour for an A4000 with 16GB; the A100 SXM 80GB is $1.60.
Sources (2) →Sources ↓
- Pricing — CoreWeave ↗
“On-Demand Price (Per Hour) Spot Price (Per Hour) Inference Single GPU Price(Per Hour)² [...] NVIDIA L40 8 48 128 1,024 7.68 $10.00 NVIDIA L40 On-Demand Price: $10.00 / Hour Spot Price: $6.27 / Hour Inference Single CPU Price: $1.25 / Hour [...] Pricing based on GPUs is exclusively applicable to the CoreWeave inference platform customers.”
Read 2026-10-17 · official pricing · Report a correction
- GPU Pricing — Hyperstack ↗
“NVIDIA A4000 16 6 24 $0.15 [...] NVIDIA A100 SXM 80 24 120 $1.60”
Read 2026-10-17 · official pricing · Report a correction
Largest GPU offered
279GB per GPU on GB300 NVL72, which is Contact sales only. HGX B300 is next at 270GB, with no on-demand price and a $35.84/hour spot price for the 8-GPU instance. The most VRAM with a published on-demand price is GB200 NVL72 at 186GB per GPU: $42.00/hour instance price for 4 GPUs, or $10.50 per GPU-hour on the inference platform.
288GB per GPU on B300 at $7.40 an hour; B200 at 192GB is $6.00.
Sources (2) →Sources ↓
- Pricing — CoreWeave ↗
“GPU Model GPU Count VRAM (GB) vCPUs System RAM (GB) Local Storage (TB) On-Demand Price (Per Hour) Spot Price (Per Hour) Inference Single GPU Price(Per Hour)² NVIDIA GB300 NVL72 4^1 279 144 960 61.44 Contact sales [...] NVIDIA GB200 NVL72 4^1 186 144 960 30.72 $42.00 NVIDIA GB200 NVL72 On-Demand Price: $42.00 / Hour Spot Price: N/A / Hour Inference Single CPU Price: $10.50 / Hour [...] NVIDIA HGX B300 8 270 192 4,096 61.44 Contact sales NVIDIA HGX B300 On-Demand Price: / Hour Spot Price: $35.84 / Hour”
Read 2026-10-17 · official pricing · Report a correction
- GPU Pricing — Hyperstack ↗
“NVIDIA B200 192 31 256 $6.00 NVIDIA B300 288 28 240 $7.40”
Read 2026-10-17 · official pricing · Report a correction
Billing granularity
- CoreWeave
- Invoice line items are per SKU with quantity in hours; usage attribution across Reserved, Flex and On-Demand capacity runs every 30 seconds
- Hyperstack
- Per minute, with enterprise-scale contracts negotiated separately.
Sources (2) →Sources ↓
- About capacity plans - CoreWeave Docs ↗
“Billing reconciles all consumption into one invoice. Each line item shows the SKU, quantity (hours), and the applied rate type. [...] Attribution runs every 30 seconds. For each SKU, usage is applied in this order:”
Read 2026-10-17 · official docs · Report a correction
- GPU Pricing — Hyperstack ↗
“competitive pricing structure, billed per minute. For larger enterprise-scale contracts, please contact”
Read 2026-10-17 · official pricing · Report a correction
Reserved pricing
- CoreWeave
- Up to 60% below on-demand for reserved compute capacity.
- Hyperstack
- Published reserved rates rather than a percentage: H200 SXM falls from $3.99 on demand to $2.79 reserved, about 30% below.
Sources (2) →Sources ↓
- Pricing — CoreWeave ↗
“Reserved Compute Capacity CoreWeave offers up to 60% discounts over our On-Demand prices for committed usage.”
Read 2026-10-17 · official pricing · Report a correction
- GPU Pricing — Hyperstack ↗
“NVIDIA H200 SXM 141 22 225 $3.99 [...] Reservation Pricing GPU Model Starting from Reserve NVIDIA H200 SXM $2.79 Reserve here”
Read 2026-10-17 · official pricing · Report a correction
SOC 2
- CoreWeave
- SOC 2 Type II is listed as a badge in the CoreWeave Trust Center; detailed documentation is available by requesting access.
- Hyperstack
- SOC 2 Type 2 attested, achieved via parent company NexGen Cloud; the same page shows SOC 2 Type 1 as certified.
Sources (2) →Sources ↓
- CoreWeave Trust Center ↗
“Browse documents, certifications, and compliance details with confidence. [...] use the "Get Access" button on the top-right corner of the page to request access to the detailed security documentation. [...] Badges ISO 27001 SOC 2 Type II ISO 27017 ISO 27018”
Read 2026-10-17 · official site · Report a correction
- Secure By Design — Hyperstack ↗
“We have achieved SOC 2 Type 2, an independent attestation of our security controls. [...] SOC 2 Type 2 achieved via NexGen Cloud, our parent company [...] SOC 2 Type I Certified SOC 2 Type II Attested”
Read 2026-10-17 · official site · Report a correction
ISO 27001
- CoreWeave
- ISO 27001 (with ISO 27017 and ISO 27018) is among the frameworks CoreWeave's compliance evidence maps to, available via its Compliance Documentation Portal.
- Hyperstack
- Not certified: ISO 27001 sits on the upcoming-certifications list, shown as "ISO 20071" (sic) in progress.
Sources (2) →Sources ↓
- Compliance programs - CoreWeave Docs ↗
“CoreWeave aligns its security and compliance programs with recognized regulatory standards so you can run sensitive and regulated workloads on the CoreWeave Cloud. [...] third-party assessments that map to major frameworks and regulations such as: SOC 2 ISO 27001 ISO 27017 ISO 27018”
Read 2026-10-17 · official docs · Report a correction
- Secure By Design — Hyperstack ↗
“SOC 2 Type II Attested ISO 20071 In progress [...] ISO 27001 A globally recognised standard for information security management”
Read 2026-10-17 · official site · Report a correction
HIPAA
- CoreWeave
- Security docs say multi-tenant isolation and encryption controls help customers address regulatory requirements such as GDPR, HIPAA and CCPA. Stated as helping customers meet requirements, not as a HIPAA attestation or an offer of a BAA.
- Hyperstack
- Not yet: HIPAA is listed under "Coming Next", the security page's roster of upcoming certifications.
Sources (2) →Sources ↓
- Security architecture - CoreWeave Docs ↗
“CoreWeave’s multi-tenant isolation and encryption features are backed by enforceable technical controls that support trustworthy, secure AI adoption at scale. They help customers address regulatory requirements such as GDPR, HIPAA, and CCPA.”
Read 2026-10-17 · official docs · Report a correction
- Secure By Design — Hyperstack ↗
“Coming Next Our upcoming certifications include: HIPAA A US healthcare standard for protecting sensitive patient data”
Read 2026-10-17 · official site · Report a correction
GDPR / data residency
- CoreWeave
- Regions are placed to meet data residency requirements, and Geos are used to ensure compliance with regulatory requirements.
- Hyperstack
- Regional data residency options across the UK, EU and other supported regions, with data centre regions in the EU, Canada and the US.
Sources (2) →Sources ↓
- About Regions and Availability Zones (AZs) - CoreWeave Docs ↗
“CoreWeave uses Geos to manage global infrastructure, ensure compliance with regulatory requirements, and optimize for large-scale redundancy and disaster recovery. [...] CoreWeave places Regions to offer low latency, high-performance connectivity, and to meet data residency requirements.”
Read 2026-10-17 · official docs · Report a correction
- Secure By Design — Hyperstack ↗
“Availability of data centre regions in the EU, Canada, and US [...] Regional Data Residency Options Choose where your data resides across the UK, EU and other supported regions.”
Read 2026-10-17 · official site · Report a correction
Encryption at rest
- CoreWeave
- AI Object Storage encrypts object data at rest by default with AES-256 at no additional cost; object metadata is not encrypted by default. Read for CoreWeave AI Object Storage; other storage products document encryption separately.
- Hyperstack
- Not server-side for Object Storage: docs tell users who need encryption at rest to encrypt client-side. The Secure By Design page separately claims data is encrypted at rest and in transit.
Sources (2) →Sources ↓
- About CoreWeave AI Object Storage - CoreWeave Docs ↗
“AI Object Storage provides encryption at rest by default and supports encryption in transit through TLS. For additional control, you can supply your own encryption keys using SSE-C. Encryption has no additional cost. [...] Only the object data itself is encrypted at rest. Object metadata is not encrypted by default.”
Read 2026-10-17 · official docs · Report a correction
- Object Storage - Hyperstack Docs ↗
“Server-side encryption is not currently supported. Users who require encryption at rest should implement client-side encryption.”
Read 2026-10-17 · official docs · Report a correction
Role-based access control
- CoreWeave
- IAM Access Policies assign roles to users or groups across CoreWeave services; CKS clusters use Kubernetes RBAC and Object Storage its own bucket policies.
- Hyperstack
- Role-based access control: custom user roles built from policies and individual permissions; invited members start with none until a role is assigned.
Sources (2) →Sources ↓
- IAM Access Policies - CoreWeave Docs ↗
“CoreWeave IAM Access Policies let administrators and privileged users define which principals (CoreWeave IAM users or groups) can perform specific actions across CoreWeave services. [...] The CoreWeave AI Object Storage S3-compatible API, which is governed by organization and bucket access policies. [...] Kubernetes authorization within CoreWeave Kubernetes Service (CKS) clusters, which uses Kubernetes RBAC after a user authenticates to the Kubernetes API.”
Read 2026-10-17 · official docs · Report a correction
- Organization - Hyperstack Docs ↗
“Role-Based Access Control (RBAC) system enables you to create user roles that specify the level of access each member has. [...] new members will not have any permissions assigned by default”
Read 2026-10-17 · official docs · Report a correction
Private networking / VPC
- CoreWeave
- VPCs define network isolation for attached products such as CKS clusters; VPC prefixes reach other VPCs, remote clouds or on-premises networks over VPN or Direct Connect.
- Hyperstack
- Each environment is allocated its own private /24 subnet; every VM in it gets a private IP from that subnet and they share one network space.
Sources (2) →Sources ↓
- Virtual Private Clouds (VPCs) - CoreWeave Docs ↗
“define the network isolation configuration for attached CoreWeave products, like CKS clusters. [...] Routing: VPC Prefixes let the cluster reach other VPCs in the same Region, remote cloud providers, or on-premises networks through a VPN or Direct Connect.”
Read 2026-10-17 · official docs · Report a correction
- Environments - Hyperstack Docs ↗
“Each environment is allocated its own private subnet, expressed as a /24 CIDR block such as 10.0.5.0/24 [...] Every virtual machine you deploy into the environment receives a private IP address from that subnet. All virtual machines in an environment therefore share the same network space”
Read 2026-10-17 · official docs · Report a correction
Dedicated infrastructure
- CoreWeave
- Dedicated Access AZs are single-tenant: no other customer's workloads run in the AZ and no other tenant shares the underlying network fabric; reserved for select customers.
- Hyperstack
- Secure Private Cloud: dedicated, single-tenant infrastructure with zero shared resources, running on dedicated hardware with no shared GPUs.
Sources (2) →Sources ↓
- About Regions and Availability Zones (AZs) - CoreWeave Docs ↗
“CoreWeave manages Dedicated Access AZs and reserves them for select customers. Infrastructure in a Dedicated Access AZ is single-tenant: no other customer’s workloads run in the AZ, and no other tenant shares the underlying network fabric.”
Read 2026-10-17 · official docs · Report a correction
- Secure Private Cloud for AI Workloads: Single-Tenant GPU Infrastructure | Hyperstack ↗
“Hyperstack Secure Private Cloud provide dedicated, single-tenant infrastructure with zero shared resources and full control. [...] Your workloads run on dedicated hardware with no shared GPUs, no noisy neighbours, no cross-tenant risk.”
Read 2026-10-17 · official site · Report a correction
When these numbers change, hear about it. Sources are re-checked monthly; a repricing goes out as a short note.
Documented by only one.
These criteria are published by one provider and not the other. An absence here means we have not found a source, not that the feature is missing — except where we say we read the other provider's own page on a given date and it does not state it. Those rows lean, and the provider can correct them with a link.
Questions this comparison answers.
Egress over allowance: CoreWeave or Hyperstack?
CoreWeave: Free. CoreWeave charges nothing for egress or for I/O operations, and runs a zero-egress migration programme for moving data in.
Hyperstack: Free. Egress and ingress traffic are not charged; a public IP address costs $0.00672043 per hour.
Should I choose CoreWeave or Hyperstack?
Pick CoreWeave if Training runs that want a whole HGX node and move large datasets — free egress and free I/O operations remove the charge that usually dominates a data-heavy job.
Pick Hyperstack if Steady European workloads that can commit — reserved H200 at $2.79 against $3.99 on demand is the clearest published discount in this set, and egress is free.
Comments are hosted by EchoThread, which sets cookies and offers social sign-in. Like the analytics on this site, it does not load until you accept — see the privacy page.