CoreWeave pricing: GPU rental rates
Updated October 2026 · updated 2026-10-06
The lowest published CoreWeave rate is $1.25/GPU-hour for the L40 (on-demand; No reservation; Not reported, updated 2026-10-06). CoreWeave is an AI cloud built around large NVIDIA GPU fleets, Kubernetes-native infrastructure, managed AI services and enterprise capacity planning. It fits teams that need serious training or inference infrastructure, especially when predictable GPU availability matters more than a tiny self-serve setup. Public pricing covers on-demand and spot GPU nodes, storage and managed Kubernetes, while committed capacity requires a sales motion.
Compare the full daily GPU market
GPU prices compared with the market
| GPU | Provider lowest firm rate | Market median | Difference |
|---|---|---|---|
| H100 | $6.16 | $3.39 | 81.3% |
| H200 | $6.30 | $4.76 | 32.5% |
| A100 | $2.70 | $1.65 | 63.6% |
| B200 | $8.60 | $6.84 | 25.7% |
| L40S | $2.25 | $1.57 | 43.1% |
| L40 | $1.25 | $0.905 | 38.1% |
| GH200 | $6.50 | $3.02 | 115.2% |
| RTX PRO 6000 | $2.50 | $2.17 | 15.5% |
The market median gives each provider equal weight. GPU variants, node size, included CPU and memory can differ. A lower rate does not establish equivalent performance.
Live CoreWeave prices (2026-10-06)
Provider, memory & availability
| GPU / provider | Billing | $/GPU-hour | GPUs / total per hour | Availability / details |
|---|---|---|---|---|
| L40CoreWeave48 GB VRAM | on-demand | $1.25per GPU-hour+0.0% over 30 days | 8× $10.00per node-hour | Not reportedView provider2026-10-06Details
|
| L40SCoreWeave48 GB VRAM | on-demand | $2.25per GPU-hour+0.0% over 30 days | 8× $18.00per node-hour | Not reportedView provider2026-10-06Details
|
| RTX PRO 6000 ServerCoreWeave96 GB VRAM | on-demand | $2.50per GPU-hour+0.0% over 30 days | 8× $20.00per node-hour | Not reportedView provider2026-10-06Details
|
| A100 (unspecified)CoreWeave80 GB VRAM | on-demand | $2.70per GPU-hour+0.0% over 30 days | 8× $21.60per node-hour | Not reportedView provider2026-10-06Details
|
| H100 (unspecified)CoreWeave80 GB VRAM | on-demand | $6.16per GPU-hour+0.0% over 30 days | 8× $49.24per node-hour | Not reportedView provider2026-10-06Details
|
| H200 (SXM)CoreWeave141 GB VRAM | on-demand | $6.30per GPU-hour+0.0% over 30 days | 8× $50.44per node-hour | Not reportedView provider2026-10-06Details
|
| GH200CoreWeave96 GB VRAM | on-demand | $6.50per GPU-hour+0.0% over 30 days | 1× $6.50per node-hour | Not reportedView provider2026-10-06Details
|
| B200CoreWeave180 GB VRAM | on-demand | $8.60per GPU-hour+0.0% over 30 days | 8× $68.80per node-hour | Not reportedView provider2026-10-06Details
|
Try another provider or reduce the minimum memory.
Tables rank by price, lowest first. Commercial relationships never affect ordering. How we collect and rank prices.
Monthly view assumes 730 hours of continuous use. Daily view assumes 24 hours.
Cheapest listed per-GPU-hour rates verified at the source. Before storage, egress and networking.
CoreWeave access and company facts
| Check | Published information | Updated |
|---|---|---|
| Smallest recorded configuration | 1 GPU | 2026-10-06 |
| Published regions in this comparison | No per-configuration region recorded | 2026-10-06 |
| Minimum account funding | not published | 2026-07-07 |
| SOC 2 | claimed | 2026-07-07 |
| ISO 27001 | claimed | 2026-07-07 |
| Data location terms | US+EU | 2026-07-07 |
A company fact or certification is not a guarantee for every product or region. Unrecorded fields are left unknown. The smallest recorded node does not establish immediate stock or account approval.
Strengths and trade-offs
Strengths
Trade-offs
- No published SLA
Computed from the verified fees and platform matrix (cells updated 2026-07-07 at CoreWeave's own documentation; verified items link to their source). Full matrices: fees · enterprise.
Pricing model
CoreWeave publishes portfolio-wide pricing for AI compute, storage, networking and managed services. GPU nodes are listed with on-demand and spot pricing where available, and some newest systems require contacting sales. CoreWeave also documents Reserved Compute Capacity discounts for committed usage, plus capacity plans that govern commitment level, billing structure and capacity guarantees. Use the live table for exact prices.
Capacity tiers
CoreWeave's public materials separate self-serve style usage from enterprise capacity planning. The pricing page lists on-demand and spot GPU compute, while the docs describe capacity plans for securing and paying for committed compute capacity. Managed Kubernetes is a first-class path through CoreWeave Kubernetes Service, and the pricing page also lists managed Kubernetes control-plane items. This is more enterprise and platform oriented than a simple notebook GPU rental service.
GPU lineup
CoreWeave publishes a large NVIDIA GPU catalog. Current public pricing and product pages include Blackwell and Blackwell Ultra systems such as GB300, GB200, HGX B300, HGX B200 and RTX PRO 6000 Blackwell Server Edition, plus HGX H200, HGX H100, GH200, L40, L40S and A100 options. Availability, region and spot support vary by SKU, so buyers should check the live pricing page before planning around a specific GPU.
Storage, egress and hidden costs
CoreWeave lists AI Object Storage tiers, Distributed File Storage and an egress and IOPS line item on its pricing page. The public pricing page marks egress and input output operations per second as free, while storage itself is billed separately and reserved storage discounts may be available. Hidden-cost checks should focus on reserved capacity terms, storage volume growth, managed service usage, and whether the chosen GPU SKU is on-demand, spot or sales-led.
Reliability and support posture
CoreWeave positions itself for large AI labs and enterprise production workloads. Its Trust Center is the source for current security, privacy, resiliency and compliance information, and its docs include security material for workload isolation and identity controls. Terms of Service include uptime language, but buyers should verify the current SLA, service-credit mechanism and support plan for their exact agreement before treating marketing reliability claims as contractual.
Free credits and referral programs
CoreWeave has publicly announced a Startup Accelerator Program for AI and machine-learning-focused companies, with GPU cloud credits, compute discounts and support. The exact credit amount and current eligibility are not publicly documented. A public self-serve referral program was not found in CoreWeave's public docs.
Who should pick CoreWeave
Pick CoreWeave if you need an enterprise-grade AI cloud with large NVIDIA GPU systems, Kubernetes-native operations and a path to reserved capacity. It is best for production inference, distributed training and teams that can work through sales or capacity planning for larger commitments. Smaller builders who only need a quick single GPU may prefer a simpler marketplace or developer-first GPU rental platform.
Similar providers
RunPod
Secure Cloud (datacenter, reliable) and Community Cloud (peer hosts, cheaper). Per-second billing, serverless option.
from $0.25/hr secure cloud
7 GPU families in common
Hyperstack
European GPU cloud with on-demand and reserved pricing.
from $0.15/hr on-demand
6 GPU families in common
Ranked by overlap in the GPU models both providers list today; not an endorsement and never affected by commercial relationships.
FAQ
Yes. CoreWeave publishes live pricing for many GPU node types, with some newest or reserved configurations handled through sales.
Yes. The public pricing table includes spot pricing for supported GPU SKUs. Spot availability varies by SKU and region.
CoreWeave's pricing page marks egress and IOPS as free, while storage capacity is billed separately.
CoreWeave has announced a Startup Accelerator Program with credits and discounts, but exact current credit amounts are not publicly documented.
Profile updated 2026-07-06. Sources: www.coreweave.com · www.coreweave.com · www.coreweave.com · www.coreweave.com · docs.coreweave.com · www.coreweave.com · docs.coreweave.com · docs.coreweave.com · www.coreweave.com