Gcore
Luxembourg-headquartered cloud, CDN and security provider selling 8-GPU bare-metal nodes (H100, H200, A100, L40S, B300; GB300 NVL72 by request), H100 Virtual GPU VMs and per-minute serverless inference.
Trust Score v0.1- Company & track record73
- Infrastructure control100
- Pricing & billing transparency100
- Reliability signals75
- Security & compliance47
- 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 | 2014gcore.com · Sep 29, 2026 | 4 / 4 |
| Headquarters published | G-Core Labs S.A.gcore.com · Sep 29, 2026 | 3 / 3 |
| Legal entity named | G-Core Labs S.A.gcore.com · Sep 29, 2026 | 4 / 4 |
| Funding or ownership disclosed | not disclosedgcore.com · Sep 29, 2026 | 0 / 4 |
| Capacity model published | operates or rents dedicated capacitygcore.com · Sep 29, 2026 | 10 / 10 |
| Regions published | 10 regions or countriesgcore.com · Sep 29, 2026 | 5 / 5 |
| Public prices without an account | yes (read on the pricing page)gcore.com · Sep 29, 2026 | 5 / 5 |
| Billing granularity published | per minutegcore.com · Sep 29, 2026 | 4 / 4 |
| Egress policy published | metered for Bare Metal: 0.0013 EUR (0.00143 USD)/GB per the pricing calculator and Edge Cloud billing doc, while the GPU Cloud doc states egress from GPU clusters is free (contradiction); free for Virtual Machinesgcore.com · Sep 29, 2026 | 4 / 4 |
| Stopped-instance billing published | Bare Metal GPU servers are charged from creation until deletion regardless of power state. Virtual GPU: shelving (power off) stops compute billing; volumes and IPs remain billed; the local data disk is wiped on shelving.gcore.com · Sep 29, 2026 | 4 / 4 |
| Public IPv4 policy published | extra: 2.5331 EUR (2.7886 USD)/month per external IP (calculator, GPU regions)gcore.com · Sep 29, 2026 | 3 / 3 |
| Status page | https://status.gcore.com/gcore.com · Sep 29, 2026 | 6 / 6 |
| SLA published | 99.9% monthly availability for GPU Cloud, Bare Metal and Spot Instances, Inference at the Edge (Cloud SLA 3.3); service credits 5% (99.89-99.0%), 10% (98.99-90.0%), 20% (below 90%)gcore.com · Sep 29, 2026 | 6 / 6 |
| SLA with a numeric uptime commitment | 99.9% monthly availability for GPU Cloud, Bare Metal and Spot Instances, Inference at the Edge (Cloud SLA 3.3); service credits 5% (99.89-99.0%), 10% (98.99-90.0%), 20% (below 90%)gcore.com · Sep 29, 2026 | 3 / 3 |
| Uptime history published | not disclosedgcore.com · Sep 29, 2026 | 0 / 5 |
| Observed availability (30 days) | not measured yet | — |
| SOC 2 | not disclosedgcore.com · Sep 29, 2026 | 0 / 8 |
| ISO 27001 | yes (ISO/IEC 27001 stated on infrastructure and security pages)gcore.com · Sep 29, 2026 | 4 / 4 |
| HIPAA or GDPR DPA | yes (Data Processing Agreement on legal page; GDPR compliance stated)gcore.com · Sep 29, 2026 | 3 / 3 |
| Aggregated public reviews | not collected yet | — |
Market report · Oct 25, 2026 · Hyperscalers and general clouds · #23 of the 200 to audit · prices read on the provider's pageWhat the market report notes about Gcore
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
- Gcore
- Cross-checkfr
- H100 SXM 1,78 $ ; H200 (8×) 3,17 $ (07/09/2026) ; partenaire NVIDIA Dynamo.
Prices by GPU · 5 GPU types with a current offerWhat Gcore 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 | $1.71 | $3.17 | −46% | 9 | on-demand | H100 prices |
| A100 80GB | $1.41 | $1.99 | −29% | 5 | on-demand | A100 80GB prices |
| H200 | $3.55 | $4.76 | −25% | 1 | on-demand | H200 prices |
| L40S | $1.21 | $1.26 | −5% | 4 | on-demand | L40S prices |
| B300 | $6.70 | $8.96 | −25% | 1 | on-demand | B300 prices |
All offers · 71 readings on fileEvery Gcore 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 | |
|---|---|---|---|---|---|---|---|---|
| L40S | 1× L40S | On-demand | — | Stock unknown | $1.21 | $1.21 | 21:35 UTCfresh | Receipt |
| L40S | 2× L40S | On-demand | — | Stock unknown | $1.21 | $1.21 | 21:35 UTCfresh | Receipt |
| L40S | 4× L40S | On-demand | — | Stock unknown | $1.21 | $1.21 | 21:35 UTCfresh | Receipt |
| A100 80GB | 1× A100 80GB | On-demand | — | Stock unknown | $1.41 | $1.41 | 21:35 UTCfresh | Receipt |
| A100 80GB | 2× A100 80GB | On-demand | — | Stock unknown | $1.41 | $1.41 | 21:35 UTCfresh | Receipt |
| A100 80GB | 4× A100 80GB | On-demand | — | Stock unknown | $1.41 | $1.41 | 21:35 UTCfresh | Receipt |
| A100 80GB | 8× A100 80GB | On-demand | — | Stock unknown | $1.41 | $1.41 | 21:35 UTCfresh | Receipt |
| L40S | 8× L40S | On-demand | — | Stock unknown | $1.42 | $1.42 | 21:35 UTCfresh | Receipt |
| A100 80GB | 8× A100 80GB | On-demand | — | Stock unknown | $1.52 | $1.52 | 21:35 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | On-demand | — | Stock unknown | $1.71 | $1.71 | 21:35 UTCfresh | Receipt |
| H100SXM | 2× H100 SXM | On-demand | — | Stock unknown | $1.71 | $1.71 | 21:35 UTCfresh | Receipt |
| H100SXM | 4× H100 SXM | On-demand | — | Stock unknown | $1.71 | $1.71 | 21:35 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM | On-demand | — | Stock unknown | $1.71 | $1.71 | 21:35 UTCfresh | Receipt |
| H100SXM | 4× H100 SXM | On-demand | — | Stock unknown | $2.64 | $2.64 | 21:35 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM | On-demand | — | Stock unknown | $2.64 | $2.64 | 21:35 UTCfresh | Receipt |
| H100SXM | 2× H100 SXM | On-demand | — | Stock unknown | $2.64 | $2.64 | 21:35 UTCfresh | Receipt |
| H100SXM | 8× H100 SXM | On-demand | — | Stock unknown | $3.05 | $3.05 | 21:35 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM | On-demand | — | Stock unknown | $3.15 | $3.15 | 21:35 UTCfresh | Receipt |
| H200SXM | 8× H200 SXM | On-demand | — | Stock unknown | $3.55 | $3.55 | 21:35 UTCfresh | Receipt |
| B300 | 8× B300 | On-demand | — | Stock unknown | $6.70 | $6.70 | 21:35 UTCfresh | Receipt |
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
- Edge Cloud needs at least EUR 4 / USD 4 on the balance to start; PAYG usage is metered about every 5 minutes and charged each time accrued costs reach EUR/USD 4. Spot clusters need a minimum balance (amount not found on the pages read).
- Stopped instance
- Bare Metal GPU servers are charged from creation until deletion regardless of power state. Virtual GPU: shelving (power off) stops compute billing; volumes and IPs remain billed; the local data disk is wiped on shelving.
- Egress
- metered for Bare Metal: 0.0013 EUR (0.00143 USD)/GB per the pricing calculator and Edge Cloud billing doc, while the GPU Cloud doc states egress from GPU clusters is free (contradiction); free for Virtual Machines
- Public IPv4
- extra: 2.5331 EUR (2.7886 USD)/month per external IP (calculator, GPU regions)
- Billing model
- both
- Minimum deposit
- EUR 4 / USD 4 to start using Edge Cloud
- Free credits
- none: no trial for Cloud (AI, Basic VM, DDoS Protection) per trial conditions doc
Account
- Payment
- card · invoice · PayPal
- Sign-up
- email, password, card, KYC above threshold
- Identity check
- ID on request
- Support
- email, chat, phone
- Enterprise support
- not published
- SLA
- 99.9% monthly availability for GPU Cloud, Bare Metal and Spot Instances, Inference at the Edge (Cloud SLA 3.3); service credits 5% (99.89-99.0%), 10% (98.99-90.0%), 20% (below 90%)
- Status page
- status.gcore.com ↗
- Published uptime
- not published
Infrastructure
- Regions
10 named regions in 7 countries
Chester (NDCH-2) (United Kingdom) · Helsinki (TH-2) (Finland) · Incheon (NHN-2) (South Korea) · London-2 (NDLO-2) (United Kingdom) · Luxembourg-2 (ED-16) (Luxembourg) · Luxembourg-3 (LUX-2) (Luxembourg) · Manassas (ANX-2) (United States) · Sines-2 (NDP1-2) (Portugal) · Sines-3 (NDP1-3) (Portugal) · Singapore (SGC-2) (Singapore)- Countries
- Finland, Luxembourg, Portugal, Singapore, South Korea, United Kingdom, United States
- Platform
- api, serverless, ssh, persistent_volumes, snapshots, spot_or_interruptible, bare_metal · SDK : python, go · Terraform : official · Kubernetes : managed · interconnexion : InfiniBand 3.2 Tbit/s (H100, H200), 800 Gbit/s (A100), 3,600 Gbps (B300) · NVLink 130 TB/s within GB300 NVL72 rack · clusters : Multi-node Bare Metal and Virtual GPU clusters with automatic InfiniBand setup · 250+ nodes in one cluster supported, limited by regional stock
- Storage
- volume · Standard 0.10 EUR (0.11 USD)/GB-month in Luxembourg-2 and Manassas; 0.05 EUR (0.055 USD)/GB-month in Sines-2 and Chester · Calculator back end (bff.gcore.pro/cloud/volume-price) used by gcore.com/pricing/cloud, period 43,800 minvolume · SSD High IOPS 0.22 EUR (0.242 USD)/GB-month; SSD Low-latency 0.15 EUR (0.165 USD)/GB-month · Same in Luxembourg-2, Sines-2, Chester, Manassasobject · Standard PAYG 0.04 EUR/GiB-month; FAST (AI workloads) PAYG 0.064 EUR/GiB-month; 5 TiB plan 120 EUR/month; 20 TiB plan 400 EUR/month · Other traffic 0.02 EUR/GiB (FAST 0.0013 EUR/GiB); 10,000 write requests 0.03 EURnetwork · None · File shares (VAST-backed in GPU regions) exist; price not found in the pages readephemeral · included · Local NVMe on bare-metal nodes and a local data disk on Virtual GPU nodes, erased on deletion (and on shelving for Virtual GPU)
- Migration out
- File shares, network volumes and object storage persist independently of clusters; local NVMe is erased on deletion. Custom images can be uploaded and VM snapshots exist; infrastructure can be re-created or exported via REST API, Python/Go SDKs or Terraform (import of existing infrastructure is documented). Bare-metal egress is billed at 0.0013 EUR/GB when moving data out.
Security & compliance
- SOC 2
- not published
- ISO 27001
- yes (ISO/IEC 27001 stated on infrastructure and security pages)
- HIPAA
- not published
- GDPR / DPA
- yes (Data Processing Agreement on legal page; GDPR compliance stated)
- Other
- PCI DSS (Cloud passed annual QSA audit); most PoPs in Tier III/IV data centers; a Trust Center is referenced
- Legal entity
- G-Core Labs S.A.
- Funding
- not published
From the study · assessmentGcore, in short
What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.
Strengths
- Publishes on-demand, spot and 12/36-month reserved prices for 8-GPU bare-metal nodes in EUR and USD
- H100 SXM spot at USD 1.80 per GPU-hour (USD 14.40 per 8-GPU node) and 36-month reservation at USD 1.93
- B300 bare metal carries a public price (EUR 48.00 / USD 52.80 per 8-GPU node-hour); GB300 NVL72 by request
- InfiniBand 3.2 Tbit/s on H100/H200 nodes configured automatically; clusters of 250+ nodes supported
- Serverless Everywhere Inference billed per minute on L40S, A100 and H100 (1 to 8 GPUs); H100 at USD 1.70 per GPU-hour
- Formal 99.9% SLA with service credits for GPU Cloud, Bare Metal, Spot and Inference
- Official Terraform provider (v2 generated from OpenAPI), Python and Go SDKs, managed Kubernetes
- ISO 27001 and PCI DSS stated; DPA published; EU (Luxembourg) contracting entity; card, PayPal, bank, wallets
Weaknesses
- Bare-metal GPUs are sold per 8-GPU node only; single-GPU access is limited to Virtual GPU (H100 only, USD 2.60 per GPU-hour) or serverless inference
- Bare Metal GPU is billed from creation until deletion, even when powered off
- Bare-metal egress (0.0013 EUR/GB) and public IPv4 (2.5331 EUR/month) are extra, and the documentation contradicts itself on GPU egress
- No free trial for Cloud AI and no cryptocurrency payment
- MSA makes contracts contingent on a KYC check and allows annual indexation (higher of 5% or Luxembourg CPI) plus cost-based fee changes
- Published tables contain inconsistencies (Virtual GPU 8xH100 USD 10.8/h; GPU memory totals in the inference table; calculator vs table differences)
- Spot and reserved tabs not visible on the rendered page; B300 reservations and GB300 are contact-sales
Ideal for
- EU-contracted teams training on full 8-GPU H100/H200 nodes with InfiniBand, especially with 12-36 month commitments
- Buyers who need a formal SLA, ISO 27001/PCI DSS and invoice or bank payment
- Teams wanting serverless inference and GPU clusters from one provider, managed with Terraform
Not published
- Spot and Reserve tab values come from the pricing page's content data served by gcore.com
- the tabs themselves were not visible in the rendered text
- Spot discount per region and the minimum balance for spot: the docs point to a billing page that was not found
- File share (VAST) price per GB in GPU regions not found in the pages read
- GPU model of Cloud VM 'GPU Instances (Gen 1)' not named
- VRAM of A2 instances not stated
- per-region availability of these VMs not listed
- H100/A100 form factor for Virtual GPU and Everywhere Inference not stated, so those rows use model 'other'
- Everywhere Inference regions and whether custom Docker containers are accepted not verified on a primary page read
- Exact KYC documents required: MSA 2.17 requires KYC documentation, sign-up docs only require email, password and a payment method
- SOC 2 and HIPAA status not stated on the pages read (a Trust Center is referenced but was not read)
- No CLI for Cloud/GPU found in the docs read (only FastEdge CLI appears in the docs navigation)
- Funding and ownership not published on the pages read
- B300 VRAM per GPU given only as 2,304 GB per 8-GPU node
- B300 and GB300 regions not listed
Observations
gcore.com/cloud/ai-gpu redirects to gcore.com/gpu-cloud; product cards show 'From' prices in EUR: L40S 1.08, A100 1.26, H100 1.64, H200 2.27, B300 6.00 per hour; GB300 'Per request price'
The product page's H100 'From EUR 1.64' and H200 'From EUR 2.27' equal the spot per-GPU figures in the pricing data, while the pricing table shows on-demand EUR 2.27 (H100) and EUR 2.73 (H200) per GPU
gcore.com/pricing/ai shows prices without an account, in EUR by default with a EUR/USD icon switch, all excluding VAT
The GPU table headline reads 'Choose a pricing model, prices update across all GPUs', but the On-demand / Spot / Reserve 12mo / 'Reserver 36mo' tabs present in the page data were not rendered as visible text on 2026-09-29
Bare-metal prices are shown per 8-GPU node per hour with a per-GPU line; B300 and GB300 use 'Contact sales' buttons, the others 'Get started'
An interactive GPU Cloud Calculator (region, flavor, IP, egress GB) gives Hour/Month/Year totals; its node prices differ slightly from the table (A100 EUR 10.08 vs 10.16, L40S EUR 8.64 vs 8.72, H100 EUR 18.18 vs 18.16)
Everywhere Inference table lists 4xA100 and 4xH100 at 240 GB and 8xA100/8xH100 at 960 GB GPU memory; Virtual GPU 8xH100 shows USD 10.8/h against EUR 18.91/h
'Sign up for free' with email, Google or GitHub; FAQ answers mention 'upcoming GB200' and H100 'up to 100GB of HBM3 memory'
Status page lists Cloud AI maintenance windows per region (Chester, London-2, Sines-2, Sines-3)
Pages read
- gcore.com/gpu-cloud — From-prices, features, FAQ
- gcore.com/pricing/ai — Bare-metal on-demand, Everywhere Inference and Virtual GPU prices, GPU calculator
- gcore.com/assets/cached/api/en/pricing-pages/ai_pLevel-4.e5aea57a.json — USD figures, Spot and Reserve 12/36 month tabs
- bff.gcore.pro/cloud/ai-item-price — Per-region node prices, GPU region list and flavors (with /cloud/regions?type=ai, /cloud/ai-items
- bff.gcore.pro/cloud/volume-price — Volume, IPv4 (/cloud/ip-address-price) and bare-metal egress (/cloud/traffic-price) prices
- gcore.com/pricing/cloud — GPU VM instances (A2, Gen 1), object storage
- docs.gcore.com/edge-ai/ai-infrastructure/about-gpu-cloud — Cluster types, configurations, storage, egress statement
- docs.gcore.com/edge-ai/ai-infrastructure/spot-bare-metal-gpu — Spot reclaim policy, per-minute billing aggregated hourly
- docs.gcore.com/edge-ai/ai-infrastructure/create-a-virtual-gpu-cluster — Virtual GPU storage, shelving, provisioning time
- docs.gcore.com/edge-ai/ai-infrastructure/create-a-bare-metal-gpu-cluster — Region grouping, file shares, API/Terraform
- docs.gcore.com/edge-ai/billing — GPU billing rules, inference per minute
- docs.gcore.com/cloud/billing — PAYG per minute, EUR 4 minimum, stopped-instance rules, egress
- docs.gcore.com/account-settings/billing/overview — Threshold billing
- docs.gcore.com/account-settings/billing/payments — Payment methods, card verification hold, top-ups
- docs.gcore.com/account-settings/account/create-account — Sign-up requirements
- docs.gcore.com/account-settings/billing/trial-conditions — No Cloud AI trial
- docs.gcore.com/account-settings/billing/data-deletion — Retention after suspension
- docs.gcore.com/cloud/bandwidth-and-traffic — VM egress free, bare-metal egress billed
- docs.gcore.com/developer-tools/sdks/overview — Python and Go SDKs
- docs.gcore.com/developer-tools/terraform/overview — Official provider
- docs.gcore.com/cloud/getting-started/request-a-quota-increase — GPU Cloud quotas
- gcore.com/legal — Legal entity, KYC clause, price indexation, 99.9% SLA, spot terms
- gcore.com/secure-infrastructure — PCI DSS, ISO 27001, GDPR
- gcore.com/infrastructure — Data centers, certifications
- gcore.com/about — HQ, founding year
- gcore.com/everywhere-inference — Serverless inference product
- status.gcore.com/ — Status and maintenance
Questions about Gcore
What GPUs does Gcore offer?
Gcore has 6 GPU types with a fresh offer today, for example H100 from $1.71, A100 80GB from $1.41, H200 from $3.55, L40S from $1.21, B300 from $6.70 per GPU-hour on-demand. The full list with prices and market medians is on this page.
How does Gcore bill?
Billing granularity: per minute. Minimum charge: Edge Cloud needs at least EUR 4 / USD 4 on the balance to start; PAYG usage is metered about every 5 minutes and charged each time accrued costs reach EUR/USD 4. Spot clusters need a minimum balance (amount not found on the pages read). Stopped instance: Bare Metal GPU servers are charged from creation until deletion regardless of power state. Virtual GPU: shelving (power off) stops compute billing; volumes and IPs remain billed; the local data disk is wiped on shelving.
How trustworthy is Gcore?
Its Trust Score is 80/100 (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 Gcore's data centres?
10 named regions in 7 countries: Finland, Luxembourg, Portugal, Singapore, South Korea, United Kingdom, United States.
How do you pay Gcore, and are there credits?
Payment by card, invoice. Free credits: none: no trial for Cloud (AI, Basic VM, DDoS Protection) per trial conditions doc. Identity check: ID on request.