WHMLAB

Bare-metal dedicated servers. No neighbours. No virtualization.

Single-tenant dedicated servers with full hardware access. AMD EPYC and Intel Xeon, ECC RAM, NVMe Gen4, 10 Gbps uplinks. IPMI included. Hourly or monthly billing. 10 global data centers. Your server, your rules.

Bare-metal lineup

Four server classes. Single-tenant. Your hardware, your kernel.

Compute

AMD EPYC 7402P

24 cores . 48 threads . 2.8 GHz base, 3.35 GHz boost

128 GB DDR4 ECC

2x 1 TB NVMe Gen4 (RAID-1)

1 Gbps unmetered

$149/mo

$0 setup . 24h deploy

Configure

Most picked

Performance

AMD EPYC 7443

24 cores . 48 threads . 4.0 GHz boost

256 GB DDR4 ECC

2x 2 TB NVMe Gen4 (RAID-1)

10 Gbps unmetered

$299/mo

$0 setup . 12h deploy

Configure

Memory

Intel Xeon Gold 6438Y+

32 cores . 64 threads . 4.0 GHz boost

512 GB DDR4 ECC

2x 1 TB NVMe Gen4 (RAID-1)

10 Gbps unmetered

$499/mo

$0 setup . 12h deploy

Configure

GPU

AMD EPYC 9354 + L40S

32 cores . 64 threads . 3.8 GHz boost

256 GB DDR5 ECC

2x 4 TB NVMe Gen5 (RAID-1)

25 Gbps unmetered

$1,899/mo

$0 setup . 48h deploy

Configure

Workloads

What a dedicated server is actually for.

Large databases

100 GB+ datasets that you would not dream of putting on shared infra. PostgreSQL, MySQL, MongoDB, ClickHouse, Elastic -- with full NVMe throughput and zero noisy neighbours.

Case: FinTech firm runs 8 TB PostgreSQL on Performance class. P99 query 47ms.

High-traffic web

Sites doing 1M+ visits/day. WordPress, Magento, custom PHP/Node. 10 Gbps link, NVMe storage, no contention with other tenants. CDN in front optional but not required.

Case: News portal handles 3M daily visitors on a single Performance server.

Game servers

Minecraft, Valheim, CS2, FiveM, Rust, Ark. Single-tenant means no tick rate loss when other tenants spike. NVMe for world saves, 10 Gbps for many concurrent players.

Case: Gaming community hosts 200-player Rust server on Compute class.

AI / ML inference

GPU-equipped servers for Stable Diffusion, LLM inference, computer vision. L40S / A100 / H100. NVMe Gen5 for dataset loading. Dedicated bandwidth for remote inference calls.

Case: AI startup runs 70B LLM inference on GPU class, 32 req/s.

Custom build

Not the right spec? We will build to order.

Custom configurations delivered in 5-7 business days. Engineering review free. Volume discounts above 3 units.

CPU

EPYC 7402P (24c)EPYC 7443 (24c)EPYC 9354 (32c)EPYC 9554 (64c)Xeon Gold 6438Y+ (32c)Xeon Platinum 8480+ (56c)Dual EPYC 9654 (192c)Custom

RAM

128 GB DDR4256 GB DDR4512 GB DDR41 TB DDR42 TB DDR4256 GB DDR5512 GB DDR51 TB DDR5

Storage

2x 1 TB NVMe2x 2 TB NVMe2x 4 TB NVMe4x 4 TB NVMe RAID-102x 8 TB NVMeSATA 16 TB HDD + NVMe cacheSATA 32 TB HDDCustom

GPU

NoneNVIDIA L4NVIDIA L40SNVIDIA A100 40GBNVIDIA A100 80GBNVIDIA H100 80GB2x L40S4x L4

Network

1 Gbps10 Gbps25 Gbps40 Gbps100 GbpsDual 10 Gbps LAGPrivate VLANCross-connect

Location

MumbaiSingaporeTokyoFrankfurtLondonVirginiaDallasAny 2+ for HA

Data centers

Tier III+ data centers across 6 continents.

Mumbai (BOM-1)

Tier III . ISO 27001 . SOC 2

STT GDC facility. 40 MW. Power: 2N. Cooling: N+1. 6 carriers. Cross-connects to Tata, Airtel, Jio.

Singapore (SIN-11)

Tier IV . PCI DSS . SOC 2

Equinix SG4. 45 MW. 2N+1 power. Liquid cooling ready. Direct connect to AWS, Azure, GCP.

Frankfurt (FRA-3)

Tier III . ISO 27001 . GDPR

Interxion. 30 MW. 100% renewable. DE-CIX on-site. Great for EU data residency.

Virginia (IAD-12)

Tier III . SOC 2 . HIPAA

CoreSite Reston. 50 MW. AWS Direct Connect. Lowest latency to US East Coast.

Tokyo (NRT-2)

Tier III . ISO 27001

NTT Data. 25 MW. JPIX on-site. Best for Japan + APAC latency.

Sydney (SYD-5)

Tier III . ISO 27001

NextDC S1. 20 MW. NZ-AU cable termination. Sole-tenanted cages available.

Customer stories

What teams do with their dedicated server.

"We moved a 6 TB PostgreSQL database from a hyperscaler to a dedicated server. Same workload now costs us $299/month instead of $2,800. P99 query latency went from 240ms to 47ms. Why did we wait so long?"

Anil Verma
CTO, LogisticsCo

"Game server hosting business. 40 dedicated servers in 3 regions. IPMI access is critical for us -- we deploy and manage remotely without filing tickets. The 10 Gbps uplinks handle our peak Friday nights."

Daniel Park
Founder, GamerHost

"AI inference on a single H100 dedicated server. Cheaper than AWS p5 instances and we own the hardware. We amortized the cost in 3 months. Now deploying 4 more in different regions."

Sarah Chen
ML Lead, NexaAI

Deploy a single-tenant server today.

From $149/month. IPMI included. 10 Gbps unmetered. 24h deploy.

Deploy a Server See Hardware