← Back to pricing

Dedicated Servers

A whole machine, not a slice of one. Pick a configuration below and we rack it, provision it, and put it behind the same Heyo control plane you already use — images, pools, networking, and TLS included. No noisy neighbours, no hypervisor tax, no cluster to babysit.

Bare metal, without becoming a datacenter team

01

Dedicated capacity

Every core, every gigabyte, and the full network port belong to you. Nothing is oversubscribed and nothing is shared, so performance on a Tuesday afternoon looks exactly like performance at 3am.

02

Bare metal

Real hardware with hardware virtualization exposed, so Firecracker micro-VMs boot at full speed. Bring your own kernel, pin your own CPUs, and run nested workloads that a shared cloud instance will not allow.

03

Managed control plane

heyctl and heyvm handle builds, deploys, routing, autoscaling, and certificates on top of your box. You get the machine; we keep the platform running on it.

Pick a machine to get started

Monthly pricing, billed per server. Every configuration includes the managed control plane, provisioning, and remote hands. Need something that isn’t listed — more GPUs, more disk, a private rack — tell us what you need.

Dedicated server configurations
Configuration Memory Storage Network Price Order
AMD Ryzen 5600X DS-01 · 6 cores 128 GB 2 TB NVMe (980 Pro) 1 Gbps unmetered $334/mo Get started →
AMD Ryzen 5950X DS-02 · 16 cores 64 GB 2 TB NVMe (980 Pro) 1 Gbps unmetered $348/mo Get started →
AMD Ryzen 5900X DS-03 · 12 cores 128 GB 2 × 2 TB NVMe (990 Pro) 1 Gbps unmetered $364/mo Get started →
AMD Ryzen 5950X DS-04 · 16 cores 64 GB 4 TB NVMe 1 Gbps unmetered $368/mo Get started →
AMD Ryzen 5900X DS-05 · 12 cores 128 GB 2 × 2 TB NVMe (980 Pro) 1 Gbps unmetered $380/mo Get started →
AMD Ryzen 5600X DS-06 · 6 cores 64 GB 24 TB + 22 TB SATA, 2 × 4 TB NVMe 1 Gbps (10G port) $520/mo Get started →
AMD Ryzen 5950X DS-07 · 16 cores 128 GB 2 × 4 TB NVMe 1 Gbps (10G port) $538/mo Get started →
AMD Ryzen 5950X + GPU DS-08 · 16 cores 128 GB 2 × 4 TB NVMe 1 Gbps unmetered $598/mo Get started →
AMD EPYC 7402P DS-09 · 24 cores 128 GB 4 TB NVMe 1 Gbps (10G port) $698/mo Get started →
AMD EPYC 7552 DS-10 · 48 cores 128 GB 2 × 4 TB NVMe 1 Gbps (10G port) $798/mo Get started →
2 × Xeon E5-2680 v4 DS-11 · 28 cores 384 GB 4 × 6 TB SATA 1 Gbps (10G port) $798/mo Get started →
AMD Ryzen 9950X DS-12 · 16 cores 192 GB 2 × 4 TB NVMe 1 Gbps unmetered $816/mo Get started →
2 × Xeon E5-2680 v4 DS-13 · 28 cores 384 GB 2 × 4 TB SSD, 2 × 2 TB NVMe 1 Gbps (10G port) $860/mo Get started →
AMD EPYC 7502P DS-14 · 32 cores 256 GB 4 TB NVMe 1 Gbps (10G port) $956/mo Get started →
NVIDIA DGX Spark (GPU) DS-15 · 20 cores 128 GB 4 TB NVMe 1 Gbps (10G port) $998/mo Get started →
AMD EPYC 7443P DS-16 · 24 cores 256 GB 2 × 3.84 TB NVMe 1 Gbps (10G port) $1,052/mo Get started →
AMD EPYC 7443P DS-17 · 24 cores 256 GB 2 × 4 TB NVMe 1 Gbps (10G port) $1,052/mo Get started →
AMD Ryzen 9950X3D + GPU DS-18 · 16 cores 96 GB 2 × 2 TB NVMe (990 Pro) 1 Gbps (10G port) $1,382/mo Get started →
AMD EPYC 9254 DS-19 · 24 cores 512 GB 2 TB NVMe (990 Pro) 1 Gbps (10G port) $1,398/mo Get started →
AMD EPYC 9254 DS-20 · 24 cores 512 GB 4 TB NVMe (9100 Pro) 1 Gbps (10G port) $1,598/mo Get started →
AMD Threadripper PRO 9965WX DS-21 · 24 cores 512 GB 2 × 4 TB NVMe 1 Gbps (10G port) $1,998/mo Get started →
AMD EPYC 9254 DS-22 · 24 cores 512 GB 4 × 4 TB NVMe (990 Pro) 20 Gbps (40G port) $3,998/mo Get started →

Common questions

How long does provisioning take?

Most configurations are racked and handed over within one business day. Anything custom — unusual disk layouts, extra GPUs, a private rack — we’ll quote a date when you ask.

Is there a contract or setup fee?

No setup fee and no long-term contract. Dedicated servers bill monthly and you can cancel at the end of any billing period.

Can I mix dedicated servers with my own hardware?

Yes. Dedicated servers join the same virtual network as the machines you run yourself, so a hybrid fleet — a box under your desk and a box in our rack — behaves like one pool.

Do I still bring my own model keys?

Always. Dedicated servers run Anthropic, OpenAI, Gemini, Mistral, or local models via Ollama — and with a whole machine to yourself, local inference is a realistic option.

Need a machine of your own?

Tell us the workload and we’ll point you at the right configuration — or build one that isn’t on the list.