SKIP_TO_MAIN_CONTENT
06.0 // Comparison Apples to apples Same YAML

machine.dev vs
WarpBuild.

WarpBuild pitches "2× faster, 50% cheaper." machine.dev is 12.5× cheaper on spot, 26% cheaper on x64 on-demand, and 40% cheaper on ARM — because WarpBuild doesn't pass through the Graviton discount. Plus a GPU catalog they don't ship.

01 Intro
[REF_001]

BuildJet's aggressive heir.

WarpBuild has been the loudest voice chasing BuildJet's customer base since BuildJet announced their March 31, 2026 shutdown. They ship managed cloud runners, BYOC mode (bring your own AWS / GCP / Azure account), and a separate Docker Builders product line. Customer roster includes Sonar, Comcast, Sky, Braintrust, LanceDB, DDEV, and Kintsugi. SOC2 Type 2.

Their managed Linux runners are $0.002/vCPU/min — the same flat rate as Blacksmith and Depot. Where the comparison gets interesting is ARM: WarpBuild charges the same per-minute rate for ARM as for x64. machine.dev passes through the AWS Graviton baseline, which is ~13-19% cheaper than the equivalent x64 instance, so the gap on ARM is uniformly 40% — much wider than the 26% x64 gap.

Honest note: WarpBuild's BYOC mode (flat $0.002/min software fee + customer's AWS bill) is structurally cheaper than machine.dev on-demand at large sizes — same dynamic as RunsOn. We're not in that game; our wedge is managed-with-spot, plus GPU.

02 Cpu pricing
[REF_002] // LINUX_X64

x64 Linear, Both Ways.

Per-minute USD. WarpBuild managed prices verified from warpbuild.com/pricing April 9, 2026.

[NOTE_X64] Both providers scale linearly per vCPU on x64. WarpBuild at $0.002/vCPU/min; machine.dev at ~$0.00149/vCPU/min on-demand. The 26% on-demand gap and the 12.5× spot gap are uniform at every size.

[NOTE_BYOC] WarpBuild's BYOC mode is a flat $0.002/min software fee with customer's AWS / GCP / Azure rates passed through. At large sizes BYOC beats machine.dev on-demand. See ARM pricing below for the wider 40% gap.

02b Arm pricing
[REF_003] // LINUX_ARM64

ARM: 40% Cheaper.

WarpBuild charges the same per-minute rate for ARM as for x64. machine.dev passes through the AWS Graviton baseline (~13-19% cheaper than equivalent x64). Result: the on-demand gap on ARM is 40%, much wider than the 26% x64 gap.

02c Gpu pricing
[REF_004] // NVIDIA

No GPU vs Five Tiers.

WarpBuild has no GPU offering at any size. machine.dev offers T4G ARM, T4, L4, A10G, L40S, RTX 6000, plus AWS Trainium and Inferentia2 on the public per-minute catalog.

02c Features
[REF_004]

Beyond price.

SPEC WARPBUILD MACHINE_DEV
Pricing model Linear $0.002/vCPU/min managed; $0.002/min flat BYOC Per-minute, USD
Spot tier (managed) No Yes (12.5× cheaper than WarpBuild)
ARM Graviton discount No (same price as x64) Yes (~13-19% cheaper than x64)
GPU runners No Yes (T4G, T4, L4, A10G, L40S, RTX 6000)
AWS Trainium / Inferentia2 No Yes
CUDA pre-installed No (no GPU) Yes (12.1 + cuDNN 9.2.1)
Docker Builders product Yes (separate product line, $0.06-0.72/min) No (use actions runners + buildx)
BYOC mode (AWS/GCP/Azure) Yes ($0.002/min flat + customer cloud bill) No (managed only)
Compliance SOC2 Type 2 In progress
Free tier No free tier $10 free compute, no card
Setup One runs-on change One runs-on change
03 What you keep
[REF_003] // SCOPE_OF_CHANGE

The only line you touch.

Your workflow files, steps, secrets, caches, and Actions syntax stay exactly as they are. The only change is the runs-on label.

BEFORE.YAML DEFAULT
jobs:
train:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- run: python train.py
AFTER.YAML EXECUTE
jobs:
train:
runs-on: [machine, gpu=T4]
steps:
- uses: actions/checkout@v4
- run: python train.py
04 Where warpbuild wins
[REF_004]

Four strengths.

[REF_010] BYOC

BYOC mode for high-volume customers

WarpBuild's BYOC mode runs in your own AWS / GCP / Azure account at a flat $0.002/min software fee, with raw cloud rates passed through to your bill. For very high-volume teams that already run cloud at scale, BYOC's flat fee is structurally cheaper than any per-minute managed service. machine.dev doesn't compete in this mode.

[REF_011] DOCKER

Dedicated Docker Builders product

WarpBuild ships a separate Docker BuildKit product (16-192 vCPU) for remote Docker layer caching. machine.dev runs Docker on the same runner as your job and relies on actions/cache.

[REF_012] COMPLIANCE

SOC2 Type 2

WarpBuild ships SOC2 Type 2 today. machine.dev's compliance posture is in progress.

[REF_013] BUILDJET

Established BuildJet migration story

WarpBuild has been the most aggressive BuildJet migration target since BuildJet announced their March 2026 shutdown. They've published migration playbooks. If you're a BuildJet customer who already evaluated WarpBuild, you're already familiar with their flow.

05 Where machine dev wins
[REF_005]

Five strengths.

[REF_020] PRICE // SPOT

12× cheaper on spot at every size

WarpBuild's managed cloud runners have no spot tier — flat $0.002/vCPU/min. machine.dev's spot CPUs are 11.4-15.5× cheaper at every size, on both x64 and ARM. If you can use spot, the math is unambiguous.

[REF_021] PRICE // ARM

40% cheaper on every ARM on-demand size

WarpBuild charges the SAME rate for ARM as for x64 — no Graviton discount. machine.dev passes through the AWS Graviton baseline (~13-19% cheaper than equivalent x64), so the on-demand ARM gap is uniformly 40%, much wider than the 26% x64 gap.

[REF_022] PRICE // OD

26% cheaper on every x64 on-demand size

Both providers scale linearly per vCPU on x64 managed. WarpBuild is $0.002/vCPU/min; machine.dev's effective rate is ~$0.00149/vCPU/min. The 26% gap holds at every size from 2 to 32 vCPU.

[REF_023] GPU // STRUCTURAL

GPU runners — they have none

WarpBuild is CPU-only. machine.dev offers six NVIDIA tiers (T4G ARM, T4, L4, A10G, L40S, RTX 6000) plus AWS Trainium and Inferentia2. If your CI pipeline touches model evaluation, training, or any CUDA workload, machine.dev is the only option between the two.

[REF_024] GPU // T4G_ARM

The only ARM GPU runner anywhere

T4G ARM 8 vCPU spot is $0.00277/min ($0.166/hr) — the cheapest GPU runner on the public internet. WarpBuild doesn't offer ARM GPU at any price.

06 Live pricing
[REF_PRICING] // LIVE_DATA

Live per-minute rates

See what you'd actually pay. Compare with GitHub's $0.006/min for 2-core, $0.012/min for 4-core, and $0.052/min for GPU.

07 Initialize
READY
> SYSTEM_READY // AWAITING_INPUT

Get started free.

$10 free compute on signup. Enough to run real workloads and compare cost before committing.