tracking 9 providers · pricing sourced from public APIs & pricing pages

Compare real GPU and bare-metal pricing. Not vendor marketing.

ComputeFare replaces hours of tab-switching between provider pricing pages. Answer four questions, get three ranked matches, skip straight to the provider that fits your workload.

computefare --workload=training --gpu=h100
$→ scoring 9 providers against workload profile...
$→ 3 matches ranked by fit
#1
Latitude.shH100 self-serve · api-first · minutes to provision
#2
HyperstackEU-native GPU · below-hyperscaler hourly pricing
#3
Vast.aimarketplace · cheapest $/GPU-hr · verify per-host SLA
providers tracked
9
affiliate confirmed
2
pricing last verified
pending — verifying now
commission-influenced ranking
no
// workflow

Three steps. No email required.

01

Describe your workload

Four questions: workload type, GPU tier, budget, region.

02

Get scored matches

Three ranked providers with a one-line 'matched because' explanation.

03

Go straight to the provider

Outbound to the provider's signup. No email wall. No sales funnel.

// preview

A slice of the comparison.

see all 9 providers →
providerstarting
DigitalOcean
digitalocean
$—.––details →
Vultr
vultr
$—.––details →
Hetzner
hetzner
$—.––details →
// why computefare

Independent, in plain language.

Providers can't pay to rank higher. Scoring is deterministic and based on your inputs.

We may earn commission when you sign up through us — this is disclosed on every outbound link.

Commission rates are never shown per-provider on the public site. That's business-internal.

Pricing is pulled from providers' public pages and APIs, then rechecked periodically.

Full detail: affiliate disclosure →

// faq

Straight answers.

// optional

Get notified when GPU pricing changes.

A short email when an H100/A100 provider drops or raises hourly pricing. Nothing else.

// stub — no backend wired yet. structured to drop in Mailchimp / ConvertKit later.