GMI Cloud SCALE

gmicloud.ai · Compute · 1 record · JSON

GMI Cloud SCALE Compute

fails: identityinstitutionno cardDwrite-gated

www.gmicloud.ai/en/company/scale · inference API · compute/gmi-cloud-scale · JSON · compare

Free tier	a six-week equity-free cohort for AI startups: USD 500 of model-as-a-service credits on acceptance, 1:1 credit matching up to USD 4,500 for money the team adds, a USD 3,000 GPU voucher on completion, and partner credits the vendor totals at 'Up to $40K+ in Structured Credits'. unit hours; metered on wall-clock; grant
Price	none stated
Per-object cap	not applicable
Retention	not applicable
Egress	not stated
Identity	institution not published
Limits	write-gated cohort admission is the gate: an application form, a selected cohort of about 25 companies a cycle, and a team that is building a product. The USD 4,500 match also requires the team to spend USD 4,500 first, so the headline credit is not the credit received. source cohort admission is the gate: an application form, a selected cohort of about 25 companies a cycle, and a team that is building a product. The USD 4,500 match also requires the team to spend USD 4,500 first, so the headline credit is not the credit received.
Binding	write-gated cohort admission is the gate: an application form, a selected cohort of about 25 companies a cycle, and a team that is building a product. The USD 4,500 match also requires the team to spend USD 4,500 first, so the headline credit is not the credit received.
Interface	rest inferred from the tooling the provider documents; not probed
Tooling	not stated
Composes with	the GPU voucher is the only part that buys iron: USD 3,000 / USD 2.60 an hour = 1,154 H100-hours if the voucher is spendable at list rate, which the page does not state.
Reliability	no reliability event recorded
Notes	The 1:1 match is a rebate on spend, not a grant: receiving USD 4,500 requires paying USD 4,500. The 40K figure aggregates partner credits from other vendors (an OpenAI for Startups allotment of USD 5,000 in API credits appears in the partner list), so it is not GMI compute. The cheapest card is H100 at USD 2.60 an hour and the page marks B200 'Limited Availability', so the cheap class is the one usually in stock. Cohort 02 ran Sep-Oct 2026 and applications for Cohort 3 were open on 2026-10-22.
Evidence rung	documented
Checked	2026-10-22 (live); stale on 2026-11-21

Evidence (2)

date	source	what it proved
2026-10-22	www.gmicloud.ai/en/company/scale	'6-week Equity-Free Startup Program for AI-native teams'; 'Up to $40K+ in Structured Credits'; 'Phase 1 - Onboarding Credits $500 MaaS Credits on acceptance'; '1:1 credit matching up to $4,500'; 'For every $1 you add to MaaS, GMI matches $1 in credits', quoted
2026-10-22	www.gmicloud.ai/en/pricing	on-demand rates 'NVIDIA H100 GPU from $2.60/GPU-hour', 'H200 from $3.20/GPU-hour', 'B200 from $5.00/GPU-hour', 'GB200 from $8.00/GPU-hour', quoted

More fields (1)

{
 "wave2": {
  "class": "startup-cohort",
  "free_tier_kind": "grant",
  "shape": {
   "vcpu": null,
   "ram_gib": null,
   "gpu_class": "H100",
   "gpu_count": 1
  }
 }
}

Unknown or withheld values (9)

$.egress.metered egress-unknown
$.egress.terms not-stated
$.identity_cost.detail process
$.per_object_cap.notes process
$.per_object_cap.stated not-applicable
$.reliability_note none-observed
$.retention.notes process
$.retention.stated not-applicable
$.tooling not-stated