Marketplace · San Francisco, CA
Hyperbolic
US GPU cloud that resells H100, H200 and B200 capacity from an aggregated supplier network, self-serve on prepaid credits, with short self-serve reservations and private clusters.
GPU types3with a current offer
Current offers1617 readings on file · last 45 days
H100 from$4.85per GPU-hour, on-demand
Entry price$4.49cheapest GPU-hour listed
Countries13 named regions
Trust Score v0.158/100
- Company & track record73
- Infrastructure control60
- Pricing & billing transparency85
- Reliability signals45
- Security & compliance20
- Customer signal—
Scored Sep 29, 2026 from the provider's own pages. Dimensions not yet measured are excluded from the denominator. Criteria ↓
Trust Score · 5 of 6 dimensions measuredThe score, criterion by criterion
Each point comes from something the provider publishes, with the page it was read on and the date. Anything not disclosed earns zero.
Company & track record
weight 15%
73 / 100
| Founding year published | not disclosedhyperbolic.ai · Sep 29, 2026 | 0 / 4 |
| Headquarters published | Hyperbolic Labs, Inc.hyperbolic.ai · Sep 29, 2026 | 3 / 3 |
| Legal entity named | Hyperbolic Labs, Inc.hyperbolic.ai · Sep 29, 2026 | 4 / 4 |
| Funding or ownership disclosed | Private; founders Jasper Zhang (CEO) and Yuchen Jin (CTO); investors not named on pages readhyperbolic.ai · Sep 29, 2026 | 4 / 4 |
Infrastructure control
weight 15%
60 / 100
| Capacity model published | third-party hostshyperbolic.ai · Sep 29, 2026 | 4 / 10 |
| Regions published | 3 regions or countrieshyperbolic.ai · Sep 29, 2026 | 5 / 5 |
Pricing & billing transparency
weight 20%
85 / 100
| Public prices without an account | yes (read on the pricing page)hyperbolic.ai · Sep 29, 2026 | 5 / 5 |
| Billing granularity published | per hourhyperbolic.ai · Sep 29, 2026 | 4 / 4 |
| Egress policy published | included (no egress fees)hyperbolic.ai · Sep 29, 2026 | 4 / 4 |
| Stopped-instance billing published | No stop state documented: on-demand instances bill for their lifetime until terminated; termination deletes on-machine datahyperbolic.ai · Sep 29, 2026 | 4 / 4 |
| Public IPv4 policy published | not disclosedhyperbolic.ai · Sep 29, 2026 | 0 / 3 |
Reliability signals
weight 20%
45 / 100
| Status page | not disclosedhyperbolic.ai · Sep 29, 2026 | 0 / 6 |
| SLA published | 99.5% uptime for On-Demand and Reserved; Private Cloud customhyperbolic.ai · Sep 29, 2026 | 6 / 6 |
| SLA with a numeric uptime commitment | 99.5% uptime for On-Demand and Reserved; Private Cloud customhyperbolic.ai · Sep 29, 2026 | 3 / 3 |
| Uptime history published | not disclosedhyperbolic.ai · Sep 29, 2026 | 0 / 5 |
| Observed availability (30 days) | not measured yet | — |
Security & compliance
weight 15%
20 / 100
| SOC 2 | Type II in progress (pursuing); Trust Center lists SOC 2hyperbolic.ai · Sep 29, 2026 | 0 / 8 |
| ISO 27001 | not disclosedhyperbolic.ai · Sep 29, 2026 | 0 / 4 |
| HIPAA or GDPR DPA | Not HIPAA-certified; BAA possible for qualified use caseshyperbolic.ai · Sep 29, 2026 | 3 / 3 |
Customer signal
weight 15%
not measured
| Aggregated public reviews | not collected yet | — |
provisional v0.1: evidence from the provider's own pages; dimensions not yet measured are excluded from the denominatorFull method →
Prices by GPU · 2 GPU types with a current offerWhat Hyperbolic charges, against the market
Cheapest current on-demand true cost here, next to the market median for the same GPU. Negative means cheaper than the market.
| GPU | Cheapest here | Market median | vs market | Current offers | Modes | |
|---|---|---|---|---|---|---|
| H100 | $4.85 | $3.17 | +53% | 2 | on-demand | H100 prices |
| H200 | $4.49 | $4.76 | −6% | 7 | on-demand | H200 prices |
All offers · 17 readings on fileEvery Hyperbolic offer, with its receipt
Current offers (last 45 days) ranked by true cost, all rental modes, each with its read date. Older readings stay listed below, in grey.
| GPU | Configuration | Mode | Region | Stock | List | True cost | Read | |
|---|---|---|---|---|---|---|---|---|
| H200marketplace floor | 8× H200224 vCPU · 2,048 GB RAM · 14,307 GB disk | On-demand | US | In stock | $4.49 | $4.49 | 21:11 UTCfresh | Receipt |
| H200 | 8× H200160 vCPU · 2,048 GB RAM · 15,202 GB disk | On-demand | Japan | In stock | $4.49 | $4.49 | 21:11 UTCfresh | Receipt |
| H200 | 16× H200448 vCPU · 4,096 GB RAM · 28,614 GB disk | On-demand | US | In stock | $4.49 | $4.49 | 21:11 UTCfresh | Receipt |
| H200 | 16× H200320 vCPU · 4,096 GB RAM · 30,404 GB disk | On-demand | Japan | In stock | $4.49 | $4.49 | 21:11 UTCfresh | Receipt |
| H200 | 24× H200672 vCPU · 6,144 GB RAM · 42,921 GB disk | On-demand | US | In stock | $4.49 | $4.49 | 21:11 UTCfresh | Receipt |
| H100SXMmarketplace floor | 1× H100 SXM20 vCPU · 240 GB RAM · 720 GB disk | On-demand | US | In stock | $4.85 | $4.85 | 21:11 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM20 vCPU · 240 GB RAM · 720 GB disk | On-demand | Canada | In stock | $4.85 | $4.85 | 21:11 UTCfresh | Receipt |
| H200 | 1× H20024 vCPU · 240 GB RAM · 720 GB disk | On-demand | US | In stock | $4.92 | $4.92 | 21:11 UTCfresh | Receipt |
| H200 | 1× H20024 vCPU · 240 GB RAM · 720 GB disk | On-demand | US | In stock | $4.92 | $4.92 | 21:11 UTCfresh | Receipt |
9 fresh offers · prices per GPU-hour · true cost includes the add-ons the provider publishesOpen in the comparison tool →
True cost & billing · read Sep 29, 2026What the hourly price does not say
Billing granularity, minimums, what a stopped machine costs, traffic, addresses, deposits and identity checks, as published by the provider.
Billing
- Billing granularity
- per hour
- Minimum charge
- No minimum charge, but balance must cover 1 hour of runtime of all running plus new instances at creation
- Stopped instance
- No stop state documented: on-demand instances bill for their lifetime until terminated; termination deletes on-machine data
- Egress
- included (no egress fees)
- Public IPv4
- not published
- Billing model
- prepaid credits
- Minimum deposit
- 5 USD one-time to unlock GPU instances (Pro tier); credits never expire
- Free credits
- Phone-verification credit banner shown in plan page screenshot (amount not stated)
Account
- Payment
- card · crypto (not published)
- Sign-up
- email, password or Google/GitHub, SSH key, 5 USD card deposit via Stripe, KYC on request (name, address, date of birth, ID document)
- Identity check
- ID on request
- Support
- in-app chat (Pylon), email
- Enterprise support
- not published
- SLA
- 99.5% uptime for On-Demand and Reserved; Private Cloud custom
- Status page
- not published
- Published uptime
- not published
Infrastructure
- Regions
3 named regions in 1 countries
North America (None) · Europe (None) · United Kingdom (United Kingdom)- Countries
- —
- Platform
- API, SSH, custom Docker images, persistent volumes, bare metalKubernetes: managed (Private Cloud only)Interconnect: InfiniBand up to 3.2 Tb/s (NDR, ConnectX-7) for multi-node; NVLink 900 GB/s H100/H200, 1.8 TB/s B200/B300Clusters: Self-serve multi-node clusters 8 to 128+ GPUs; Private Cloud with managed Kubernetes or Slurm
- Storage
- network · not published (rates vary by region) · Storage volumes 100 GB to 10 TB, billed hourly on provisioned size, select regions only
- Migration out
- Standard Linux VMs or bare metal with SSH; data on storage volumes survives instance termination and can be copied out with no egress fee. Reserved instances cannot be terminated early.
Security & compliance
- SOC 2
- Type II in progress (pursuing); Trust Center lists SOC 2
- ISO 27001
- not published
- HIPAA
- Not HIPAA-certified; BAA possible for qualified use cases
- GDPR / DPA
- DPA on request, EU data residency options
- Other
- not published
- Legal entity
- Hyperbolic Labs, Inc.
- Funding
- Private; founders Jasper Zhang (CEO) and Yuchen Jin (CTO); investors not named on pages read
Read on the provider's own pages · 16 pages read · facts marked "not published" were looked for and not foundPages read ↓
Assessment · documents read Sep 29, 2026Hyperbolic, in short
What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.
Strengths
- No charge for instances that fail to become ready; billing starts only when SSH-accessible
- Per-GPU rate locked for the life of the instance
- Self-serve reservations from 7 days, and multi-node InfiniBand clusters without sales
- No egress fees; VM or bare-metal choice
- 5 USD minimum deposit; credits never expire
Weaknesses
- On-demand H100 SXM from 3.19 USD/GPU-hr, H200 from 3.99, B200 from 5.99
- Capacity comes from third-party suppliers; the app showed all GPU instances in use on 2026-09-29
- Hourly billing, and no stop state: instance termination deletes on-machine data
- Balance must cover one hour of all instances; instances can be terminated when the balance hits zero
- SOC 2 still in progress; not HIPAA certified
- Storage volume prices not published and only in select regions
Ideal for
- Teams needing H100/H200/B200 multi-node clusters self-serve for days to weeks
- Researchers wanting prepaid, no-contract access with rate lock
Not published
- Per-configuration on-demand prices (only 'starting from' published
- app showed no available configurations)
- Storage volume price per GB
- Exact data center countries and cities (only North America, Europe, United Kingdom)
- Crypto payments (not documented)
- Founding year, HQ city and investors
- B300 and RTX PRO 6000 prices
- Status page URL
Observations
Registry pricing URL /pricing returns 404; prices appear only as 'starting from' figures on /marketplace
Marketplace comparison table still mentions RTX 4090, RTX 3080 and RTX 3070, which the app no longer lists
App GPU page requires no login but showed 'All GPU Instances Currently In Use' with no configurations
Pages read
- hyperbolic.ai/marketplace — Starting-from GPU prices, payment by card/Stripe, no charge for failed instances
- hyperbolic.ai/docs/on-demand/pricing.md — On-demand and reserved billing mechanics, storage billing
- hyperbolic.ai/docs/general/billing-payments.md — 5 USD minimum, prepaid credits
- hyperbolic.ai/docs/general/account-management.md — Plan tiers, phone verification credit
- hyperbolic.ai/docs/overview/platform-comparison.md — GPU catalogue, SLA, interconnect, bare metal, no egress fees
- hyperbolic.ai/docs/on-demand/overview.md — Regions, 99.5% SLA, card payment
- hyperbolic.ai/docs/reserved/overview.md — Reservation mechanics
- api.hyperbolic.xyz/v2/on-demand/reserve-options — Reservation terms and discounts
- app.hyperbolic.ai/gpus — RTX-6000-PRO filter, full capacity, up to 10% reservation banner
- hyperbolic.ai/docs/general/security-compliance.md — SOC 2 in progress, HIPAA, GDPR
- hyperbolic.ai/docs/general/support.md — Support channels and hours, SLA
- hyperbolic.ai/terms — Legal entity, KYC clause, California law
- hyperbolic.ai/docs/overview/quickstart.md — Stripe payment, 5 USD deposit
- hyperbolic.ai/about — Founders
- trust.hyperbolic.ai/ — SOC 2 listing, AWS subprocessor
- hyperbolic.ai/privacy — Headquarters
Questions about Hyperbolic
What GPUs does Hyperbolic offer?
Hyperbolic has 3 GPU types with a fresh offer today, for example H100 from $4.85, H200 from $4.49 per GPU-hour on-demand. The full list with prices and market medians is on this page.
How does Hyperbolic bill?
Billing granularity: per hour. Minimum charge: No minimum charge, but balance must cover 1 hour of runtime of all running plus new instances at creation Stopped instance: No stop state documented: on-demand instances bill for their lifetime until terminated; termination deletes on-machine data
How trustworthy is Hyperbolic?
Its Trust Score is 58/100 (v0.1), read from its own documents. Weaker on infrastructure control, reliability signals, security & compliance. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.
Where are Hyperbolic's data centres?
3 named regions in 1 countries: —.
How do you pay Hyperbolic, and are there credits?
Payment by card, crypto. Free credits: Phone-verification credit banner shown in plan page screenshot (amount not stated). Identity check: ID on request.