Can you own it?
This page is a projection of the one entry record, the Use & modify and Transparency factors that Assess covers. The full verdict is set by all four factors together, floor-weighted so the weakest caps the whole.
- AssessUse & modify + Transparency
- ImplementData control + Doesn't fail you
- UseDoesn't fail you
- SupportTransparency
Data governance - retention, ZDR & training
The governing documents - the Groq Services Agreement and the your-data docs, both read this session - say inference is not retained by default and is never trained on: Groq "is not permitted to use Inputs or Outputs for training or fine-tuning any AI Model Services or other models" and "will never access Customer Data for training purposes." Zero data retention is available to all customers through Data Controls, not gated to an enterprise tier. Because Groq serves open-weight models on its own LPU hardware, there is no closed third-party model whose data terms would attach to routed inference.
The caveat is in the non-inference features: batch processing retains input/output files for 30 days unless deleted, and fine-tuning retains weights and datasets until the customer deletes them, with that retained data stored in US GCP buckets. So the "not retained" headline holds for streaming inference but not automatically for batch or fine-tuning workflows - plan those data lifecycles explicitly.
Data & IP ownership
The Services Agreement is explicit: the customer "retains all Intellectual Property Rights in Customer Data (including in Inputs and Outputs)," and Groq "does not access, use, store, or retain Inputs or Outputs except as necessary to provide the Cloud Services." Combined with never-training and open-weight-only serving, your data and the knowledge derived from it stay yours. The DPA discloses a published sub-processor list at trust.groq.com/subprocessors with a 15-day written objection right - a correction to the earlier "not disclosed" reading. The caveats are lifecycle rather than training: the batch (30-day) and fine-tuning features retain data in US GCP storage until you delete it. Contractualise the specifics through the DPA and BAA.
Compliance & attestations
The trust centre states SOC 2 Type II is maintained (2025 report). The DPA - read this session - covers GDPR, CCPA and PDPL, and a HIPAA BAA is offered (the BAA excludes the Compound system as a non-covered service). ISO 27001 / 27701 are unconfirmed, and the SOC 2 report itself is gated behind request rather than read here. For ISO-dependent or healthcare procurement, request the current attestation letters and countersign the BAA directly.
Data residency & jurisdiction
Default storage is US, with a European data centre in Helsinki, Finland launched (with Equinix) on 6 July 2025. The EU footprint is therefore real but very recent, region pinning is unconfirmed, and the DPA expressly permits Processing in the United States and other countries - so treat EU residency as an emerging capability to validate in onboarding rather than a mature guarantee. As a US-headquartered company, Groq also carries standing US CLOUD Act exposure regardless of region.
Security controls
Documented controls include SOC 2 Type II scope (trust centre), Data Controls for enabling ZDR, and the Services Agreement's commitment that Groq does not access, store or retain Inputs or Outputs except to provide the service. No independently reviewed penetration test surfaced, and the SOC 2 report itself was not read this session - so this is assessed as solid-but-partial: day-to-day controls are documented, but the deepest independent assurance is not visible.
Pricing & cost model
Pricing is per-token with a rate-limited free/developer tier, read from Groq's own pricing page. Confirmed rates include Llama 3.3 70B $0.59/$0.79 and gpt-oss-120B $0.15/$0.60 per M in/out; other example rates (Llama 3.1 8B $0.05/$0.08, DeepSeek R1 Distill Llama 70B $0.75/$0.99) are approximate, so treat them as an order-of-magnitude guide and confirm live rates before committing.
Reliability posture
The trust centre attests operational controls, but Groq is a fast-scaling hardware startup with no public uptime SLA or status history independently verified this session. Reliability is therefore assessed conservatively: adequate on available evidence, but without an observable status/SLA record to lean on for availability-critical workloads.
How this scores
The ownership factors this domain covers, drawn from the one entry record.
Use and modify freelyCan you use it freely and leave without lock-in?
StrongOpenAI-compatible API (docs read: base URL https://api.groq.com/openai/v1) over open-weight models keeps workloads portable - the same checkpoints run elsewhere or self-hosted - and the Services Agreement confirms the customer retains all IP in Inputs/Outputs; the custom LPU and narrower catalogue are coverage limits, not exit blockers.
TransparencyAre the binding terms published, legible and independently checkable?
ModerateThe governing Services Agreement and DPA were read this session and are legible - they state customer IP retention, never-train, retention windows and a published sub-processor list - but the SOC 2 report itself is gated (not read) and ISO 27001/region-pinning are unconfirmed.
Sources
The same evidence records as the entry sheet. Read means the text was verified; unverified means it is known to exist but not yet read.