GROQ (GROQ) Competitors & Peers
Companies tied to GROQ through its disclosed supply-chain and competitive relationships. Numbers are code-computed, never estimated.
How these relationships were derived
Every relationship below carries its provenance. A confirmed edge is sourced from a public filing or disclosure and links to it; an AI-inferred edge is surfaced by the relationship model and is unverified — shown for color, not as evidence.
Confirmed relationships (cited)
Sourced from a public filing or disclosure — each links to its origin.
- NVDALicensor
NVIDIA licensed language processing unit (LPU) technology from Groq under a non-exclusive license agreement.
Confirmed — SEC filing / public disclosuresec.gov ↗
Model-inferred relationships (unverified)
Suggested by the relationship model, with no public source on file. Treat each as a lead worth checking, not as evidence.
- 005930.KSSupplier
Samsung Foundry manufactures Groq's next-generation Language Processing Unit (LPU) chips on its 4nm node.
Model-inferred, unverified — treat as a lead, not evidence - AMDCompetitor
AMD (with its MI300 series) is a major competitor fighting for market share in the AI inference hardware space alongside Groq.
Model-inferred, unverified — treat as a lead, not evidence - NVDACompetitor
Groq LPU vs NVIDIA GPU in inference.
Model-inferred, unverified — treat as a lead, not evidence - GFSSupplier
GroqLPU fabricated at GlobalFoundries 14nm; future generations to TSMC.
Model-inferred, unverified — treat as a lead, not evidence - CBRSCompetitor
Cerebras is a key startup competitor also designing custom silicon (Wafer-Scale Engines) for AI workloads.
Model-inferred, unverified — treat as a lead, not evidence - TENSTORRENTCompetitor
Tenstorrent is another highly valued AI chip startup competing with Groq for inference data center deployments.
Model-inferred, unverified — treat as a lead, not evidence - METAPartner
Groq extensively utilizes and optimizes for Meta's Llama models to demonstrate its hardware's inference speeds to the open-source community.
Model-inferred, unverified — treat as a lead, not evidence - GFSSupplier
GlobalFoundries fabricated Groq's earlier generation AI chips before the transition to Samsung's 4nm.
Model-inferred, unverified — treat as a lead, not evidence - EARTHLINKCustomer
EarthLink has partnered with Groq as an enterprise customer to deploy AI inference capabilities for its users.
Model-inferred, unverified — treat as a lead, not evidence