COMPANIES · AI CHIPS & NETWORKING

Groq

Designs LPU (language processing unit) chips optimized for fast AI inference, and operates a cloud inference service.

DirectAI chips & networking
What it does

Designs LPU (language processing unit) chips optimized for fast AI inference, and operates a cloud inference service.

Why it matters

It built one of the fastest AI-inference chips on the market, enough that NVIDIA itself moved to license its technology in December 2025 rather than compete with it outright.

How it makes money

Sells inference cloud services and licenses its chip architecture.

Geography

United States

AI stack position

AI chips and networking. The companies designing the processors, accelerators and networking equipment that do the actual computing behind AI, including GPUs like NVIDIA's, and the high-speed connections that link thousands of them together.

Connections (10)

Every connection below is tied to a readable source, date, classification and statement of its limits. A company source establishes what that company stated; it does not necessarily provide independent confirmation.

Current relationships (2)

Equinix provides data center infrastructure to Groq
Provides infrastructure toCurrent
Groq's May 28, 2025 newsroom post identifies Equinix as the data center partner hosting a Groq inference facility in Dallas (alongside a separate DataBank facility in Houston).

Source: 2025-05-28 · reviewed 2026-09-22

Evidence: Primary-source statement · Reported operating

Limits: Confirms colocation at an Equinix facility; does not disclose contract value, capacity (MW) or contract duration.

Groq newsroom: sovereign, fast AI inference infrastructure ↗
Groq provides fast inference for Meta Platforms
Provides computing capacity toCurrent
Meta said at LlamaCon that early experimental access to Llama 4 models powered by Groq was available by request through its Llama API developer platform.

Source: 2025-04-29 · reviewed 2026-09-24

Evidence: Primary-source statement · Reported operating

Limits: Describes an experimental, request-only preview as of April 2025. It does not establish commercial terms, volumes, that the option is still offered, or that Groq runs any of Meta's own internal workloads.

Everything we announced at our first-ever LlamaCon ↗

Dated or announced (6)

NVIDIA licenses inference technology from Groq
Uses technology fromDated / announced
Reporting on a deal disclosed December 24-26, 2025 states NVIDIA agreed to pay nearly $20 billion to license chip technology from Groq and to hire Groq CEO Jonathan Ross and other engineers, structured as an asset/IP licensing deal rather than a full acquisition of the company.

Source: 2025-12-26 · reviewed 2026-09-22

Evidence: Secondary reporting (unconfirmed by primary parties) · Announced

Limits: Reported deal structure and price come from press coverage, not a joint NVIDIA/Groq statement fetched directly in this research; does not establish that Groq ceased operating as an independent company, and closing status as of the audit review date is unconfirmed.

Nvidia to License Groq Technology in $20 Billion AI Chip Deal ↗
Groq licenses inference technology to NVIDIA
Partners withDated / announced
Groq and NVIDIA entered a non-exclusive technology licensing agreement covering Groq's inference technology.

Source: 2025-12-24 · reviewed 2026-09-17

Evidence: Primary-source statement · Licensing agreement announced

Limits: Primary source frames this as a non-exclusive licensing agreement; widely covered in the press as a roughly $20 billion 'acquisition' or 'acquihire' of Groq, a framing this record does not adopt without primary confirmation.

Groq and NVIDIA enter non-exclusive inference technology licensing agreement ↗
Cisco Systems invested in Groq
Invested inDated / announced
Groq's own August 5, 2024 announcement of its $640 million Series D round names Cisco Investments among the participating investors, led by BlackRock Private Equity Partners.

Source: 2024-08-05 · reviewed 2026-09-22

Evidence: Primary-source statement · Announced

Limits: Does not disclose Cisco's individual investment amount; a separate law-firm summary describes Cisco Investments as also participating in Groq's later Series E.

Groq Raises $640M to Meet Soaring Demand for Fast AI Inference ↗
BlackRock invested in Groq
Invested inDated / announced
The same August 5, 2024 release names BlackRock Private Equity Partners as the lead investor in Groq's $640 million Series D round, which valued the company at $2.8 billion.

Source: 2024-08-05 · reviewed 2026-09-22

Evidence: Primary-source statement · Announced

Limits: Does not disclose BlackRock's specific investment amount or resulting ownership stake; BlackRock Private Equity Partners is a private-markets unit of BlackRock, Inc.

Groq Raises $640M to Meet Soaring Demand for Fast AI Inference ↗
Samsung Electronics invested in Groq
Invested inDated / announced
The same release names Samsung Catalyst Fund -- Samsung Electronics' strategic investment arm -- among the participants in Groq's $640 million Series D round.

Source: 2024-08-05 · reviewed 2026-09-22

Evidence: Primary-source statement · Announced

Limits: Investment was made through Samsung Catalyst Fund rather than Samsung Electronics' balance sheet directly; individual amount not disclosed. This is distinct from Samsung Electronics' separate foundry/manufacturing relationship with Groq.

Groq Raises $640M to Meet Soaring Demand for Fast AI Inference ↗
Samsung Electronics manufacturing Groq's next-generation LPU chips at Samsung Foundry Groq
Manufactures forDated / announced
Groq selected Samsung Foundry as its silicon manufacturing partner to produce Groq's next-generation LPU chips on Samsung's SF4X (4nm-class) process.

Source: 2023-08-15 · reviewed 2026-09-20

Evidence: Primary-source statement · Manufacturing agreement announced in 2023 as a forward-looking commitment; current production status not reconfirmed by a newer joint statement

Limits: The original announcement describes a forward-looking manufacturing commitment for next-generation chips; it does not by itself confirm current wafer volumes, exclusivity, or that production is presently active.

Groq Selects Samsung Foundry to Bring Next-gen LPU to the AI Acceleration Market ↗

Comparisons (2)

Cerebras Systems competing as alternative AI inference chip and cloud providers Groq
Competes withComparison only
Cerebras and Groq both market purpose-built chip architectures and cloud inference services positioned as faster alternatives to GPU-based AI inference, and are commonly grouped together as NVIDIA challengers in inference hardware.

Source: 2025-10-21 · reviewed 2026-09-20

Evidence: Analytical comparison · Ongoing market competition in the AI inference hardware/cloud segment

Limits: Reflects a category overlap in inference-chip and inference-cloud positioning; does not reflect market share, head-to-head customer wins, or financial performance.

Cerebras, Groq And SambaNova Line Up To Compete With Nvidia ↗
Groq competing as alternative AI inference chip and cloud providers SambaNova Systems
Competes withComparison only
Groq and SambaNova both market purpose-built chip architectures and inference services positioned as faster or more efficient alternatives to GPU-based AI inference, and are commonly grouped together as NVIDIA challengers.

Source: 2025-10-21 · reviewed 2026-09-20

Evidence: Analytical comparison · Ongoing market competition in the AI inference hardware/cloud segment

Limits: Reflects a category overlap in inference-chip and inference-cloud positioning; does not reflect market share, head-to-head customer wins, or financial performance.

Cerebras, Groq And SambaNova Line Up To Compete With Nvidia ↗
Evidence for this profile

Groq newsroom ↗
Reviewed

Explore it interactively

Open this profile in the database See it on the relationship map

Inclusion in this database is not an endorsement of a company or its securities. Nothing here is a recommendation to buy or sell anything, and none of it is individualized investment advice. See the disclosures. Data: AI Map v3.22, released .