← All companies

Chips / Company profile

Groq

What Groq does

Groq is a private U.S. AI infrastructure company that builds a purpose-designed AI inference processor architecture (its LPU / GroqChip line) and delivers those capabilities through its GroqCloud “neocloud” offering. The company’s positioning focuses on ultra-low-latency, cost-efficient inference for production workloads, with the product surface delivered via an API (Groq API) and a developer console (GroqCloud). Groq describes an integrated “from silicon to cloud” approach: it controls both the hardware (inference-focused accelerators) and the software stack, and it operates global inference capacity through data centers that customers and partners can access for running AI models.

GroqCloud is aimed primarily at developers and AI-native enterprises building real-time or latency-sensitive applications across modalities; Groq’s API is designed to be mostly compatible with OpenAI client libraries to reduce integration friction. In addition to the public GroqCloud service, Groq also references enterprise deployment options such as dedicated/private instances (depending on customer needs), and it emphasizes operational governance and control for enterprise and regulated use cases.

Strategically, Groq’s most material recent shift is an increased focus on scaling inference capacity as an operating business (“AI inference cloud”), alongside deeper ecosystem alignment with NVIDIA through a non-exclusive licensing agreement (Dec 2025) and later NVIDIA Cloud Partner certification (Aug 2026). In 2026, Groq announced two large capital raises that explicitly tie proceeds to expanding its inference footprint and scaling capacity over time, including a stated path toward 200MW+. Overall, Groq is attempting to convert its inference technology differentiation into an inference-capacity platform with large-scale data-center operations and a growing API/developer ecosystem.

News

Company record

Aug 17, 2026 · Official · Groq NewsroomGroq Closes $350 million Series A, Building the World's Leading AI Inference Cloud

Groq announced a $350M Series A led by Disruptive (with planned NVIDIA participation) valuing the company at $3.5B, tied to scaling global inference capacity.

Aug 12, 2026 · Official · Groq NewsroomGroq Becomes an NVIDIA Cloud Partner

Groq announced it joined NVIDIA’s NVIDIA Cloud Partner (NCP) program, positioning certification as a formal step in its collaboration to design, deploy, and operate NVIDIA accelerated computing to NVIDIA operational standards.

Jun 22, 2026 · Official · Groq NewsroomGroq Raises $650M to Scale Its AI Inference Cloud Business

Groq announced $650M in growth capital led by Disruptive and Infinitum, describing an operating plan around scaling GroqCloud capacity and targeting growth toward 200MW+.

Dec 24, 2025 · Official · Groq NewsroomGroq and Nvidia Enter Non-Exclusive Inference Technology Licensing Agreement to Accelerate AI Inference at Global Scale

Groq announced a non-exclusive licensing agreement with Nvidia for Groq’s inference technology, stating Groq will continue to operate as an independent company with leadership changes tied to CEO transition.

Dec 18, 2025 · Official · Groq NewsroomGroq Partners with U.S. Department of Energy to Advance AI Inference and Next-Generation Computing Infrastructure

Groq announced a partnership with the U.S. Department of Energy, describing work intended to advance AI inference and next-generation computing infrastructure.

Nov 17, 2025 · Official · Groq NewsroomGroq Expands to Asia-Pacific with Sydney Data Center to Power the Next Generation of AI Inference

Groq announced an Asia-Pacific expansion with a Sydney data center, positioning it as closer capacity for Australian organizations and the public sector.

Oct 28, 2025 · Official · Groq NewsroomGroq Powers HUMAIN One, a Real-Time AI Operating System for Enterprise

Groq announced that HUMAIN selected Groq to power HUMAIN One, positioning low-latency inference as critical for real-time enterprise voice-driven operations.

Jul 06, 2025 · Official · Groq NewsroomGroq Launches European Data Center Footprint in Helsinki, Finland

Groq announced its first European data center footprint in Helsinki, describing expansion in collaboration with Equinix and emphasizing low latency and scalability.

Show 2 earlier updates
Source map · 2 recurring channels · 19 references

Still resolving: Newsroom · Official X

Funding

Latest disclosed valuation$3.5B
Tracked capital$2.1B7 sourced rounds

What it builds

GroqCloud

GroqCloud is Groq’s inference platform delivered for running AI models via a managed cloud service, presented by Groq as the “premier neocloud for fast inference.”

source ↗
Groq API (GroqCloud API)

Groq provides an API for creating and running chat completions and other AI inference workflows using GroqCloud-backed inference, with GroqDocs describing OpenAI-compatible client library behavior.

source ↗
Groq LPU™ Inference Engine

Groq’s LPU inference engine is Groq’s inference technology intended to deliver real-time inference performance, and it is described as part of GroqCloud’s end-to-end offering.

source ↗
GroqChip™ Processor / GroqNode™ server

Groq sells/describes inference-focused compute components including the GroqChip™ processor and GroqNode™ server chassis as part of its hardware stack for inference capacity deployments.

source ↗

Milestones & partnerships

Aug 12, 2026
Joined NVIDIA Cloud Partner (NCP) program

Groq announced it joined NVIDIA’s NVIDIA Cloud Partner program to design, deploy, and operate NVIDIA accelerated computing to NVIDIA’s reference architecture and operational standards.

NVIDIA
Dec 24, 2025
Non-exclusive inference technology licensing agreement with NVIDIA

Groq and Nvidia entered a non-exclusive licensing agreement for Groq’s inference technology; Groq stated it would continue operating independently and announced CEO transition timing.

NVIDIA
Dec 18, 2025
Partnership with U.S. Department of Energy (DOE)

Groq announced a partnership with DOE to advance AI inference and next-generation computing infrastructure.

U.S. Department of Energy
Nov 17, 2025
Sydney data center expansion (Asia-Pacific)

Groq announced a data center footprint in Sydney to expand low-latency inference capacity for organizations across Australia and the public sector.

Jul 06, 2025
First European data center footprint in Helsinki with Equinix collaboration

Groq announced its European footprint in Helsinki, describing collaboration with Equinix to deliver scalable, low-latency inference capacity.

Equinix
May 28, 2025
Exclusive inference provider for Bell Canada’s Bell AI Fabric

Groq announced an exclusive partnership with Bell Canada to power Bell AI Fabric, described as a sovereign AI infrastructure project across multiple sites.

Bell Canada
Apr 29, 2025
Partnership for fast inference for the official Llama API

Groq and Meta announced collaboration to deliver fast inference for the official Llama API, referencing Groq LPU acceleration.

Meta