AI chip startup Groq is in advanced talks to raise nearly $600 million in a new funding round that would value the company at around $6 billion, according to Bloomberg. he funding, still being finalised, is led by Austin-based venture firm Disruptive, which has reportedly committed more than $300 million to the deal. While terms could still change, the raise would mark a significant leap in Groq’s valuation, nearly doubling it from the $2.8 billion valuation just under a year ago.
A fast-growing funding trajectory
This latest funding round comes on the heels of Groq’s August 2024 raise, where the company secured $640 million at a $2.8 billion valuation. That round was led by BlackRock and included high-profile backers such as Neuberger Berman, Type One Ventures, Cisco, KDDI, and Samsung Catalyst Fund.
With the upcoming round led by Disruptive, Groq’s total funding will exceed $1 billion, solidifying its position as one of the most well-capitalised startups in the AI hardware space. The potential doubling of its valuation in under a year reflects growing market confidence in Groq’s strategic direction and technology.
From Google to Groq: A founder’s journey
Groq was founded in 2016 by Jonathan Ross, a former Google engineer best known for developing the Tensor Processing Unit (TPU) chip at the tech giant. With a vision to design purpose-built chips optimised for AI, Ross took Groq out of stealth mode with a clear focus: to accelerate AI workloads with unmatched speed, quality, and efficiency.
What Groq builds: Hardware for AI inference
At its core, Groq specialises in AI inference, the process of running trained models in real-time. The company has developed its own proprietary Language Processing Units (LPUs)—custom-designed chips that process complex natural language tasks efficiently. These LPUs are tailored for applications where real-time response is critical, such as autonomous vehicles, advanced robotics, and large-scale data centres.
Unlike general-purpose GPUs, Groq’s chips are built specifically for inference, offering an edge in energy efficiency, speed, and scalability. The company provides both cloud-based and on-premises solutions, appealing to enterprises looking for flexible, high-performance AI infrastructure.
Strategic partnerships signal enterprise demand
In April, the company teamed up with Meta to support Llama 4 inference, helping to scale one of the most powerful large language models currently in development. Then, in May, Groq announced an exclusive partnership with Bell Canada to power the telecom giant’s large-scale AI infrastructure.
These collaborations show that Groq’s chips are being chosen by some of the biggest names in tech and telecom to power their AI ambitions. While Groq and Disruptive have not commented publicly on the latest funding report, these deals suggest that Groq is becoming a trusted player in delivering cutting-edge AI infrastructure.
A Nvidia challenger?
While Nvidia remains the dominant force in the AI chip market, Groq is carving out a distinct position by focusing exclusively on inference. Its purpose-built hardware and software stack offers specific performance advantages that general-purpose chips may not match, especially in low-latency, high-throughput environments.
Groq’s growing presence and rapid valuation increase underscore the opportunity for niche players to thrive in the AI hardware ecosystem, particularly as companies require tailored compute solutions rather than one-size-fits-all chips.
Positioned for the next phase of AI growth
With a strong foundation in both technology and funding, Groq is poised to continue scaling. The additional capital from this raise is expected to support expanded production, global partnerships, and further integration of hardware and software for enterprise-grade deployments.
As generative AI continues to evolve and demand for real-time inference grows, Groq’s LPUs could become a central part of the AI infrastructure landscape. The company’s next chapter, backed by billions in valuation and support from industry-leading investors, may help redefine how enterprises build and deploy AI at scale.
