(Reuters) - Semiconductor startup Groq said on Monday it has raised $640 million in a Series D funding round led by Cisco Investments, Samsung Catalyst Fund and BlackRock Private Equity Partners, among others, bringing its valuation to $2.8 billion.

The Silicon Valley firm, founded by a former Alphabet engineer, specializes in producing AI inference chips - a type of semiconductor that optimizes speed and executes commands of pre-trained models.

Besides big firms such as Advanced Micro Devices, many startups including Groq have been trying to nibble away at Nvidia's dominant position in the booming AI chip industry.

Last year, Groq adapted Meta Platforms' (META.O), a large language model, LLaMA, to be able to run on its chips rather than those of Nvidia's (NVDA.O),. Meta researchers built LLaMA using Nvidia's chips.

Cloud service providers racing to develop their own AI products are also seeking alternatives to Nvidia's top-of-the-line processors due to high demand but limited supply.

In 2021, Groq was valued at $1.1 billion after funding from Tiger Global Management and D1 Capital.

"Groq will use the funding to scale the capacity of its tokens-as-a-service (TaaS) offering and add new models and features to the GroqCloud," the company said about the latest round of funds.

Groq will deploy more than 108,000 Language Processing Units manufactured by Global Foundries (GFS.O), by the end of the first quarter of 2025.

Groq has also announced the appointment of Stuart Pann, a former senior executive at Intel and HP Inc, as its chief operating officer, while Meta's chief AI scientist Yann LeCun was named its newest technical adviser.