This chip company plans to do something big

Latest update time：2024-03-17

Reads：

Source : Content compiled from axios by Semiconductor Industry Observation (ID: i c bank ), thank you.

????If you want to see each other often, please mark it as a star???? and add it to your collection~

Groq is a Mountain View, Calif.-based artificial intelligence chip startup backed by the likes of Tiger Global and Lee Fixel's Addition. The company told Axios it is preparing for a new round of funding next quarter after recently receiving a small cash infusion.

Why it matters: The company, which claims its artificial intelligence chips can be 10 times faster than traditional chips, was the focus of a recent firing at venture capital firm Social Capital.

The big picture: Groq, founded by Ross in 2016, designs chips specifically for inference, that is, running generative AI models. The company says its chips, known as "language processing units" (LPUs), are not only faster but also cost one-tenth the cost of traditional AI hardware.

The company currently has about 4,500 chips deployed, and Ross said it will have 1.5 million by the end of next year.

While at Google, Ross helped invent the company's Tensor Processing Unit (TPU), a chip designed specifically for machine learning.

To date, Groq has raised US$367 million, with the most recent round of financing in 2021 at a valuation of US$1.1 billion, led by Tiger Global and D1 Capital.

The startup's strategy is two-pronged. It sells access to developers who want to run artificial intelligence software on its cloud infrastructure and hardware to customers who run their own data centers.

It recently acquired Definitive Intelligence to solidify a new business unit focused on expanding its offerings to customers and developers.

The startup was founded in 2022 by Sunny Madra and Gavin Sherry and is backed by the likes of Social Capital. (Madera and Shirley also founded at least one other social capital portfolio company, Autonomic, before.)

Groq's only other acquisition was Maxeler Technologies in 2022, although Ross said there are certainly more to come.

The implication: As generative AI technologies, applications, and users explode, demand for specialized computer chips is skyrocketing, but supply cannot keep up.

Groq and other upstarts are betting they can get a piece of the market, and Ross predicts that inference will surpass training when it comes to demand for computing resources—a perfect opportunity for his company.

Groq is also positioning itself as a greener alternative to traditional infrastructure as concerns grow about the carbon footprint of artificial intelligence.

What they said: "If you go out and say you're going to spend $7 trillion, that's smart marketing and it's going to make people say 'Gosh, maybe we can't do it,'" Ross said of Sam Altman ( Sam Altman said in a public estimate of the capital required for the OpenAI chip business.

Ross said Groq would only need $700 billion to do the same thing because its chips are much cheaper.

Yes, but: in addition to funding, Groq needs partners to execute its vision (and bear some upfront costs).

The company has signed a deal with an unnamed partner to deploy more than 10% of its 1.5 million LPUs by the end of 2025.

Insider: On the topic of Social Capital, Groq said in a statement that the events "are internal company matters and we know very little about them at this time," adding that it will "determine how Jay Zaveri's departure from Social Capital will impact Groq Board".

The company declined to clarify whether Zaveri, a former Social Capital partner, still serves on its board.

What to watch: Ross predicts that Nvidia's big annual conference next week will include a significant discussion about inference.

For information about the Groq inference chip, interested readers can read " Challenging Nvidia requires a different approach "

Original link

https://www.axios.com/2024/03/16/artificial-intelligence-groq-social-capital

Click here to follow and lock in more original content

END

*Disclaimer: This article is original by the author. The content of the article is the personal opinion of the author. The reprinting by Semiconductor Industry Watch is only to convey a different point of view. It does not mean that Semiconductor Industry Watch agrees or supports the view. If you have any objections, please contact Semiconductor Industry Watch.

Today is the 3707th issue shared by "Semiconductor Industry Observation" with you. Welcome to pay attention.