Rebellions secures $124 million in funding to develop new AI Rebel chips with Samsung | TechCrunch

rebelliona Korean fabless An AI chip startup said today that it has completed a $124 million (165 billion won) Series B round of financing to develop its third AI chip, Rebel. The startup will also use the new funds, oversubscribed to an initial target of $90 million, to ramp up production of its Atom data center chips and recruit.

Rebellions Chief Financial Officer Sungkyue Shin told TechCrunch in an exclusive interview that the three-year-old startup’s post-Series B funding valuation is about $658 million (880 billion won). The latest capital infusion brings total funding to approximately $210 million Rebellions has raised since its founding in 2020.

South Korean telecom giant KT led the latest funding round as a strategic investor, with previous investors Temasek Holdings’ Pavilion Capital and Korea Development Bank also participating, as well as new investors including Korelya Capital and DG Daiwa Ventures .

Rebellions’ funding comes at a critical time for the chip industry, particularly surrounding the development and use of chips for artificial intelligence.

Nvidia is the leader in the AI ​​chip market, and its name is synonymous with the AI ​​craze currently sweeping the tech world. many have Observed How about NVIDIA Flourish Part of the reason is moat Create an ecosystem around hardware and software. But the rest of the field is far from finished. Data processing and the associated high costs remain a major issue for AI applications, so people continue to look for innovations to groundbreakingly improve these.

Development comes from many fronts.Big tech giants, e.g. Google, Amazon, apple and Microsoft Open AI CEO Sam Altman reportedly develops or has his own chips to integrate AI into their products and services Visited South Korea last week Meet Samsung and SK Hynix, leaders in the country’s chip industry. In addition, Open AI Said to have raised billions of dollars Set up a chip manufacturing factory, Make your own AI chipIn addition to Rebellions, there are many new startups bringing new concepts to speed up processing while increasing efficiency.

Cooperate with Samsung

This fundraiser—has Rumors for months — followed by other moves by the startup. Last October, Rebellions declare It will work with Samsung Electronics to develop the latest Rebel chips to establish a partnership with Samsung Electronics initially forged The two companies aim to complete the development of Rebel by the end of this year and start mass production in 2025, Shin said, adding that the next generation of AI chips will target the generative AI market (LL.M.) running large language models and ultra-large scale enterprise.

Shin told TechCrunch that Rebel will use Samsung Electronics’ 4nm manufacturing process, and its AI chips will be deployed on Samsung’s advanced storage chip technology HBM3E, designed to handle high-bandwidth memory for building and running large language models. Rebellions’ unique selling point is the claim that its technology and products are more versatile than custom AI chips, meaning they can support a variety of generative AI models that require AI accelerators.

The company’s CFO emphasized that Rebellions will cooperate with Samsung from joint development, chip design to Rebel mass production. Samsung has a second motive for doing things here: in addition to its efforts in chips, South Korea’s largest memory chip manufacturer has been Committed to research and development in the chip field. own generative artificial intelligence model, Samsung Gaussian.

atoms and ions

It has also been working with customers using previous generations of chips. In May 2023, KT, a strategic investor in Rebellions, installed Atom, Rebellions’ data center-targeted artificial intelligence chip, in its cloud-based neural processing unit (NPU) infrastructure. It expects to generate revenue from Atom in the second half of this year and will continue to produce the chip model through Samsung’s 5nm manufacturing process. Atom is designed for data centers and language models with up to 7 billion parameters, while Rebel targets larger-scale language models, Shin noted.

At the same time, the startup’s first artificial intelligence chip, Ion, was launched in November 2021 and is currently undergoing qualification testing in the United States and has not yet signed up any commercial customers. Ion is designed for edge computing and one key use case, which the company believes will be in financial services applications, is that large institutions building their own hardware could use the chips to support stock forecasting and trading applications.

Rebellions CEO Sunghyun Park (a former quantitative developer at Morgan Stanley in New York) and four co-founders founded this AI chip startup in 2020.

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button