
"Cerebras Systems is preparing for a new IPO in the United States. The AI chipmaker plans to file its application with the SEC next week, with its sights set on a stock market listing in the second quarter of 2026. The news was reported on first by Reuters. The company withdrew its previous IPO in October, a few days after raising more than $1 billion at a valuation of $8 billion. This was already the second delay: Cerebras submitted paperwork to the US stock market regulator in 2024, but postponed it after a national security investigation."
"The earlier delay was related to an investigation by the Committee on Foreign Investment in the United States (CFIUS). It focused on the minority stake held by G42, a tech conglomerate from the United Arab Emirates. US authorities feared that companies from the Middle East could give China access to advanced US AI technology. Sources report that G42 is no longer listed as an investor in the new application. The reason for this is unknown. Cerebras announced earlier this year that it had received approval from CFIUS."
"Cerebras is known for its wafer-scale engines, chips designed to accelerate the training and inferencing of large AI models. The company competes directly with Nvidia and other AI chipmakers. Currently, Nvidia has a huge lead in terms of adoption, with the added advantage that it now has the deepest pockets in the industry thanks to billions of dollars in quarterly revenue. Only Google could operate independently of Nvidia's GPUs from one day to the next without any major problems thanks to its Tensor Processing Units (TPUs)."
Cerebras Systems plans to file an SEC application next week and aims to list on the stock market in the second quarter of 2026. The company withdrew a previous IPO in October after raising more than $1 billion at an $8 billion valuation, following an earlier postponement after a national security investigation. The delay related to a CFIUS probe into a minority stake held by G42; sources say G42 is no longer listed and Cerebras received CFIUS approval earlier this year. Cerebras develops wafer-scale engines to accelerate training and inference and competes with Nvidia, claiming higher efficiency and speed.
Read at Techzine Global
Unable to calculate read time
Collection
[
|
...
]