Chip-making startup Cerebras Systems Inc. has selected Citi as the lead bank for its initial public offering (IPO), according to people familiar with the matter.
The Zhitong Finance App learned that, according to people familiar with the matter, chip manufacturing startup Cerebras Systems Inc. has selected Citi as the lead bank for its initial public offering (IPO).
People familiar with the matter said that the Silicon Valley company chose the bank after discussions with potential advisors about the US listing. According to reports in January, Cerebras plans to go public as early as the second half of this year and may seek a valuation of more than $4 billion in 2021 financing in an IPO.
People familiar with the matter said that the details provided by Cerebras's IPO may change, and more advisors may join the company's lineup. Representatives from Cerebras and Citi declined to comment.
Last month, social media platform Reddit (RDDT.US) and semiconductor connectivity company Astera Labs (ALAB.US) successfully went public in the US, laying the foundation for more companies' IPOs in the future. Cloud and data security startup Rubrik Inc. (RBRK.US), digital marketing software company Ibotta Inc. (IBTA.US), and commodity dealer Marex Group Plc (MRX.US) have all submitted IPO applications in the past two weeks.
According to a statement at the time, Cerebras raised $250 million in Series F funding in 2021, with a valuation of over $4 billion. The funding round was led by Alpha Wave Ventures, Abu Dhabi Growth Fund, and G42. According to the statement, Cerebras' current investors include Altimeter Capital, Benchmark Capital, and Coatue Management.
Benchmark Nvidia
According to information, Cerebras introduced its latest Wafer Scale Engine 3 chip (WSE-3 chip) earlier.
According to the company, the WSE-3 chip was specially built to train the industry's largest AI model. The new device is built using TSM.US's 5 nm process and contains 4 trillion transistors and 900,000 AI cores. In contrast, the Nvidia (NVDA.US) H100 chip contains 80 billion transistors.
It is worth noting that when introducing the performance parameters of the WSE-3 chip, Cerebras fully benchmarked this product against the Nvidia H100. The information on the company introduction page shows that in terms of accelerated artificial intelligence training, the chip's performance is 8 times that of the H100.
“You've probably seen Cerebras say its platform is easier to use than Nvidia's,” the company said in a press release. An important reason for this is the way Cerebras stores weights and activations, and it doesn't have to scale to multiple GPUs in the system and then to multiple GPU servers in the cluster.”