Cerebras Systems Inc., a designer of massive AI accelerators and operator of data‑center facilities, filed paperwork to launch a U.S. IPO that could fetch as much as $3.5 billion. The filing places the company among a wave of semiconductor firms courting public markets after a surge in demand for artificial‑intelligence hardware. Investors will weigh the offering against rivals such as Nvidia and other emerging chipmakers.

Cerebras’ business model hinges on scaling wafer‑scale engines that promise orders of magnitude faster training for deep‑learning models. By monetising both the chips and the associated cloud‑based services, the firm hopes to capture a slice of the multibillion‑dollar AI spend that has accelerated since 2020. The IPO proceeds would fund expanded fab capacity and broaden its data‑center footprint.

Wall Street analysts see the filing as a litmus test for how capital markets value niche AI hardware players versus established giants. A successful listing could validate Cerebras’ claim that its wafer‑scale approach offers a defensible edge, while a tepid reception might pressure the company to reassess pricing or seek private funding. Either outcome will shape the competitive dynamics of the AI chip sector.