Cerebras CEO Reiterates Intent to Go Public

Cerebras CEO Andrew Feldman has addressed the recent withdrawal of the company’s IPO registration, acknowledging the company’s misstep in failing to immediately explain the rationale behind the decision.

In a LinkedIn post published late Sunday, Feldman stated that Cerebras remains committed to going public but has undergone significant transformation since its initial filing a year prior. The company intends to amend its prospectus to reflect these changes before proceeding with a public offering.

“Given that the business has improved in meaningful ways we decided to withdraw so that we can re-file with updated financials, strategy information including our approach to this the [sic] rapidly changing AI landscape,” Feldman wrote.

The withdrawal announcement followed shortly after Cerebras secured $1.1 billion in pre-IPO funding at an $8.1 billion valuation. Notable investors in this round included Tiger Global and 1789 Capital, where Donald Trump Jr. is a partner, none of which were disclosed in the initial 2024 IPO filing, according to Feldman.

“We made this call because it’s in the best interest of our investors, partners, and team — and it will allow potential investors to better understand the value of the business when we enter the public markets,” Feldman stated, without specifying a timeline for refiling.

Cerebras has positioned itself as a provider of large-scale chips tailored for training and deploying AI models. This year saw the company’s expansion into cloud services, operating data centers designed to handle AI model requests directly. This transition underscores a strategic pivot towards capturing recurring revenue streams and offering a more complete AI infrastructure solution.

At the heart of Cerebras’ value proposition is the assertion that its hardware surpasses the performance of graphics processing units (GPUs), a market currently dominated by Nvidia, with Advanced Micro Devices (AMD) vying for market share. AMD has recently announced that OpenAI committed to setting up to 6 gigawatts’ worth of the company’s AI processors and could end up owning 10% of the chipmaker, highlighting the intensifying competition in the AI accelerator space.

Cerebras’s approach, centered on its wafer-scale engine (WSE), aims to deliver superior computational density and memory bandwidth compared to traditional GPU-based solutions. While Nvidia holds a strong lead in the GPU market due to its established software ecosystem (CUDA) and extensive customer base, Cerebras is betting on the potential for its architecture to unlock new levels of performance and efficiency, particularly for large-scale AI models that push the limits of existing hardware.