Cerebras IPO Filing Reveals Strong Financials and Massive Backlog

Cerebras, known for its unique Wafer-Scale Engine technology, now seeks public investment after years of private development. This specialized processor uses an entire silicon wafer, containing 4 trillion transistors and 900,000 cores, a departure from conventional AI systems that link thousands of smaller Graphics Processing Units (GPUs). This design aims to streamline AI computing, particularly for inference, where models process new data. Founders notably declined an acquisition offer from Elon Musk in 2018, choosing to pursue this distinct chip architecture independently.

The IPO filing details Cerebras' rapid financial growth. The company generated $510 million in revenue during 2025, marking a 76% increase year-over-year. This growth reflects a substantial shift in profitability, as Cerebras transitioned from a $481 million net loss in a prior period to an $87.9 million net income. Furthermore, as of December 31, the company disclosed a significant $24.6 billion backlog in remaining performance obligations, indicating substantial future contracts.

These financial disclosures validate Cerebras' distinct approach within the competitive AI hardware market, which NVIDIA currently leads. The reported revenue and profit figures, coupled with a large backlog, demonstrate commercial adoption for its wafer-scale technology across diverse customers. This IPO positions Cerebras as the first major AI infrastructure company to test public investor interest in the core components powering the artificial intelligence boom. The market's reception of Cerebras' debut will likely influence how future AI hardware IPOs are received.