small compared to SpaceX, very large compared to other similar specialized companies
Cerebras Systems executed the largest semiconductor Initial Public Offering (IPO) in history, raising $5.55 billion. Pricing above the expected range at $185 per share, the dinner-plate-sized chipmaker achieved a fully diluted valuation exceeding $63 billion. [1, 2, 3, 4]
Cerebras' unprecedented market debut follows massive strategic deals, including a $20 billion capacity agreement with OpenAI and a deployment deal with Amazon Web Services. [1]
The Story Behind Cerebras’ $63 Billion IPO with Founder and CEO Andrew Feldman - YouTube @NoPriors podcast
Cerebras founder and CEO Andrew Feldman discussing the company's journey to a $63 billion IPO and their mission to revolutionize AI computing.
Key takeaways:
- Wafer-Scale Innovation: Cerebras differentiates itself by building massive chips (the size of a dinner plate) specifically for AI, offering 15x-20x faster inference speeds than traditional GPUs (1:36 - 3:45).
- Overcoming the Chasm: The company survived a difficult period of being ahead of market demand by proving their technology in supercomputing centers and partnering with G42 (8:42 - 10:59).
- Market Explosion: The recent surge in demand for AI, particularly for fast inference, led to massive growth, a $20 billion deal with OpenAI, and a partnership with AWS (1:45 - 2:38, 24:00 - 24:50).
- Operational Culture: Feldman emphasizes the importance of remaining a "professional David" against tech giants and maintaining a fearless engineering culture despite scaling to over 800 employees (14:30 - 17:47).
Nvidia acquired Groq (competition)
Cerebras: Cerebras CS-3 vs. Groq LPU
The Cerebras CS-3 outperforms Groq’s LPU-based solution across almost all key metrics, delivering ~6x higher inference speeds on frontier LLMs, enabling more generation in the same amount of time, with higher accuracy and lower power consumption – at similar cost. With Cerebras, developers can build the fastest and most intelligent conversational AI, real-time code generation, instant reasoning, and agentic applications.
