Cerebras ($CBRS.O) Equity Research Report
Nvidia Killer or AI Headfake?
Irrational Analysis is heavily invested in the semiconductor industry.
Please check the ‘about’ page for a list of active positions.
Positions will change over time and are regularly updated.
Opinions are authors own and do not represent past, present, and/or future employers.
All content published on this newsletter is based on public information and independent research conducted since 2011.
This newsletter is not financial advice, and readers should always do their own research before investing in any security.
At the time of publishing, I have no economic interest in Cerebras.
Feel free to contact me via email at: irrational_analysis@proton.me
Hello wonderful subscribers. Today’s content is attached as a ~10MB PDF. Not sure if it will make it successfully to your email inbox. Please check the website if needed.
The attached report has been typeset in LaTeX because this is a special occasion. Cerebras is a company I have followed since 2019. Huge fan of their unique technology. The wafer-scale engine really is a wonder of engineering.
Unfortunately, the IPO structure and their marketing behavior indicate that management seems to be more interested in seeking exit liquidity on public markets rather than fixing the issues with their technology and becoming a real competitor to Nvidia. My report is broadly negative because I want them to win and genuinely believe they are currently on a path to irrelevance.
I would like to take a moment to thank the nine peer reviewers who provided extremely valuable feedback on the technical, financial, and tone/vibe/structural aspects of this report.
Note: Several people have contacted me about paid consulting/advisory roles. I don’t do paid anything. That is against my employment contract and various ethical rules in my head. I love my day job and prefer to keep it. Happy to help out for free in my spare time if your query is sufficiently interesting.
Send me an email: irrational_analysis@proton.me
What if they could cut out the interface FPGAs and implement standard electrical SERDES PHYs on the WSE, with optical transceivers outside - like others are doing. Do you think this would make engineering and economic sense, if on-wafer optics is so difficult?
Latest News from cerebras on llama-405b : https://cerebras.ai/blog/llama-405b-inference