Was genuinely excited about this one. Now ran the numbers and decided the upside doesn't justify the price anymore. The tech is legitimately impressive, their chip does AI inference 20x faster than Nvidia's H100, and that speed advantage is real, not just marketing. They have OpenAI, AWS, and Meta as customers with $24.6B in signed contracts.
But the bankers have cooked it.
IPO expected price range history this week alone:
– Monday: $115–125
– Thursday: $125–135 (Bloomberg)
– Sunday night: $150–160 (Reuters)*
Three raises in under a week. 20x oversubscribed apparently. Great for the company, bad for us.
At $160 the numbers don't work for me:
– Nvidia, PS 24x, gross margin 75%, hundreds of customers
– Broadcom, PS 31x, gross margin 68%, 6 major customers
– AMD, PS 9x, gross margin 55%, dozens
– Cerebras (at $160), PS 67x, gross margin 43%, 3 major customers
Cerebras will be more expensive than all of them on trailing revenue, has the worst margins, and the most concentrated customer base.
The only justification is forward revenue, if OpenAI executes and 2026 revenue hits $1.1B+, then the forward P/S drops to ~30x (Broadcom territory). But OpenAI is reportedly burning $14B this year and their CFO has apparently warned internally about paying future compute contracts.
Will it pop on day one? Probabably, as stock float is tiny and demand insane. But it's not based on any real numbers anymore.
Reuters news: Cerebras to raise IPO price range to $150-$160 as demand surges, sources say: https://www.reuters.com/legal/transactional/cerebras-raise-ipo-price-range-150-160-demand-surges-sources-say-2026-05-10/
Cerebras IPO is Thursday and I'm out at $160
byu/charon-the-boatman instocks
Posted by charon-the-boatman
12 Comments
If it makes you feel better, fidelity wont even let me apply for shares because I dont have 500,000$. Lmao
That sucks I was excited for this one
Don’t you think they’re able to gain more than 3 customers now they’ll have all this funding?
More customers = more sales = lower P/S?
Companies don’t go public to raise funds to support current business levels.
Cerebras revenue for 2025 was $510M. Thats only 10x to $5B. They’re going to be at $50B within months,
100x revenue is the same as nebius. Iren is at 50x 2025 revenue ($500m / $24.4B market cap)
Why would you compare these “small cap” ai names with trillion dollar companies. There is absolutely no reason Cerebra’s won’t trade at similar numbers as the other smaller names
It’s going to pass $10B easily on day one. The rich will be richer, and I’ll get allocated 0 stocks as usual
Thanks for this analysis, I had also an eye out on this one but will wait until the 6 month lock up period is up!
I think that’s the right way to look at it. A great business and a bad entry price can both be true. Once an IPO turns into an auction powered by scarcity and FOMO, it stops being about conviction and starts being about whether you want to outguess the next buyer. Walking away when the numbers stop working is a win, not missing out.
Every single IPO is overpriced and pumped up before the average Joe can buy in. It’s always just a gamble if you buy in.
What’s the market cap at $160? Share price means nothing on its own.
Speed advantage is only partially real. Only if the models are small enough to benefit from fitting on chip.
They’re going to have to prove that they can scale before I invest in Cerebras. Good luck with the IPO and let’s see how many more racks they can ship after a couple of years. I might not get in at the group floor, but I am not convinced that the Cerebras is not just benefiting from companies buying any and all compute they can find. I’d rather get in at a higher price and have better visibility of the company’s future.
100% agreed. Withdrawn, waited for heat, AI heat is here. That’s what you’re buying. But, honestly? It could still be amazing. Peolpe have wildlly underestimated demand for tech in the past. [https://ipogrid.com/news/2026-05-11-cerebras-ipo-ai-capacity-buildout](https://ipogrid.com/news/2026-05-11-cerebras-ipo-ai-capacity-buildout)
> and that speed advantage is real, not just marketing
Is it that hard to articulate your own words on a reddit post instead of just writing it with chat gpt? I swear to god it’s impossible to find a legitimately written post these days on here.