@intjudo Thanks for the feedback!
I actually had Credo next up to look into after that Marvell earnings that came out to understand the competitive landscape better. Credo is reporting earnings this week on September 4. There’s a few reasons I wouldn’t invest in Credo or strongly prefer investing in Astera,
- On their last earnings (May 29), they had guided for 59-62M revenue and came in at 60.8M
- Next quarter they guided for only 58-61M or 2% sequential decline - Astera is projecting 24-30% sequential growth
- They have a legacy business and their AI portion of the business is still ramping
- Credo was founded in 2008 before AI/cloud - Astera was founded in 2017 with cloud/AI in mind and already has more revenue
- Credo has 25M in inventory, while they are fabless, I believe they take delivery themselves whereas Astera integrates their product in without taking the inventory delivery
- Credo was asked about Astera and Marvell on the call and they said they are confident in their solution - this is in contrast to Astera saying they put competitors out of business with their software and competitors are using their software to debug their own issues
- Credo has lower gross margins than Astera
- Credo has ~400M in cash, Astera has ~800M
- Credo has 500 employees, Astera has less than 300
I was looking into the Blackwell GB200 specs trying to wrap my head around how Astera plays into the Nvidia solutions. Interesting that the GB200 does have include a “PCIe Gen 6 Bandwidth Bidirectional - 2x 256 GB/s Gen6”
I think this is what Astera is taking about with Nvidia transitioning to Gen6 PCIe. This may be the way that the Aries product interfaces with Blackwell.
I’m still trying to track down how Nvidia themselves is a customer of Astera. There seems to be little information about what products they are buying, but the hardware manufacturers like to maintain confidentiality so they don’t reveal their product specs to competitors.