In consumer AI, the battle is for the one agent to rule them all. Consumers will only pick one main agent as their primary. It will need to work with other agents but those will not be their main agent/interface. Place bets accordingly on who the masses chooses.

My biggest concern remains being able to get power, then build data centers (regulatory approval). Two very hard hurdles to clear. We anticipate a potential gap of global accelerator compute demand to energized GW of 20-30%. BUT even if you get the GW you still need approval to build...

All the research we do, all field notes, event coverage, exec interviews, data models, surveys, and so much more goes into CS Atlas. We do sooooo much research and analysis it just needed a home. @DiligenceStack subs even get entry tier access :)

Had so many really interesting conversations/meetings this week across compute, interconnect, power/land/shell. The challenge is how to have formats to share those deeper observations and insights that I have not seen publicly analyzed yet. Some in podcasts but others have to get out of my brain/notes and into written form lol.

My conversation with Penguin Solutions ( $PENG) and Astera Labs ( $ALAB) from Ai Infra Summit: Both working on CXL related solutions Astera is a $50B market cap company, and Penguin is a $2B market cap company, so there’s a huge size difference between the two I’m lumping them together because they do similar things at least in one of the lines of their business (Astera’s “Leo” CXL product, and Penguin’s compute solutions division) Penguin uses Astera Labs “Leo” controllers for their CXL offering (SMART) (although I guess the next generation will use Qualcomm if I remember correctly), but they layer on top more of a finished product experience primarily for enterprise users (not CSPs) who need some extra support beyond what Astera will provide. So their product is more suitable for Fintech/financial services, health care, and other industries where enterprise on-prem is in-demand These boards have DIM slots for LPDDR5 memory, which can be used to augment memory from GPUs (meaning a server tray can have less compute vis expensive GPUs by pooling in this board’s memory via LPDDR5) instead of having to either A. Have more GPUs ($$$) or B. Going all the way up to memory on the CPU which causes a big slowdown in memory access speeds. The next generation Astera “Leo 2” and Penguin’s next generation equivalent will both also be compatible with LPDDR4 Cool product. Can see the need, makes sense. I don’t know how defensible Penguin’s business here is though, I guess it’s niche enough that Astera is focused elsewhere, but I imagine that if on-prem got lucrative enough maybe Astera would come after it (many such cases across the showroom floor)

It will be consensus, if it isn't already NPO is going to ramp HARD first. This was from our report on the copper to fiber inflection in March. I know the sub is expensive but you would have known early if you read it :)

Creative Strategies has worked inside the technology industry since 1969. Our research follows how companies compete and which product or technology choices create durable differentiation. We connect those decisions to changes in market structure and the larger paradigm shifts that determine where value moves next. CS Atlas gives clients a new way to work with that full body of research. Ask a company or market question and see how our view has developed as the technology and evidence changed. Every answer links back to the supporting work, making it easier to test an assumption against our broader research and bring the relevant evidence into a decision.