What The Fudge for May 24, 2026: The Paradigm Shift in Compute Infrastructure

It's time for What the Fudge. This time it's What the Fudge for Sunday, May 24th. I'm getting to get into the habit, so let's see if I can get it done again in under 5 minutes. First thing this week, if you were listening to your favorite podcast, reading your favorite blogs, you probably noticed some of the stories around quantum computing and the funding around that. This is somewhat related to the billions in grants that are up for various companies that are part of this.. It does tie back into the 2022 CHIPS and Science Act. It does appear that the largest of the large are probably going to benefit. I think if you're familiar with IBM and the notion of quantum supremacy and something around the order of most qubits, it seems like they're going to probably pull down quite a bit. I think that some of the hand-wringing and, uh, concern around suddenly AI eating all of the SaaS companies also probably, uh, was deflated a little bit as most of the shocking Shocking, shocking, SaaS companies are gonna be shipping AI features. I think the other thing that was probably big this week was just around stories related to what will happen as the IPOs for things like SpaceX and stories around what would happen when you combine space, data centers in space, potentially processors in space, but also still terrestrial AI infrastructure being built out and trading hands. I think the next biggest one was probably If you were any— anywhere close to Google I/O, there was all the things related to Google Gemini, new models, new ways of using AI within the AI world of Google. And then finally, the last couple things that probably stood out were, if you are not familiar with Blackstone, uh, if you're not familiar with Google TPUs, uh, these are the two stories I think that kind of fused a little bit of what was part of Google I/O, but also more importantly thinking of this in terms of data and infrastructure space. So if you're familiar with Blackstone picking up things like QTS and some of the other data center companies that were out there, but now imagine that with an equity investment that says billions and billions of dollars, which again is a big number, but in the terms of AI infrastructure, it is a relatively currently big number. It may not be currently big number, say, 3 years from now, but it is interesting that Blackstone and Google are doing a compute joint venture, it seems. And then the last couple of things I wanted to call out were just overall this concept around infrastructure investment for AI. This notion of the US government being involved is interesting, but I also think that some of the challenges we're going to run into aren't related to infrastructure. It'll be related to how we think of when and where particular AI stories start to play out in other industries. I also remember asking the question of when and how does the power shift occur between these companies. So right now it does appear that the largest of the large are the ones that are in charge, but ultimately I think that you'll have stories around smaller, more nimbler startups that do begin to use AI come of age. I also think that the tokenomics, if you will, which is this study of what is the true cost of the inputs, the output tokens related to these AI outcomes, what that's going to mean. And so in the modern world of FinOps and in the modern world of unit costing, that Gartner definition, we're going to see more and more emphasis on where the dollars come in, what value is derived from that platform if it is an AI feature, and then ultimately what will come from that. So that's a bit of a roundhouse of the week. I was also traveling. I will be traveling again this week, so Even putting all these things together, I think it's fascinating that you can talk to your Siri phone and say, Siri, ask ChatGPT to answer this question. That's a pretty cool override. I'm not sure if you've used that before, but I'm also enjoying using Gemini, and I'm also enjoying using Anthropic Claude Cowork, as well as some of the other tools that allow you to have a conversation using just your speakerphone and just basically handing off this concept of transcription and asking questions and then getting a response. And it is interesting how even things like when you have a cough, the perception of the LLM is that you are trying to say something. So I'm looking forward to AI, if you will, accounting for that I might have a tickle in my throat or a cough. But I'm almost up to the 5-minute mark, but I wanted to thank you again for listening. For those of you sticking around. And that's it again. That's again another episode of What the Fudge for Sunday, May 24th. Take care.