This is the takeaway from our conversation with Big Technology expert Alex Kantrowitz, one of our favorite frequent guests who sat down with us this week to talk all things hyperscaler. He talks to the CEOs and CTOs of the most important companies in hardware, software and AI. Alex knows more than most. And amid all the uncertainty on Wall Street about all of the money being spent to scale the data centers, he’s fairly certain that this compute will not go unused. More importantly, there is a belief out there that we are nowhere near built-out to answer the cascading amount of demand coming when the agentic future takes hold. If we’re going to have a billion people ordering around their own AI agents to carry out tasks and handle transactions, the current state of compute simply cannot support it. Which means compute prices are going higher and the infrastructure build will have to continue - or, quite possibly, even accelerate.

Most people who are reading all the articles and absorbing all the Wall Street commentary about capex bubbles are thinking in terms of the technology we have today. Coding assistants and chatbots and user-generated artwork. These are the use cases they see in front of them. They’re obvious and familiar and tangible. But they’re not thinking about what it’s going to take to facilitate the existence of a hundred million autonomous cars and trucks, humanoid robots, automated manufacturing operations and other such things that are a little further out on the horizon. However, the people making these investments with trillions of dollars and multi-decade time horizons are thinking precisely this way. More. The bottom line is we will definitely need more.

This is the disconnect behind the volatility in tech this summer. The bond market is trying to rationalize all of the debt issuance being used to augment the free cash flow spend that’s currently taking place. Yes, they can see the revenue growth numbers for AWS, Azure and Google Cloud (37%, 43% and 82% during this latest quarter, respectively). That’s helpful. Yes, they understand that Anthropic is now approaching a $90 billion ARR number for the full year. Yes, they can see that OpenAI has likely reached a billion regular users.

But they can’t make the leap - mentally - into a world where we are looking back at Claude Sonnet and ChatGPT 5 a few years from now and laughing at how primitive they were - how utterly manual and annoying they were to use. They are not picturing a world where we are talking to our agents all day rather than typing to them, and they in turn are doing useful things for us while we sleep. A world in which robots and computers are talking to each other continuously and working together to construct a new world where nobody does dangerous or filthy jobs ever again. All of the constraints of the cost and safety of human labor come off of humanity for the first time in the history of the world and anything is possible.

Elon Musk thinks we’re going to 10x global GDP in the coming years and end up with enough wealth that all of the things we worry about today simply will not matter. You don’t have to buy into that. But just imagine if, directionally, his instinct isn’t totally wrong. How much compute will something like that require?

Yes. More. Lots more.

So we don’t know for sure how far out the agentic stuff is, but it’s getting close every day. We don’t know the exact definition of “super intelligence” - Mark Zuckerberg’s favorite term - but we know that the smartest and most highly capitalized people on earth are currently working on it around the clock. We don’t have an exact date on the calendar upon which all of these predictions will arrive, but what we do know is that it’s inevitable. They’re spending trillions of dollars to build it and their investors and customers are demanding it. It’s going to happen, even if there are a whole bunch of crashes and corrections and panics along the way.

Not every project will bear fruit. Not every investment will be a good one. Not every expenditure will prove justified in time. The conversation we should be having, however, is about the entirety of the forest, not each individual tree. Sitting with Alex this week gave us a helpful reminder about the larger story versus the quarterly earnings per share figures.

You can listen to the whole show below or watch it on YouTube to get caught up. I hope you enjoy checking it out as much as we enjoy recording these episodes each week.

THE COMPOUND & FRIENDS

Josh and Michael are joined by Alex Kantrowitz of Big Technology to discuss: OpenAI and Anthropic's explosive growth, the threat to traditional software companies, soaring cloud demand, agentic AI, and the diverging strategies of Microsoft, Meta, Amazon, Apple, Google, and other tech giants.