So, what am I going to talk about today? Three things: what's really going on in the AI market, gold's performance and whether we missed it, and what's happening in credit markets. Plus some extra goodies at the end.
Welcome to thoughtful money. I'm its founder and your host Adam Tagert. Today we have the return appearance of one of the most respected capital allocators, Jan Van Eck. Jan is CEO of VanEck, with over a hundred billion dollars in assets under management. Jan, thanks so much for joining us today.
It's good to see you again, Adam.
It's a privilege to have you. The audience really looks forward to these quarterly outlooks. Before we get to your chart deck, if you had to pick a word or phrase to describe the macro situation right now, what comes to mind?
Happy? As in healthy and happy or manic and happy?
Happy with eyes wide open. There are so many things changing so fast.
Your slides will answer that better. Let's jump into your material.
Thanks for having me. A quick reminder of our perspective: we look at the same big drivers but with a historical perspective that change can be dramatic and rapid. We focus on multi-year trends. There are four big crazy things: the worst peacetime fiscal deficit, AI as the second biggest technology after railroads, India becoming a major consumer economy, and a dramatic deregulatory act in payments. In a nutshell: fiscal deficit likely improving, AI supporting mega-cap profitability, gold trend early, and we'll talk about credit. Agenda: what's going on in AI, gold, and credit. Does that sound like an okay agenda?
It does. Those are all questions the thoughtful money audience is asking right now. Very timely.
Great. For AI, I want to talk about what's driving the huge demand for compute and the deal frenzy. Token demand is exploding 38 times a year – that's the amount of processing power needed. Costs are coming down exponentially: hardware costs dropping 90% per year from Nvidia, and software costs (LLMs) falling fast. This is why we're seeing a frenzy to lock down compute. Nvidia's demand is visible into 2027, and their valuation is high but justified by growth. We're facing a compute shortage. OpenAI is becoming a MAG8 company – it has the fastest user adoption and is competing with Google and Amazon. They're referring traffic to e-commerce and have multiple chat types: social, research, shopping, gaming. OpenAI's revenue is 20 billion this year, growing to 40-50 next year and 100 billion after. They're borrowing heavily but the prize is huge. The MAG7 earnings growth has been a straight line up due to AI.
Hey Jan, just so people understand, tokens – are those the inputs used to train AI or the queries?
Good point. Tokens are units of compute. A simple query might require many tokens to compute the answer. There's also reasoning capability emerging, and software itself is generating new demand for AI processing.
We've trained on human content and now AI is generating its own data. How long until it gets so smart it kills all of us? We'll save that. But you said token demand exploding and costs dropping – why isn't this a bubble?
I focus on token demand. If that keeps rising exponentially, I'm not worried. The compute shortage supports Nvidia's earnings. Their P/E is high but normal for a company growing 40-50% a year. The multiple might come down later, but near-term demand is massive. Also, nuclear stocks are up 90%, Oklo up 7x with no revenue – take profits but stay in for the multi-year trend. Electricity costs are rising, which is regressive and a political issue.
Will efficiency gains moderate energy demand?
No, in the short term we have a compute shortage. We need all energy sources. Consumer electricity bills are rising, and that contributes to pocketbook pain despite low inflation.
After your last appearance I bought the VanEck uranium ETF and it's up 76%. Victory lap.
That's why I love multi-year trends. You knew the puck was going there, but not how much would be priced in this year.
Now, two AI winners: OpenAI I think is a new MAG8 company. They have the fastest user adoption and are winning against Google in monthly active users. They are competing with Amazon for shopping via Shopify. Switching costs are emerging as they learn user preferences. The MAG7 earnings chart shows a straight line up – that's what's driving the market. I think Wall Street is underestimating the compute shortage and how earnings can grow over the next two years. We're living in Nvidia's hardware world, and OpenAI is building compute infrastructure. The revenue question narrows to OpenAI, which is printing 20 billion this year, 40-50 next, maybe 100 billion after. Everyone else has revenue to pay for compute.
Why isn't it a winner-take-all ecosystem?
There's scale advantage, but looking at the earnings growth, all MAG7 can win against the rest of the economy. The compute shortage and global demand for AI are so large that the prize justifies the investment. OpenAI is the one firm that doesn't have huge revenue yet, but the dream is enormous – that's why they're getting funding.
Well, they're winning. They have the consumer. The question is, can they monetize the consumer? But they do have the consumer. So I'm trying to reframe the whole AI debate. Partly because OpenAI is not a public company, I think we're missing a little bit what's going on.
Yeah. And I was going to ask about that. Do we know if they have plans to go public soon? I know when they initially announced thinking about it, Elon had a lot of issues with that, but...
There's no announcement at all on that front.
So, anyway, I'm sorry, I got to go fast here to finish the AI section. The last winner that emerged to me is OpenAI and can they get their financing because that's the weak link. The second winner is gaming. The biggest take-private in the history of capital markets happened last quarter. I love highlighting these things because things are happening so fast. Investors are saying gaming is a huge AI winner. The video experience will be way better and the costs of programming and QA are dropping dramatically. The public markets are missing that gaming is a big AI winner.
Okay. And that makes sense. We're entering the era of Ready Player One where the gaming experience becomes amazing and immersive, but the cost is driven down by the technology. If you can own one of these companies, you're going to ride that wave.
I think that's what this headline tells us. So we're finishing up the AI story. Who cares? Number one, focus on tokens. Number two, focus on whether OpenAI can raise money for compute. Also watch it compete with Amazon and other Mag7 companies. The stock market is underappreciating the compute shortage from demand for tokens. Mag7 earnings continue to explode. This is a chart of US stock market history. The growth in the Mag7 is not unprecedented.
Sorry, back to that chart. The key takeaway is you would expect the share of the chart made up of the dark and light blue – the transports are darker – but the two blues near the bottom are going to continue to get bigger as a percentage of the pie.
Very well could be. Yeah. Likely. The question is when does the earnings growth of the rest of the market catch up? Some people look at 2026 estimates: Mag7 expected 13% growth, small cap could be 22%. Maybe there are other parts of the market. People have been holding their breath for years, but it's conceivable. Value doesn't look like it will surpass Mag7, but the thing to watch is whether the market smooths out. You saw a lot of that in Q3.
Okay. And I will say that's consistent with recent experts like Mike Cantorowitz and Anna Wong becoming more optimistic that profits will broaden beyond big tech.
