Where’s the Moat?

In Where’s the Moat?, Andrew Wilkinson and Steve Sosnick break down the seismic shift in the AI landscape, exploring how DeepSeek’s disruptive model challenges NVIDIA’s dominance and reshapes investor expectations. Is the AI gold rush facing a reckoning, or just evolving into a new era?

Summary – IBKR Podcasts Ep. 222

The following is a summary of a live audio recording and may contain errors in spelling or grammar. Although IBKR has edited for clarity no material changes have been made.

Andrew Wilkinson 

Welcome to today’s podcast, everybody. This is Andrew Wilkinson in the studio with Steve Sosnick, Chief Market Strategist from Interactive Brokers. Welcome, Steve. 

Steve Sosnick 

Good to see you, Andrew. 

Andrew Wilkinson 

Likewise. Now, we thought we’d hop on a quick call to discuss the entire DeepSeek rout that cost, I guess, NVIDIA about half a trillion dollars. And then, with all the other associated hardware, software providers, and utility companies, it had a massive impact of probably around a trillion dollars in market cap. Pretty brutal, Steve. What do you think are the lessons from this episode? 

Steve Sosnick 

The lessons I take away here are that even disruptors get disrupted. And I think part of the reason we’ve come to really make AI a theme—and I say “we” as in the investment world right now—is that it is a potentially disruptive technology that will change the way we do things, similar to how the internet came into play, call it 25 years ago. It brings promises of productivity and changes to the way things are done. 

But it doesn’t mean that just because some of these companies have been staggeringly profitable, that will continue indefinitely. It’s also important to remember that this whole AI theme developed about two years ago—when ChatGPT made its debut around October or November 2022. 

What happened was that, as this AI gold rush occurred, we sort of assumed that the ChatGPT type of model was the way it needed to be done—massive investments in hardware, data centers, and the electricity required to power all that. That was also the route taken by other competitors, whether it was Alphabet, Meta—Meta just last week announced a $65 billion capital investment—or even the government, which basically committed to $500 billion, although Elon Musk would argue that number is a lot smaller. But the bottom line was, that was the paradigm. 

What DeepSeek threatens to do here is— and I say “threatens” because it’s impossible to verify whether all their claims are true about coming to life in two months and costing only $6 million—there’s not a lot of transparency about that, so it’s tough to take them 100 percent at their word. 

But if AI models can be developed without that sort of massive investment in all the things I just discussed, that changes the game. It doesn’t change the game for AI—I’d actually argue it may accelerate the adoption of AI and make it more egalitarian because it will be easier for people to use without having to rely on the services of those who spent billions of dollars building their AI infrastructure. 

If you’re NVIDIA, using the gold rush analogy, NVIDIA is the pick-and-shovel manufacturer for this gold rush. And we’ve just learned that there’s a different way to get a different outcome—one that doesn’t require constantly buying billions of dollars worth of NVIDIA chips. 

That’s why the stock got hit. 

Andrew Wilkinson 

And I think NVIDIA, after the close on Monday—a couple of days ago from when we’re recording this—came out with a press release after the rout and said, basically, that the demand for AI is there, it’s only going to build, and it will require more chips from them. 

But Steve, what do you think this means for investors who are bullish on NVIDIA and all the other players in the same game, now that DeepSeek’s AI has been built at a fraction of the cost of ChatGPT? 

Steve Sosnick 

Investors need to understand that maybe a lot of the expectations that have been priced into some of these AI companies need to be reevaluated. I divide the world into AI providers and the pick-and-shovel companies—the NVIDIAs, the Broadcoms, but also companies like Digital Realty, Constellation Energy, and Vistra, as well as utilities and housing companies that have benefited from AI-driven demand. 

It changes the game if we don’t need all that infrastructure. And it’s a little early to say, but this industry, which had been spewing profits and growth projections into the foreseeable future, may now have to temper those expectations. 

And the problem is, going into Monday, NVIDIA was the largest company in the S&P 500, and Broadcom was the sixth largest. If you change the paradigm for growth for those kinds of companies in a highly concentrated market—which is what the S&P 500 has become, and the NASDAQ 100 even more so—then it has no choice but to spill over into the larger market. 

That’s really the key consequence. It’s very important to keep in mind. I use the analogy of the internet 25 years ago. When the dot-com bubble burst, neither Google nor Facebook existed. 

At that point, we were trading options electronically. Our programming and data were housed on Sun Microsystems workstations. When we needed a PC, we used Compaq. Our search was primarily Yahoo. 

See where I’m going here? 

You can’t assume that current winners will remain winners. 

For example, today is Wednesday, and we’ve got earnings coming out tonight. People talk about Meta, Tesla, and Microsoft—all of which are very important earnings reports. IBM is reporting too. And IBM was the stock at one point. 

And you know where I’m going with this, Andrew, because I’m going to let you use your new catchphrase—some of these companies that seem unassailable can be disrupted. 

Andrew Wilkinson 

I remember back in 2001, Steve—it’s etched in my memory. There was a visual on the front page of the Financial Times showing the market cap of Yahoo at its peak against one of the biggest oil companies, and Yahoo eclipsed that market cap. And, very, very quickly, that all went away. 

Steve Sosnick 

Okay, since you’re not taking the bait—Andrew’s catchphrase, which he messaged me through Teams, is basically, “Where’s the moat?” And I used it in the piece I wrote Monday. 

By the way, I have a feeling that should be the title of today’s podcast. I thought it was a perfect three-word synopsis of what had occurred. 

A lot of the largest companies had built these big competitive moats—or anti-competitive moats, depending on your point of view. And the idea of DeepSeek may have shown that the moat is shallower than people thought. 

Andrew Wilkinson 

Right, right. We typically talk about moats in the context of value companies that have built up brand equity over years—investors like Warren Buffett look for that. I think people thought that was happening with NVIDIA, that nobody could come near them or catch up quickly. 

And then all of a sudden, bam! DeepSeek becomes the number-one download on the app store over the weekend. It’s front and center and a real challenge to that moat. That’s the concept here. 

Steve, we’ve seen tech stocks traditionally exhibit more volatility than blue chips—probably for that reason. Was this a classic example of why NASDAQ volatility is greater than S&P 500 volatility? 

Steve Sosnick 

Typically, it’s not exactly apples to apples, but the two things I always watch—and they’re right up there on my market monitor—are VIX, the CBOE Volatility Index, and VXN, which is the same calculation but for the NASDAQ 100. 

It’s not a given that VXN always trades above VIX, but it typically does. I don’t want to go down the rabbit hole of correlation and dispersion, but in general, NASDAQ is a more highly concentrated index. 

These stocks are in many ways more volatile. The market loves them because they show strong growth, and for the most part, they are relatively highly valued. But if you grow earnings (“E”) fast enough, you can get away with a high price-to-earnings ratio (“P”), as long as the “E” keeps growing. 

With NVIDIA specifically, the idea was that their technology was unassailable. DeepSeek didn’t invalidate their technology, but it may have reduced the need for quite so much of it. 

Andrew Wilkinson 

Steve, you put out a really nice piece on Monday during the rout. You called it an inflection point rather than a turning point. Could you explain that? 

Steve Sosnick 

Sure. The trend hasn’t necessarily turned, but the pace of that trend has likely changed. And as a result, some of the winners and losers may have changed. That’s the story we’ll need to resolve. 

Andrew Wilkinson 

Steve Sosnick, Chief Market Strategist at Interactive Brokers—thank you for joining me today. And to our audience, remember to subscribe wherever you download your podcasts. 

Steve Sosnick: Thanks, Andrew. Take care. 

Andrew Wilkinson

Bye for now. 

Go Source