Episode Transcript
00:00 Intro
01:28 A sharp AI sell-off and market rotation
03:30 A broader lens on defensive sectors
05:43 Evolving from the AI land grab to a focus on cash flows
06:54 Has the market really been irrational overall?
08:23 Risks that investors have turned a blind eye to
10:27 Leveraged hedge funds are struggling
12:29 Understanding why AI valuations are so volatile
14:45 Is it a rotation, or is everything broken?
16:55 Outro
The Finance Ghost: Welcome to episode 299 of Magic Markets. Some pretty serious market movements in the past week on the global stage. That blue wave washed over the markets and made large parts of it red, at least for a couple of days. Who knows where it goes from here? Investors are weighing up what the impact could be on their beloved AI stocks.
Moe, you're going to walk us through some of the big rotations that we've seen since the election results came out. I'll be talking about some of the characteristics of the AI sector that have gotten us into this sudden risk-off situation.
And Moe, I have to mention it. Good old Jensen Huang of Nvidia fame sadly appears to have made some rotations of entirely the wrong kind [laughing]. But even before that little risk emerged from way out of left field, investors were already hitting the sell button on AI stocks. They just weren't necessarily hitting it a little bit harder on Nvidia.
So perhaps just give us the lay of the land here.
Mohammed Nalla: [laughing] Ghost, there's so many funny things we could say right now. No comments on Jensen Huang's rotations. I knew you'd bring it up, though. And I mean, with all the honeypots out there, maybe we should just be long honey futures. That's probably the most PC thing I can say out there.
Jokes aside, let's jump into the market moves last week. Let's start with the scale of it because I think that context is very important. Thursday, specifically, was absolutely brutal.
We had the Nasdaq down almost 5.5%. We had the S&P down 4.5%, the Dow slightly less, but still 3.5% lower. And then if you just look at the semiconductor index, that was down 12%. It's the second-worst decline on record. Within that, Nvidia down 9% (remember, this was pre-Jensen Huang stories). AMD, down 17%. So, some pretty large moves.
And for the Nasdaq, it was the worst day since the tariff tantrum in April last year. It also left the index firmly in correction territory.
But I said the headlines only tell you a part of the story. We do need to look beyond just Thursday's move.
And if you look at the full week, for example, the S&P was down 2.7%, the Nasdaq was down 3.4%. Yes, lots of tech names there. And then the Dow, that was only down less than 1%. So, this is showing you that it's actually much bigger than just the headline moves. The overall market, certainly not the kind of panic stations that you saw on Thursday.
Technology was down more than 8%, but consumer discretionary shares gained. They actually gained around 2%. Staples were positive. Financials, those gained just over 1%. Real estate was also up around 2%. So, rather than this uniform risk-off move on equities, what we actually saw was capital moving elsewhere.
Now, consumer discretionary, that is a cap-weighted sector. So, we have to be careful about reading too much into an index that is dominated by just a few large names as well. But the breadth overall supports this rotation theory that I've got out there. And roughly two thirds of the consumer stocks actually rose during the week last week.
Now, we also have to zoom out because I think it's more than just a week. We've got to look beyond that.
And that's where a sector like healthcare certainly makes an interesting longer-term point. It's because it was basically flat last week, so not a lot of action, but it has gained around 10% over both the past month and the past three months.
And so, this shows you that it's not just one bad day for technology stocks, but that we've actually seen a bit of a change in the leadership in the market that may have been developing for a while.
In fact, healthcare is a sector I've been watching. It's a sector that offers defensive earnings. It also has improving earnings momentum. It's got this underlying narrative that goes beyond what we see on the index and whether another trillion dollars gets spent on data centres.
And then the consumer businesses point that I raised earlier. The appeal there, they're offering current cash flows rather than these profits that are sitting way into the future, if they're even there, like we see in big tech.
So, in a nutshell, that's the rotation thesis. And the market isn't necessarily rejecting growth, but it is becoming a lot more demanding about what price it's willing to pay for that growth, hence the focus on cash flows that we start to see coming through in the markets.
Now as we speak, we did see a rebound on Friday just following that massive sell-off on Thursday. And there are early signs of support for those deeply sold tech names at the time of recording this early on in the week.
And so maybe (I mean, we'd exclude some of the noise around Jensen and Nvidia), the market is showing that it's willing to own AI, but it is going to be more selective – what parts of the theme do we own? Do we own memory? Where do we play in this whole thing?
We've got CoreWeave reporting this week. They're a neocloud operator. That's going to be quite interesting.
And then lastly, just to separate the timelines, the Jensen Huang story actually arrived after that big sell-off on Thursday, and so it's not really related to that. This morning we are seeing Nvidia down quite sharply. That's because the board has actually opened an independent investigation following those reports that he's been targeted by a foreign intelligence effort.
They say they have no reason to believe that national security or Nvidia secrets (very important there) were compromised, and they've expressed confidence in him. But it's early days there.
There's some key person risk, and it's why Nvidia… I mean, they're this giant. They have a big bearing on the headline index. It will be important to see what happens there, beyond just the initial sell-off, as well as some of this rotation story that we've unpacked.
Now, Ghost, I've gone into a bit of detail there, but without getting us lost in the numbers, can you explain as simply as possible how we got to this point that we're at right now, where the market maybe felt the need to panic?
The Finance Ghost: Yeah, for sure. So, there's a lot to unpack here. But I think it's important to comment on how we've actually evolved from what I would call a ‘land grab’ phase in AI to a stage where the cash flows really do matter.
Now, it's anecdotal. I can't prove it, and frankly my Copilot gets very touchy when I point this out. But I've been saying publicly for weeks now that free AI has gotten a lot worse recently. It really, really has.
And I'm convinced that the free models are now being throttled so that resources can be unlocked for the paid tiers. Now, that's not the kind of strategy you'll see in the widespread adoption phase of any kind of technology.
So, if you see AI as just another freemium model (which is essentially what it is), then the free tier just got worse, and the providers of these models are hoping that this encourages more paid seats.
Suddenly, revenue matters, not just the number of people using the thing. So, we are evolving from an environment of hopes and prayers for investors into one where perhaps the cash flows really do matter.
But that got me thinking. Is this really so new? Have the public markets been so reckless, or were there already some sensible moves on display?
If you think about it, cash was already flowing up the chain to the likes of Nvidia, AMD, SK Hynix, and even adjacent names like SanDisk. Investors wanted the shovels in this gold rush. Yes, some of these valuations got very silly, and it does feel like everyone suddenly forgot that memory has always been a cyclical industry (pun fully intended there). But it's not a completely bonkers approach to follow the cash up the chain and invest there.
And, as free cash flow deteriorated at the hyperscalers, we then saw the market start to move out of those names in favour of some of the other places in the value chain. Another data point that I think indicates some degree of sanity is Apple, with the market celebrating it for not participating to any great extent in AI.
Apple's whole claim to fame recently was, “We're not doing this. We're just staying out of this. We're focusing on devices and services.” And that protected its free cash flow along the way.
So, on the whole, perhaps not completely nuts in terms of the market trying to follow the cash flow. But Moe, I do think that there are some areas where the market did actually turn a blind eye to some of the risks.
Mohammed Nalla: Ghost, I think those are important data points. Because, yes, Apple, that's a nice one there. They did indicate on the latest earnings call that they had some pricing pressures from memory, so pay attention to that.
Last week, Apple stood out for me. Microsoft stood out as well. And so again, maybe not a focus just on the cash flows, but maybe on the players that have some balance sheet flexibility. But in aggregate, I’ll bite. What are the examples that you're thinking of here, in general?
The Finance Ghost: So, one of them would be cannibalisation of existing business models. I'm thinking about Alphabet specifically here. We covered it in last week's show in the context of Reddit, so go back and listen to that if you haven't gotten there yet.
I guess I could just never really understand how people were not considering that Alphabet winning in AI would mean that the traditional search model is then in trouble. It feels like you can kind of have one or the other, but it doesn't feel like they can win at both.
It's pretty brave to assume that the company that dominated the previous era of the Internet will just dominate again, by default. It's just not that simple.
And the other thing that the market seemed to ignore more broadly now is the concentration risks – not just per company, but for AI as a whole. When you've got huge demand for infrastructure and data centres out there creating a bottleneck, the situation changes very quickly (or at least it can change very quickly, if the demand goes away or it diminishes significantly).
The problem is that if you have one big theme sitting behind all this demand, then a change in sentiment can have an impact that ripples throughout the sector. And that is ultimately what has happened here, which is why the memory stocks have been whacked so hard.
Nvidia is now hurting for reasons beyond just AI. All jokes aside, it is obviously potentially going to be a really big issue.
But there are other risks playing out right now in other markets. Like the private markets, for example. Things like data centre commitments and the amount of cash sitting behind that, that may now just not flow as expected.
What about commodities, as well? I'm thinking about copper and tin specifically here. What could happen there? I don't know at this point, but a lot of the narrative there has been around AI.
And of course, all of this is the backdrop for Anthropic's IPO [laughing]. So, good luck to them.
Mohammed Nalla: Yeah, Ghost. I mean those AI contagion risks are quite material. It's something we kind of unpacked even when we covered stocks that fall outside of the sector, like Caterpillar and Cummins. Those are effectively construction stocks, but now they've got AI exposure through power generation. It's quite widespread.
And you mentioned the point around private markets, but we don't even have to look at just the private markets. If you look at what's happened in listed markets, it's where our previous show on Situational Awareness becomes relevant again.
Let's just look at what happened with hedge funds last week. Tech-focused long/short hedge funds lost more than 4% on Thursday, and their returns for the year are now roughly flat – that's massively underperforming the market – having been up as much as 24% at the peak in June.
Now, specifically last week, a company called Walleye Capital lost around 5.9% in October. They are now returning around $2 billion to clients. They're cutting back the internal equity business, and they are reducing their investment teams.
Now, this isn't another situational awareness-style collapse, but the plumbing is quite similar in that we've got crowded exposures, leverage, and everyone needing to reduce risk at the same time.
And the market impact is important because, when you have these things playing out, forced selling isn't really a view on fair value. It's the fact that you need liquidity. Funds sell what they can sell. And when several large funds all own the same stocks, those same positions get hit repeatedly. And, arguably, we may have seen that in the tech sector.
Now, in aggregate, we've said this on the previous show when we looked at situational awareness. Leverage doesn't necessarily make the original investment thesis wrong, but by magnifying those moves, it turns what should be a manageable correction into a liquidity event. And I think that helps explain why the scale of the move can look so disconnected from the amount of genuinely new information entering the market.
There was nothing specific that broke the markets last week. Yes, there were the midterms. Maybe the Democrats are going to get tougher on data centres. But that in and of itself doesn't explain the magnitude of the move. I think that's where leverage comes into play.
So, with that context, you mentioned earlier the changing economics in AI. I just got shown something on ChatGPT this morning saying, "Would you like to advertise on ChatGPT?" So, that's telling you the pressure is starting to come through.
With that context, sentiment changes can have a big impact. Maybe the business model is changing, but what is the technical valuation or fundamental reason for why we've seen this outsized impact on AI names that goes beyond maybe the leverage point I just made?
The Finance Ghost: Yeah. So, it really comes down to valuation methodologies. And I'm definitely not going to spend a ton of time on this. I'll just give you the high level for those who are interested.
So, when a company is trading at a high multiple, relative to its trailing earnings or sales, for example, it's because of big expectations for the future. So that's why people will sometimes say, well, the forward multiple is really low to try and justify the valuation. But the reality is the "forward anything" is a bunch of promises and hopes, whereas the trailing numbers are actual audited, banked numbers.
Now, at an extreme, you can easily end up with a scenario actually - in this market where hype seems to be a feature, not a bug - where all the value sits far in the future, particularly in the terminal value of a discounted cash flow analysis, if you've ever built one of those.
And if you haven't built one of those, that's basically the far right-hand column where you say, “Okay, cool. I'm done with actually forecasting this year by year. I'm just going to assume that after, say, five years or even 10 years (if you really want to be very spicy on your forecasts), there's just going to be a terminal value sitting there.”
The problem is, if most of the value of the thing sits in that terminal value, it makes it incredibly sensitive to the discount rate, which is ultimately a measure of risk. So, if risk goes up, if you're worried about making that money, then the required rate of return for taking the risk also goes up.
That means the discount rate goes up, which means the present value of those future cash flows comes down. And the further in the future those cash flows can be found, the bigger the impact of changing the discount rate.
If you've built one of these things, you know you can just make it dance and give you whatever answer you want, based on changing a few inputs. It's like an inside joke in investment banking, obviously.
And so, what we're watching right now is, I think, just the risk premium going up on AI. People are worried about these cash flows actually materialising. Down comes the valuation because down comes the terminal value. And that's what's happening out there right now.
Mohammed Nalla: That really does kind of sum it up. At the end of the day, we do a lot of this detailed work in Magic Markets Premium. If you're not a subscriber, go and check that out. It's a deep dive into a global stock every single week for only R99 a month. We think that is phenomenal value.
I just want to bring this all together because we discussed some of the macro, some of the micro, and my own base case still remains one of rotation rather than “everything's broken”. If the optimist in me is right, I think the healthier version of the story is that market leadership will broaden, and we'll see some of that sectoral rotation come through.
This is where investors will reward current earnings, cash flows, balance sheet strength, maybe some of what we saw last week, reasonable valuations. But it does mean that AI investment will continue, but at a more disciplined pace.
Now, this is obviously the ideal scenario. A more dangerous version is where data centre delays become outright cancellations. You might see hyperscaler capex starting to fall. Maybe the guidance falls and credit spreads start widening. We're seeing some of this already.
And then effectively, seeing those crowded AI positions being faced with forced selling of leveraged players – those are some of the indicators I would watch, not just whether the Nasdaq has another bad day.
So, I would say it isn't necessarily the death of the AI thesis, but you have to stay close to the door here. Maybe own some higher-quality names. Because the market is applying a higher cost of capital to that thesis, as you've mentioned.
I would say it's correct, right? You've got to apply a higher cost of capital, higher risk premium, and it's asking the tougher questions which have always, in my view, been central. They needed to be there. Where are the cash flows? You can't just rely on that terminal value forever.
I think the headline last week was “a market massacre”. I think that was maybe a little bit overblown. The context is important.
Just for a quick one to wrap, from the all-time highs, tech has been hit the hardest. The Nasdaq, down around 11% off its all-time high. That was in October. The S&P, 6% lower, also an October high. The Dow, just a hair above 6% below its all-time high, and that was from slightly earlier in August.
So, if you contextualise that, yes, that's a bit of a correction. It's not panic stations for me right now. I am holding my overall equity exposure but will continue to focus on that rotation theme and on repricing, rather than thinking everything is completely broken at this stage.
I don't think we've seen signs of that. But like I say, you've got to stay cautious. You've got to stay close to the door.
Unfortunately, that's all we have time for this week. So let us know what you thought of the show. Hit us up on social media.
It’s @MagicMarketsPod, @FinanceGhost and @MohammedNalla, all on X, or go find us on LinkedIn. Pop us a note on there.
Until next week – same time, same place. Thanks, and cheers.
The Finance Ghost: Ciao.
