Transcription
Even the high-flying Russell 2000 was down, but only 1 and 1/2%. It's still up 17% so far this year.
And in fact, even though all these markets are down, they're still relatively close to their all-time record highs. Despite the fact that you have this looming, uh, decline, significant decline, in the stocks that have been leading the way. And there's been a lot of bad news, too, that came out this week. I'm going to get into it.
Even though the overall stock market was down on the week, I think investors are much too complacent regarding the risks that are increasingly rearing their heads, not just for the markets, but the overall US economy.
Let's start by talking about what happened with these AI hyperscalers on the week. These are the companies that are spending enormous, uh, CapEx for the AI build-out, all the data centers, and the, the chips, and everything that they're buying in order to be able to deliver, uh, AI to customers in the future. The biggest news, I think, came out of Alphabet. That's Google. They announced on Thursday their earnings, and the important part of the announcement was the increase in in CapEx spending, I guess, above what investors had had been expecting. Oh, I think investors were expecting a CapEx to go up, and and and that's what happened, but the reaction was not favorable.
Once upon a time, every time one of these companies announced they would be spending more money on AI CapEx, the stocks went up, but they haven't been going up, and neither have their bonds. Uh, because investors are now finally starting to question whether or not these investments are in fact going to pay off. And so, uh, Google shares, Alphabet, stock was down 10% on the week. Almost all of that decline happening, I think, on Thursday, the day that the the announcement came out. Now, uh, Alphabet is still clinging to a 1% gain so far in 2026, but that's not much to show for all the money that they're now spending and committing to spend on on CapEx.
Look at Oracle. Oracle was down almost 8% on the week. They're another big spender, 7.9%, I think, to be exact. But, that stock is now down 41% on the year. Markets are punishing, not rewarding, Oracle for all of this spending. Similar situation with Meta, that's Facebook, down 7.3% on the week, and another stock that's down on the year, down 8 and 1/2% year-to-date. Amazon also got beat up on the week, down 6.8%, but year-to-date still up a bit, but only 2 and 1/2%. Microsoft lost less than the other, uh, hyperscalers on the week, but still down 2.7%, but year-to-date, Microsoft is down 19.3%. So, almost in a bear market. Uh, Oracle already well into a bear mar- bear market because of all this spending. And remember, this spending is what is driving a good chunk of US GDP, and there's probably a lot of ancillary GDP that is not directly, uh, this spending, but results from it. You know, I think they spent collectively all these su- hyperscalers close to a half a trillion dollars last year.
Alphabet wasn't alone. Oracle, Meta, Amazon, and Microsoft, the companies pouring the most money into AI data centers and chips, all posted losses in a single week. With Oracle now deep in bear market territory, and Microsoft trailing not far behind. Collectively, hyperscalers spent close to half a trillion dollars on AI infrastructure last year, with plans to push that closer to 3/4 of a trillion this year. Markets that once cheered every spending announcement are now punishing it. A signal that patience with unproven AI returns is running out faster than executives anticipated.
On AI-related CapEx alone, not not other CapEx, but just what is specifically related to artificial intelligence, the the data center build-outs and all that. And I think this year they're supposed to spend maybe 3/4 of a trillion dollars. Um, you know, look at what happened to SpaceX this week. And I've been talking about SpaceX even before the SpaceX IPO, but this is another AI-related company, mainly because of xAI. But SpaceX down another 7.7% this week, on top of the prior losses. SpaceX barely held a 115 handle. It closed at 115.07. That's 15% below the $135 IPO price. But even more dramatic is the decline from the high. Because that stock, I think maybe 3 days after the IPO, traded above 225. It's now 49% below that high. And by the way, not just SpaceX got clobbered on the week, but Tesla had multiple bad days. And I think again, this was related to the spending that they're doing on AI. Tesla shares down 18% one of the biggest losers on the week. Now 35% below its 52-week and I guess all-time high. You know, if you thought you had a bad week last week, Elon Musk lost almost $100 billion on these two stocks alone. And you know, obviously, you know, Elon's not hurting. You know, he's still got plenty of money left over, but there a lot of people that aren't as well-positioned as Elon Musk who might have bought into these stocks closer to the highs based on all the hype. And now they have some big losses and those losses are going to get bigger because both these stocks are still outrageously expensive and have a long way to fall, especially as I said with SpaceX when all these shares that haven't even been able to sell yet, although some of them may have sold short but once they're freely trading, the float on this stock is going from 5% where it's still is up to 40% before the end of this year, which is not that far off.
And of course, all of this spending by these hyperscalers, this is what's generating all the earnings for a lot of the tech stocks that still have big gains. Uh, like uh, um Nvidia, which actually was flat on the week and its annual gains are being reduced. It's only up about 9 and 1/2%. Micron, right? Was up 3% on the week, so still up on the news but not up a lot compared to prior gains on similar announcements and in fact, year-to-date Micron is still up about 200%. Uh, and look at Taiwan Semiconductor, which makes all this stuff, was down on the week. Down 1.2% despite the increase. Still up on the year, but it's only up 26%. So, it's not fantastic. In fact, even SK Hynix, uh, which has been a, you know, trading in Korea for a long time, chipmaker over there, but a recent addition to US listings, was down 4.6% on the week.
Two Musk-linked stocks bore an outside share of the damage. SpaceX shares slid further, now sitting well below their IPO price and roughly half of their post-listing peak. A steep fall made sharper by the fact that most shares still aren't even freely tradable yet. Tesla had one of its worst weeks in recent memory, dropping over a third from its all-time high. Between the two stocks, Musk's paper losses ran into the tens of billions. A reminder that even the market's most dominant personalities aren't immune when AI spending sentiment sours broadly.
Because the problem is if the customers of the chip companies are in trouble because they're spending too much money on AI CAPEX, and the market is starting to question whether or not the spending is going to end up being recoverable, whether these investments will be profitable. That's what I've been saying all along. I don't doubt the potential of AI. In fact, I think the markets are likely underestimating the impact long-term that AI is likely to have, but I think they're overestimating the returns on the investments that are being made by the hyperscalers. I think this is going to be similar, uh, to, uh, the internet. I mean, I think that AI will be a bigger advancement, ultimately, uh, than than the internet, but the similarity is that so many companies made all these investments in internet related, you know, CapEx that turned out to be malinvestment. Most of the companies, the early favorites that investors were pouring money into, went bankrupt. They never made any money off of all of their spending related to, uh, building out the internet. An early beneficiary of that was Cisco Systems and and and the stock collapsed because their earnings collapsed because their customers went broke. And in fact, a lot of their customers went broke owing them money because there was a lot of vendor financing. And you're getting similar type of relationships going on now between the hyperscalers and the companies that are producing the products that they are buying. And I think what you're going to also have potentially this week is another scare, a kind of a deep seek type scare related to what's going on with Moonshot and and and Kimmy K3, which is another a model that potentially will compete with Anthropic and Open AI. But also bringing down the costs and further potentially commoditizing AI. Because if companies can buy lower cost AI agents or whatever they're getting from Chinese companies, they're not going to be using more expensive models offered in the US, which means to be competitive, a lot of these companies that have been making all these investments are going to have to lower the prices that they charge their customers and they won't be able to recover the cost of these investments in any reasonable return or maybe there'll be a negative return. They'll be losing money. Now, there are a lot of people that think, oh, there's there's nothing to worry about because the US companies are way ahead. We got like a 6-month lead on China.
A historical parallel worth remembering. The internet boom. When companies borrowed and spent heavily on infrastructure that later proved to be malinvestment. Vendor financing left equipment makers holding debt from customers who later went bankrupt. Cisco's collapse was a direct casualty. Today's AI buildout shows early signs of a similar dynamic between hyperscalers and their chip and hardware suppliers. The concern isn't whether AI matters long-term. Most analysts think its impact will exceed the internet's. But whether today's spending levels can ever be recovered through future profits.
And that may be significant in some respects, but not in all respects. Yes, there there is going to be some applications where a 6-month lead is is significant, where the company that is 6 months behind their AI won't do the job that you need. And so you have to pay up to use the more advanced model. But for a lot of applications, that won't be the case. I mean, if you're a businessman and you need AI to do a specific task for you, and there are Chinese companies where the AI is good enough to do that task as well as a US company might do that particular task, that there's not a meaningful improvement in the results, but it's way cheaper, then that's the one you're going to choose. You're going to You're going to buy whatever AI you need to get the job done. You don't have to overpay. You don't have to have a model that is more than capable. It just has to be capable enough. And I think that could be a scare again that goes through the market that causes some significant declines in in these stock prices that has broader implications for the overall market. Because the market has been held up by these AI-related stocks, these tech stocks. Because there's a lot of negatives that have been ignored as investors have been crowding in to these, uh, smaller, uh, universe of stocks that are still going up.
Yes, the the overall market was down last week. But not a lot. The Dow was down .7%. That's not that much. S&P down 1.3%. The Nasdaq, despite all these big blow-ups, was only down 3.1% on on the week. Year-to-date, Nasdaq's gain is now just 7 and 1/2%, which is about the same as the Dow. It was a big lead for a while. Uh, but it's still up. Even the high-flying Russell 2000 was down, but only 1 and 1/2%. It's still up 17% so far this year. And in fact, even though all these markets are down, they're still relatively close to their all-time record highs. Despite the fact that you have this looming, uh, decline, significant decline, in the stocks that have been leading the way. And there's been a lot of bad news, too, that came out this week. I'm going to get into some of that bad news after I do this break. But before I do that, I want to turn my attention to markets that were positive on the week. And that is gold. Gold and silver, precious metals, had a good week. Gold was up about 1%. Nothing spectacular. But gold rose 1% even though bond yields rose significantly. And even though oil prices rose significantly. Now, you might say, "Well, yeah, well, rising oil prices, you know, should be good for gold because gold and oil be correlated, but they haven't been correlated recently. Ever since the war broke out, the Iran war, there's been a a divergence. It's been opposite. Gold and oil have gone in the opposite directions. And I've been saying that eventually that's going to change and oil and gold are going to move in the same direction and that might be happening. Now, we didn't have a big move in in, uh, in gold this week. We did have a big move in oil. Uh, but
A quieter risk is brewing. Chinese AI models undercutting US pricing power. Cheaper competing models threaten to commoditize AI services, forcing American companies to slash prices just as their infrastructure bills peak. Squeezing margins from both directions. The common assumption that a 6-month lead protects US firms only holds for a narrow set of advanced use cases. For most everyday business applications, a cheaper model that's simply good enough wins the sale. That dynamic could trigger the next major sentiment shock across AI-exposed stocks.
Gold didn't go down. And the fact that it didn't go down, in fact, managed to rise, I think shows that that is happening and that is very significant, uh, for precious metals investors. And look what happened. Silver was up 2.4% on the week, still down 20% on the year. Gold is barely down, um, on the year. No, uh, no, excuse me, it's down about 7% on the year. Um, but considering the huge gain that both metals had in 2025, it's not that big a decline, but it's enough to create a buying opportunity. And in fact, a lot of people got scared that the war was going to be, uh, bearish for gold, that rising interest rates were going to be bearish for gold, and they are wrong. They're actually bullish for gold. And this week may be an indication that investors are figuring that out. In fact, if you look at what happened to the mining stocks, GDX was up 5.6% on the week and GDXJ up 5.8%. These are decent moves where gold was only up 1%. A lot of leverage, uh, this week in the miners relative to the metal. I think that's a good sign of a bottom that we're moving a lot higher. These stocks are still great buys. They're down on the year. The GDX is down 12 and 1/4% so far this year and the GDXJ is down 14%.
And another week where Bitcoin didn't really go anywhere. I mean, I know the week is not quite over yet because we still have the whole weekend, but I think as of when the markets closed on, uh, on Friday, Bitcoin was down close to 1% on on the week. Make it It ended last week, uh, close to 60, 65,000 or just below and this week it barely managed to gain or close above 64,000. In fact, as I'm recording this on Saturday afternoon, Bitcoin is about 64,300 and change. So, it's gained a little bit, uh, since the close in New York, but still an uneventful week for Bitcoin. Didn't move up with the precious metals, went down with the risk assets. Year-to-date, Bitcoin was down about 29% as of, uh, the close on Friday and about the same today still. That's a big drop, 29%. MicroStrategy even worse, down to 4 and 1/2%. Excuse me, strategy. Again, they should have They should have just called it micro based on where the stock is going rather than strategy because they have none. But, down 4 and 1/2% on the week, now down 42% year-to-date. Stretch managed a slight gain on the week of 0.4, but still down 13% on the year. The price is 86.88, meaning the current yield, because the coupon is 12%, the current yield is just under 14%, 13.8%. Investors don't believe in Bitcoin anymore, at least institutions. Apparently, according to Saylor, institution picked you know, stepped up and bought some stretch over the last week or two as retail sold. Obviously, retail sold for a loss, but this is still overwhelmingly a retail product, but I don't read anything bullish into this, the fact that institutions, you know, put their toe in the water. Maybe that's a short-term trade, or maybe they're making some bearish bets where this is a hedge. Maybe they're shorting strategy and buying, um, stretch, or shorting Bitcoin and buying stretch to get that yield and have a hedge. These are bearish bets if in fact they're being made, and I think Wall Street's, uh, rhetoric doesn't match their action. They're still trying to get customers into Bitcoin, but they have no interest in it themselves.
While tech sold off, precious metals quietly gained. Gold rose about 1% even as bond yields and oil prices climbed, breaking a pattern where gold and oil had moved in opposite directions since the outbreak of the Iran war. Mining stocks outperformed the metal itself, often a signal of a developing bottom. Bitcoin, by contrast, stayed flat and disconnected from the gold rally entirely, continuing a rough year. The divergence suggests investors are starting to treat gold, not crypto, as the preferred hedge against economic uncertainty.