How AI Bailed Out the Stock Market VIEW IN BROWSER By Michael Salvatore, Editor, TradeSmith Daily In This Digest: The charts are full of lies… Look at any chart of the S&P 500 since 2020, and you’ll see a stock market that has almost doubled in five years. In a normal world, this would be great news for investors in U.S. stocks. It would be an extraordinary return. The long-run annual return of the S&P 500 was previously about 9%. To go up 97.1% in five and a half years would mean an average annual return of around 17%. It would be more striking when you consider the fact that we had a bear market during one of those years that saw stocks fall 25% over the better part of a year… a black swan-style crash during the pandemic that saw a drop of more than 30%… and the “liberating” shenanigans of April 2025 when stocks fell about 20%. When you consider that, you can argue that the last five years of stock market returns really took place over about four years. Viewed this way, it would be even more impressive. Would be. The problem is we’re not in a normal world… We’re in a world with exploding liquidity, rampant inflation, and the biggest decline in U.S. currency strength in many years. And we can only celebrate the market doubling in five years if we’re looking at nominal returns… While ignoring the massive loss in purchasing power of the dollar. All we need to do is divide the S&P 500 by the M2 money stock – a useful analog for the amount of dollar liquidity, which exploded in 2020 – to see an alarming truth… Measured against the amount of liquidity created in the market, stocks have not gone up 97.1% since the start of 2020 (blue line below). In fact, when you adjust for the purchasing power of the dollar being crushed by money printing and inflation, the return is more like 37.5% (red line below): 
That takes the 2020s’ once-beautiful annual return of around 17% and chops it to less than half: 6.8%. And do you know what the long-run annual REAL return of the S&P 500 is? About 7.5%. Yes – when taking inflation into account, the 2020s have actually been a laggard. A disappointment. That’s a hard pill to swallow after back-to-back years of what looked like stunning gains in stocks. But the math doesn’t lie. The 2020s have had worse real returns than the long-run average. | Recommended Link | | | | Volatility. Uncertainty. Complexity. Ambiguity – or what The Army War College calls “VUCA.” VUCA is reordering the stock market before our eyes – sending reliable giants crashing 64%, while obscure names skyrocket as much as 980%. America’s Top Trader, Eric Fry, built his career by forecasting which stocks rise and fall in turbulent times. Now, he’s sharing 7 FREE trade ideas to help folks move beyond Amazon, Tesla, and Nvidia and into better alternatives – before it’s too late. Access Eric’s brand-new broadcast here. |  | | Looking more closely at the chart, we can see an even uglier situation… See that black horizontal line cutting across? The “real” S&P 500, adjusted for added liquidity, hit that major resistance level three times – in 2020, before the pandemic… in late 2021, before the bear market… and in mid-2023, before the late summer drawdown. What’s shocking is that the real S&P 500 was actually closer to making new highs than the nominal stock market in mid-2023. And that’s precisely because the Fed was embarking on a short-lived campaign of quantitative tightening – removing liquidity from the market. In other words, the stock market was organically trading much better than the nominal performance suggested. The “real stock market” entered a new bull in April 2023, while the “fake stock market” didn’t enter one until June. But the Fed is back to its old tricks. M2 is at its highest level ever. And the gap between real returns and nominal returns is at the highest level of the past five years – a spread of more than 30% – naturally surpassing the gap from the 2021 peak. What does this mean for you? I don’t aim to just sit here and opine about the Fed’s raucous currency intervention of the 2020s and how it’s distorted markets. That’s fun but doesn’t get us anywhere productive on its own. What is worth doing is looking at the types of assets that are saving the stock market… the kind of ideas and technologies that preserve wealth and grow productivity faster than the Fed can diminish it with money printing. Productivity is the beating heart of any economy. It’s how we measure whether products (outputs) are worth more than the sum of their parts (inputs). Currency debasement and inflation are the antithesis to productivity – it fosters a world where the output (value) shrinks relative to inputs (currency) over time. These two forces are pulling at each other in every market, every company, and every asset at any given time. Certain assets are so productive, they beat out the pull of inflation. These are the most important ones to own. Two areas come immediately to mind here in 2025: Bitcoin and artificial intelligence. Bitcoin’s resistance to money-printing and inflation is well documented, here and elsewhere. Being the first and only truly scarce asset, it directly converts energy into economic productivity via mining and has helped preserve wealth better than anything else since it was invented in 2009. Gold bugs may be fuming about my claim Bitcoin is the only scarce asset, but as soon as they can tell me exactly how much gold is and ever will be in existence, I’ll retain that definition. While new gold deposits are getting harder to find, miners are still out there doing it… and could end up mining asteroids as well. Meanwhile, AI is more similar to Bitcoin than you might think. Large-language models (LLM), first debuted in late 2022 and more widely applied toward the end of 2023, convert energy into a tool that (in theory) helps individuals work more productively in a wide array of tasks. We can look again at the chart above and see just how impactful AI has been to the stock market since the end of 2024. Despite the continuation of money printing by the Fed, the real stock market broke higher in 2024 as the application of AI became more widespread. In this sense, AI bailed out the stock market… its investors… and the Fed. It allowed for productivity – or, at least, the perception of productivity – to flourish at a pace faster than the Fed printed money. Now, why do I say the “perception of productivity”? AI could see diminishing returns from here… This will ruffle some feathers, but I think there’s a strong chance we’ve already seen AI’s peak productivity… at least of the LLM variety. There are a few pieces that can help us paint this picture. - For one, a March survey revealed that a shocking 42% of surveyed companies (1,000 respondents across North America and Europe) decided to abandon most of their previously set AI initiatives in 2025. That was up from 17% in 2024.
- Just two months back, a study of 25,000 workers across 7,000 companies by the National Bureau of Economic Research in Denmark found that “AI chatbots have had no significant impact on earnings or recorded hours in any occupation.”
- A report by Vivaldi Group notes that 90% of executives are failing to see real results from AI – only about 10% of companies are unlocking meaningful value.
AI is not useless. Heck, I used AI to discover and source these examples. But we humans still had to make sure the AI wasn’t “inferring” things that weren’t true or stumbling over the nuances of this research. And that’s just one bit of evidence that it isn’t the massive productivity hack it was originally sold as. At least, not for the 90% of companies that are struggling to see real results from it. There’s a good chance that number will diminish with time, too. One problem with an increasing reliance on AI tools is that the primary source of an AI’s knowledge is the entire sum of current human knowledge. An AI summary of human knowledge by definition is reductive, and anyone who’s used AI and has seen it make things up or produce errors knows how reductive it can be. So what happens as more AI content continues to proliferate? The primary source becomes less human knowledge and more flawed AI output… reducing the quality of the output further. It’s the same result as if you’d saved an image as a JPEG, emailed it to yourself, saved it again, and repeated the process about 20 times. Each time you send the image, the compression eats away at the image quality until you’re eventually looking at a pixelated, blurry mess. That could be what happens with AI as more and more people use it to feed the knowledge base. This isn’t me trying to call the top in AI as an investment theme. The continued performance of stocks like Nvidia, Microsoft, and more show us that people are still very much believers in it. And to be clear, these are the ultra-productive stocks that are carrying the market higher and bailing out investors. But we should be wary of these holes in the AI plot, lest we get caught unaware as many investors were back when the dot-com bubble burst. If you’re sitting on big gains in AI stocks, now’s not a bad time to take them… Earlier this month, I wrote you with an ominous message: “You Have 10 Trading Days Left.” Monday was day 10. Back then I was talking about the seasonal peak for the stock market which, based on the 15-year seasonality chart of the S&P 500, ended on July 28. From the span of June 28 to July 28, stocks were up 100% of the last 15 years for average returns of 3.35% (before turning down nearly 2% from now through the start of October). Here’s an update of the chart I shared in that article. You’ll never guess what the S&P 500 did. Through July 28, it returned 3.37%: 
As you see below, that’s almost exactly what TradeSmith Seasonality forecast. 
The 10 days are up, and I hope you took some profits as I recommended. Because now we’re in for a couple months of lag. Setting aside a seasonally strong period from August 23 to September 7, when stocks have returned an average of 1% and been positive for 80% of the last 15 years, stocks wear a heavy weight around their neck over the next two months: 
Taking some profits before that happens is a wise move. But just as wise is getting access to our seasonality software today and finding the stocks with the most consistent seasonal trends – up or down – during this period. One of the best examples of a downtrending stock during this period is Lam Research (LRCX). The stock has fallen an average of -5.89% from July 31 (tomorrow) through August 25. Take a look: 
That slide shows us LRCX’s Q2 earnings reports tend to not go well, and investors carry that selling throughout the month. Access to this software is invaluable during seasonally strong periods. Whether you want to trade options, stocks, or even find opportunistic times to go to cash, our seasonality charts show you the best times to buy and sell. Just as great is the Seasonal Edge strategy, which delivered four winners out of six trades this month, including 128% gains on Blackstone (BX) calls and 129% on Alexandria REIT (ARE) calls. Today is the day to get on the list for Seasonal Edge alerts, not to mention access to seasonality charts on just about any ticker you want to see. We’re holding the door open for one last day so our CEO Keith Kaplan can get you fully briefed on the situation ahead of the seasonal slide. Watch his webinar here; because the information is specific to July 30, it’s coming down after today… including the special offer for viewers to join our Trade Cycles seasonality service. To building wealth beyond measure, 
Michael Salvatore
Editor, TradeSmith Daily P.S. Hey TradeSmith Daily reader… I want you to “sanity check” me. Do you think there’s even the slightest chance that we’ve seen the peak in AI? Short-term, long-term… as a technology, as an investment… I want to know where you stand. Especially if you think I’m crazy. (The last time I thought I might be crazy was when I was aggressively buying Bitcoin in summer 2023. That turned out to be not so crazy.) Another thing. Did you take profits at the seasonal peak as I suggested? If not, are you looking at the Tuesday price action and hoping you did? Write me at feedback@TradeSmithDaily.com. I really love getting emails – yes, even the mean ones – and would love to establish a more regular dialogue with you. (Disclosure: Michael Salvatore held a position in BTC and NVDA at time of writing.) | 
No comments:
Post a Comment