Original text: David Lin, X
Compiled & organized by: White55, Mars Finance
History does not repeat itself simply, but it is always remarkably similar. In this episode of renowned YouTube host David Lin, we will witness Gareth Soloway, a seasoned trader with 27 years of experience, who will interpret the similarities between the current market and historical peaks, having witnessed the internet bubble and the 2008 financial crisis. He believes that driven by the AI concept, market valuations have become extremely overextended, and a critical turning point has arrived, with a correction of 10% to 15% being just the beginning.
Gareth not only issued a warning about the stock market but also holds a cautious attitude towards the short-term trends of Bitcoin and gold. He systematically elaborated on why he believes investors need to prepare for the impending market turmoil from multiple dimensions, including macroeconomics, industry insider information, and technical charts.
The relationship between the AI bubble, the labor market, and monetary policy
Host: There's no doubt that the current market is in a very interesting period. I noticed an article where former senior economic advisor to Trump, Kevin Hassett, made a point that he believes artificial intelligence may be causing a “quiet period” in the labor market. He mentioned that despite strong GDP growth in the second quarter of 2025, companies might reduce their demand for hiring college graduates because AI has increased the productivity of existing employees. I've seen some reports that large consulting firms like McKinsey have also lost some business as clients turn to cheaper, more streamlined AI consulting companies. Do you think AI is the main reason for the stock market rise, but at the same time, it is causing a slowdown in the labor market?
Gareth: Firstly, AI has undoubtedly played a role in driving up the stock market. Data shows that 75% of the S&P 500's increase over the past two years is directly related to AI concept stocks, which have been leading the market higher. However, I do not agree that the current partial weakness in the labor market is due to the prosperity of AI. AI will eventually have more and more impact on employment, but the real reason at the moment is the huge uncertainty in the business sector.
Of course, the stock market seems to be rising, but the reality is that people are still suffering from the pain caused by inflation, and I believe that the actual inflation is higher than reported, which has led consumers to start slowing down their spending (for example, the performance of Cava and Chipotle). The reduction in consumer spending has forced these companies to pause hiring.
Host: The Federal Reserve is about to stop quantitative tightening (QT) and may inject more liquidity through continued rate cuts. Do you think this will further boost the entire stock market? In other words, will the AI bubble get even bigger?
Gareth: I believe the AI bubble is at a turning point. I think the stock market has peaked and will face a correction of at least 10% to 15%. In fact, I mentioned before that a 10% correction could start in October, and the market did peak at the end of October. Although the current decline from historical highs is not significant, I do believe that a downward trend has indeed begun. I want to discuss some of the reasons for this, as they are very noteworthy.
First, there are the valuation indicators. The current valuation incorporates projected revenues over the next five years (until 2030) and even further into the future into the stock price, which means that the price has pre-empted earnings that have yet to be realized, and the risks are very apparent.
Secondly, we must talk about how funds flow back and forth. For example, AMD received billions of dollars in investment from OpenAI, but in return, AMD gave OpenAI warrants to purchase 100 million shares of AMD stock. Nvidia also provided some funds to OpenAI, allowing OpenAI to purchase or borrow chips from Nvidia. This is somewhat of a typical “Ponzi scheme,” aimed at maintaining this positive momentum, while the reality is that this ecosystem is not as stable as it seems. Many companies also acknowledge that “AI is great, but it’s currently difficult to monetize.”
I truly believe that AI is the future, but the question is, at this current point, is it really worth such a high valuation?
Another important issue is the data centers. The construction of data centers is currently on hold. Remember, the surge in AI stocks is largely due to the need for all these data centers to be built, and they all require chips. But the key point is that Microsoft has paused the construction of two data centers, and Micron has also paused one. Why? Because there is not enough power. They need to obtain power, and they can't simply draw from the existing grid because doing so could triple the electricity bills for residents. So, if you channel all the energy to these data centers, it would actually overwhelm the average consumer.
In the end, those hyperscale data center companies are using a seven-year depreciation period to calculate the value of their chips. This is absurd. Data shows that due to the rapid advancement of technology and the heavy load of chips running at high intensity for two consecutive years, a chip purchased at full price is worth only 10% of its original value after two years. When you spread the depreciation over seven years, the annual depreciation amount becomes very small, making their financial reports appear more profitable. But in reality, these hyperscale companies are seriously overestimating their profits.
Host: The questions you've raised have indeed persisted for some time, but no one knows how long this situation will last, nor when it will stop. So as a trader, how do you make practical investment decisions based on this information? After all, most people may agree with your view, but they would also say, “We don't know when the music will stop.”
Gareth: This is precisely the terrifying part. If you ask the average person, most of them would say we are in a bubble, yet they continue to buy in because they don't want to miss out on the opportunity for prices to rise and believe they can successfully exit before the market turns. This mentality was especially evident in the cryptocurrency market and other assets at their peak in 2021.
Looking at the weekly chart of SMH (VanEck Semiconductor ETF), it essentially covers all the major semiconductor companies such as Broadcom, NVIDIA, and AMD. The yellow line represents the 200-week moving average. Looking back to 2020 and 2021, we see a pattern of deviation from the 200-week moving average: in the past, at peak points, it deviated from the 200-week moving average by as much as 102%, followed by a significant pullback of 45%.
In the 2024 data, the deviation also reached 102%, followed by a 40% correction in the industry. Recently, SMH once again reached a deviation of 102% a few weeks ago, indicating that the market may face new adjustment pressure. The 200-week moving average acts like the market's “base camp”; when stock prices deviate too far, they will eventually return. According to current data, the semiconductor industry may face a significant correction.
Market Correction and Bitcoin Analysis
Host: This is indeed concerning, as you mentioned earlier, the semiconductor industry is closely tied to other tech stocks and the broader economy, exhibiting a phenomenon of cyclical financing. I suspect that if any major semiconductor stock declines, it could drag down the entire market. Do you agree?
Gareth: I completely agree. When 75% of the S&P 500's gains over the past two years have come from AI stocks, if these stocks decline, they will inevitably drag down the entire market. Moreover, you have to consider other factors. Currently, 90% of GDP growth expectations are driven by the capital expenditures of these large tech companies. Now imagine, if these big companies slightly cut back on capital spending, the U.S. economy could slip into recession. Therefore, we are on the brink of a potential pullback that could be larger than expected. This precarious state of the stock market, especially in the tech sector, has already begun in the past few weeks. We've seen several very bad sell-offs, and just last Friday, we experienced the largest single-day sell-off since April.
Host: The 10% to 15% decline expectation you mentioned earlier has also been echoed by leaders of some major banks. Goldman Sachs CEO David Solomon stated a few weeks ago in Hong Kong that the stock market is likely to experience a correction of 10% to 20% within the next 12 to 24 months. The CEO of Morgan Stanley agreed and suggested that we should welcome this possibility, as a 10% to 15% correction is a natural part of the cycle and not driven by some macro cliff effect. They seem to believe that such magnitude of correction is common even in a bull market. So, structurally speaking, are you still bullish?
Gareth: As a short-term trader, I have been more bearish recently, and I have been shorting some stocks like Nvidia and SanDisk.
Let me show you this chart of the S&P 500 index, and I'll demonstrate why we are likely at a top. A clear trend line has formed from the low point of the COVID-19 pandemic in 2020 to the low point of the bear market in 2022, and the S&P 500 index has currently reached the upper trend line parallel to the bull market peak in 2021. Historical data shows that when the market touches the upper line of this channel, bear market-level corrections have occurred.
Based on this, I believe the pullback has already begun, and the top of the S&P 500 has formed. The reason the market is currently shaky is that those “buying the dip” investors have been brainwashed by large institutions, the government, and others to believe that the market will never drop more than 2% to 3%. When a real pullback of 10% to 15% arrives, they will be very surprised.
Host: Which tech stock do you think is the most overvalued? You mentioned that you are shorting NVIDIA.
Gareth: I prefer to short the semiconductor industry broadly. I think there is some risk in shorting with NVIDIA's earnings report coming up, as it could potentially rise by $10 or $20 after the report and then drop again. Long-term, there is a significant pullback risk based on valuation and technical analysis.
I am also shorting and observing individual stocks that have moved very vertically this year, including SanDisk — these weekly charts suggest a pullback of 20%–30% is reasonable. I am not denying the fundamentals of these companies, but the technicals, valuations, and market structure present very high short-term risks.
Additionally, it is important to remember that funds do not completely exit the market at the very beginning of the peak. In the early formation of the peak, such as in 2007, we would see a pullback followed by a strong rebound, then another pullback, and then another rebound, because buyers have been trained to buy on dips. The largest declines occur at the end of the cycle when everyone gives up and falls into panic. Therefore, the early formation of the peak is usually slow because there are still people continuously buying, but as prices decline, the speed of the drop will accelerate.
Host: Alright, let's turn to Bitcoin. We'll come back later to ask Gareth for his year-end prediction for the S&P 500. Bitcoin has now seen a significant pullback, having dropped below $100,000 and is currently under $95,000. What do you think about Bitcoin right now and the key support levels coming up?
Gareth: From a professional standpoint, the recent high was actually easy to determine. We've discussed over the past few months that if you connect the 2017 bull market high and the first high of 2021, this trend line perfectly predicted every recent peak.
So it is clear that the white line here is a resistance level. If we can rise again and break through this line, then the views of those shouting absurdly high prices may become valid.
If the stock market declines and triggers panic, people unfortunately will sell Bitcoin. Currently, Bitcoin's key support is around $73,000 to $75,000 (many tops/breakout points are supported in this range). If the bears win the battle, Bitcoin may return to $73,000 to $75,000, or even lower; if the bulls can hold this line, we can return to $127,000, $128,000, or even $130,000.
Host: You mentioned the risks in the semiconductor sector, and we know that Bitcoin and tech stocks are interconnected. Why do you think Bitcoin's performance this year has been worse than many semiconductor stocks and the entire tech sector? Bitcoin has been essentially flat this year, while the Nasdaq index continues to rise.
Gareth: There are several reasons:
Bitcoin has recently become a “boring” asset. I know this sounds crazy, but when you see some chip stocks rising by 30%, 40%, or even 100%, they look like the new “altcoins” with ridiculous gains. In contrast, Bitcoin seems to be less “sexy.”
Another factor is that we are beginning to see institutional buying power is not as aggressive as before. Some crypto companies that used to build up Bitcoin reserves are now facing financing difficulties, leading to a decrease in buying pressure. We have even seen this with MicroStrategy. Due to changes in lending conditions, MicroStrategy is unable to borrow the same level of funds for large-scale purchases as it did in the past. So MicroStrategy is still buying, but the orders are much smaller than before.
The last factor is de-risking. If you look back, before the stock market peaks, risk assets usually peak first. Bitcoin peaked in December 2017, and the stock market peaked in January 2018. Bitcoin peaked in November 2021, and the stock market peaked at the end of December that same year. When people start to de-risk, at least for large institutional funds, they first pay attention to the highest risk assets, which are cryptocurrencies. Therefore, the process of de-risking starts from there and then spreads to the stock market like a cold. I believe we are on the verge of seeing this happen.
Host: If Bitcoin and tech stocks have had a very close correlation in history, and Bitcoin has outperformed the stock market in past bull market cycles, but this time it has not. Does this mean that Bitcoin is undervalued?
Gareth: I still believe that Bitcoin will ultimately outperform the stock market, as it remains a reserve digital gold asset. So when panic strikes and risk-off begins, prices will be impacted. But once the dust settles, people will realize that the stock market needs to drop more, while Bitcoin can become a recipient of some of that capital flow. I want to be clear that I still think Bitcoin could drop to $73,000-75,000, or even lower, but I will gradually buy in during the decline to accumulate a long-term holding position.
Host: Are you more optimistic or pessimistic about the altcoin market?
Gareth: I have a cautious attitude towards altcoins. They are always changing, and there is always a new popular technology emerging. In my opinion, Ethereum needs to drop a bit more; I have set the swing trading buy price for ETH between $2800 and $2700, which is an important support level.
Gold, Risk Comparison and Long-term Outlook
Host: Let's talk about gold. Gold is currently still firmly above $4000 and is undergoing consolidation. Interestingly, I've noticed that videos about gold on my channel are starting to lose traction. A month and a half ago, when the price of gold surged above $4000, people were very excited. Now it has basically formed a bottom around $4000, and my interpretation is that people have accepted this as the 'new normal.' Is this the new normal? Is $4000 the bottom price now?
Gareth: Personally, I believe there is still a bit of downward space for gold, as the “weak hands” have not been washed out yet. The market tends to wash out these hesitant holders before starting the next bull market. Comparing the gold price movements in 1979 with those in 2025 reveals a nearly similar pattern: first a wave of increase, then a sideways consolidation, followed by several consecutive weeks of increase (in 1979, there were 9 consecutive weekly bullish candles, and 2025 also saw 9 weeks). Historically, the consolidation in 1979 retraced to a key support level before embarking on a new round of increases.
Based on analogy, I believe gold might drop to $3,600–$3,500 before starting the next major rise. But importantly, this time is different from 1979. In 1979 and the 1980s, the then Fed Chairman Volcker was raising interest rates. Now, Powell is cutting rates. In 1979, the debt-to-GDP ratio was 32%, and now it is 130%. The government is recklessly spending money now. So the difference is that in 1979, it took us 20 to 30 years to see historical highs again. This time, I believe we will return to historical highs by next year. Reaching $5,000 next year is beyond doubt for me.
Host: In the short term, which asset has a greater downside risk: gold, Bitcoin, or stocks? If calculated by percentage, which one should decline the most?
Gareth: In terms of percentage, Bitcoin has the highest volatility and the greatest short-term downside risk. If Bitcoin drops to my target prices of $75,000 or $73,000, that would represent a decline of about 23% from current levels. If gold drops to around my set level of $3,600, that would be about a 12% drop. For the stock market, we discussed a pullback of 10% to 15%. This pullback would take us back to around 6,100 points, which is the previous breakout pivot high that has now become technical support. The stock market is the most uncertain for me; it may be at a cyclical high, and we could see declines of up to 30% to 40% in the coming years, although I believe there will be a rebound after a pullback of 10% to 15%. If I were to allocate, I would lean more towards gold at these mentioned target prices because it is relatively the least risky, followed by Bitcoin.
Host: You are more optimistic about gold structurally compared to Bitcoin, why?
Gareth: The main issue with Bitcoin is that there is a lot of leverage in the system. People can invest with large amounts of money. When entities like MicroStrategy hold so much Bitcoin and also use leverage, it concerns me. As someone who makes a living analyzing trading risks, it does make me a bit nervous. If they get into trouble and are forced to liquidate, they could cause a crash in Bitcoin that exceeds anything we've seen before.
Gold is even more diversified, as central banks around the world hold it. They will not sell it off in a panic, as central banks can print their own money. So ultimately, at least for me, gold prices have more safety.
Host: I would like to share a post by Ray Dalio that echoes what you mentioned earlier about “stimulating in a bubble.” He pointed out that the Federal Reserve announced the halt of quantitative tightening (QT) and the start of quantitative easing (QE), which is, in any case, a loosening measure. How do you think this will affect the market?
Gareth: I agree with Dalio's point of view: The Federal Reserve's technical shift in operations (stopping balance sheet reduction and implementing quantitative easing or similar actions) is injecting liquidity into a system that already has bubbles. Historically, the pattern of increasing debt during expansion and deleveraging during recessions has not occurred in this cycle - we have been continuously accumulating debt, which has created larger bubbles, and larger bubbles mean larger crashes. The current situation may be even worse than during the financial crisis, and I think many people find it difficult to fully understand its scale. The U.S. is approaching or entering a problem on a century-scale level, which will 'educate' many young investors who have never experienced a major crash. People often say that someone in every generation must go through one. Most of those who experienced the Great Depression are no longer around, and we seem to have forgotten those lessons - be prudent with finances, avoid excessive spending, and do not accumulate massive debt.
Host: What warning advice would you give to young traders who have not yet experienced a significant downturn or bubble burst?
Gareth: Stay alert. Many new investors entered the market after the COVID-19 pandemic, and since then, we have only experienced a V-shaped recovery, with the market hitting new highs within a month. They might think that the market will only go up. I have been trading since 1999, and I remember that it took the Nasdaq over 15 years to reach a new high. But history isn’t like that; the market may take longer to recover. Protecting capital is key—trading with discipline, controlling risk, and being aware of the systemic risks posed by the current monetary and debt situation. At least 60%–70% of American households may already be in a recession, and the market's rise masks this reality; once the stock market falls and high-end consumption slows down, in my view, regardless of how much capital expenditure AI has, the economy will decline.
