This article is reprinted with permission from Phyrex_Ni, and the copyright belongs to the original author.
In the past week, although there hasn't been as much volatility as in the second week, discussions about bull and bear markets have increased, and the divide between the two camps has grown larger. Last week, I provided insights from a macro and data perspective, indicating that while we are not in a typical bull market, the AI-driven U.S. stock market has not shown signs of decline. Moreover, Bitcoin (BTC) in the cryptocurrency space is highly correlated with tech stocks in the U.S. stock market, and AI is a major driving force within those tech stocks.
Thus, we can derive a basic logic: unless the AI bubble in the U.S. is burst, the likelihood of a significant pullback in the U.S. stock market is low. So, what could burst the AI bubble? I believe the main risks lie in investment returns, the Federal Reserve's monetary policy, and excess energy and computing power.
First, AI companies have been seeking massive funding for data center construction, but revenue is expected to be only about $13 billion by 2025. This cash-burning model, if it cannot quickly convert into sustainable profits, will lead to a collapse in investor confidence. Currently, investment in AI infrastructure has reached a historical peak, but there is a huge revenue gap.
Companies like NVIDIA, Super Micro, and AMD have very high profits, but true AI service providers, such as OpenAI, Anthropic, and Cohere, are still losing money while burning GPUs. Although enterprise clients are deploying AI, most projects do not yield investment returns and are still in pilot phases. If financial reports fail to meet the expectations of the "Fourth Industrial Revolution" for an extended period, the market will reprice.
A larger concern comes from the debt structure. Morgan Stanley and Forbes predict that by 2028, AI data center-related debt could reach $1.5 trillion, with most of it coming from capital expenditures by hyperscale cloud providers like Microsoft and Amazon, often financed through high-yield bonds with rates of 5% to 7%. In the short term, if the Federal Reserve can continue to lower interest rates in 2025, it could provide some buffer.
However, under the influence of new tariffs, inflation has already shown signs of rebounding. If the Federal Reserve is not resolute in lowering interest rates, it could lead to rising financing costs, potentially triggering a chain of defaults that could penetrate the entire credit system. The valuation logic for AI currently rests on the assumptions of perpetual growth and low discount rates. If the Federal Reserve delays rate cuts and Treasury yields remain high, the present value of future cash flows will be severely compressed. AI companies generally consume a lot of cash, and once capital costs rise again, issues such as financing difficulties, order stagnation, and declining profits for hardware manufacturers will all erupt simultaneously.
In addition to investment returns and interest rates, excess energy and computing power also need attention. The U.S. Department of Energy predicts that by 2028, electricity consumption by AI data centers will triple, while the power grids in California and Texas are already experiencing bottlenecks. If approvals for nuclear or renewable energy are delayed, capital expenditures could increase by an additional 20%, directly bursting the myth of perpetual growth. At the same time, delays in the delivery of NVIDIA's Blackwell chips have allowed AMD to gain an edge. If there is indeed an oversupply of computing power, hardware gross margins could drop from 70% to 50%, leading to a rapid collapse of profits in the supply chain.
Although I see these three risk points, I believe it is unlikely they will explode in 2025. This is because, whether in autonomous driving or AI-generated voice and video, there is still significant imaginative space. Although there is still a considerable gap in demand for the Fourth Industrial Revolution, we can already see the potential for AI to replace human labor in certain areas. Especially if NVIDIA's financial reports continue to strengthen, the market's consistency in maintaining the AI bubble remains strong.
Regarding monetary policy, the Federal Reserve is constrained by the labor market and fiscal deficits, making the likelihood of slow interest rate cuts much higher than that of raising rates again. Therefore, from a temporal structure perspective, 2025 seems more like a stabilization period for the AI bubble. Although valuations are high and risks are lurking, the narrative around AI persists. As long as NVIDIA's profit expectations continue to be met and the Federal Reserve does not withdraw liquidity, the structural strength of the AI-driven U.S. stock market will continue. What truly warrants caution is when profit growth stagnates, energy bottlenecks remain unresolved, and inflation rises alongside refinancing pressures; this could signal the turning point for the AI bubble. In this regard, the financial reports in the fourth quarter of 2025 may be particularly important.
Thus, in my view, as long as the U.S. stock market remains highly correlated with cryptocurrencies, particularly BTC, and as long as the AI narrative has not burst the bubble in the U.S. stock market, the U.S. stock market can continue to rise, and BTC's performance will not be too poor. So why do I say that the U.S. stock market and cryptocurrencies are highly correlated?
The fundamental reason for the high correlation between the U.S. stock market and cryptocurrencies is not that they belong to the same type of asset, but that they are both driven by the same liquidity logic. Whether it is tech stocks on the Nasdaq or Bitcoin, they essentially belong to risk assets. When the Federal Reserve releases liquidity, excess funds in the market seek higher returns.
Typically, high-growth tech stocks and highly volatile cryptocurrencies, especially the more consensus-driven BTC, are pushed up first. When U.S. dollar liquidity is abundant, tech stocks and BTC in the U.S. stock market rise almost in sync, while when liquidity tightens, yields rise, and funds flow back to the bond market, they both decline simultaneously. The risk appetite of funds acts like a thermometer, while the Federal Reserve's monetary policy serves as the thermostat.
Another reason this correlation has been reinforced over the past few years is the overlap in funding structures. Since 2021, more and more traditional institutions have entered the cryptocurrency market, including hedge funds, family offices, and even pension funds that are allocating BTC or other related assets. This means that the same pool of funds is operating in both the U.S. stock and cryptocurrency markets, using similar risk control models, drawdown thresholds, and risk-adjusted mechanisms.
When inflation exceeds expectations and U.S. Treasury yields rise, these institutions collectively reduce their exposure to risk assets, leading to sell-offs in the U.S. stock market and simultaneous pullbacks in BTC. Thus, short-term fluctuations in the cryptocurrency market and the U.S. stock market form a kind of "synchronous resonance."
Structurally, the cryptocurrency market has gradually become a derivative reflection of tech stocks, especially an extension of the Nasdaq. Over the past two years, the correlation coefficient between BTC and the Nasdaq has consistently remained between 0.7 and 0.9, particularly during phases when the AI narrative dominates the market, with both fluctuating almost in sync. An increase in the U.S. stock market often indicates a return of risk appetite, while an increase in the cryptocurrency market represents an overflow of that appetite.
Many times, BTC even reacts ahead of the U.S. stock market because it is a 24/7 trading market and serves as a leading indicator of fund sentiment. Additionally, psychological expectations and narrative cycles further enhance the correlation between the two. AI, technology, and cryptocurrency currently form a single speculative logic chain in the market, where technological innovation brings about a productivity revolution, which in turn drives the revaluation of growth stocks, while digital assets become the future reflection of technological dividends.
Ultimately, the correlation between the U.S. stock market and cryptocurrencies is driven by liquidity and is a result of the resonance of risk appetite. When interest rates fall and liquidity is abundant, both rise together; when yields rise and liquidity tightens, both fall together. It can be said that BTC is like a leveraged tech stock.
So, is there still a logic for the U.S. stock market to rise? There are three data points to pay attention to. First, the cash allocation of global fund managers has recently dropped to 3.8%, while the historical low for this data is 3.5%. This indicates that the funds available to global fund managers are limited, which is both good and bad. The good news is that the funds being spent are more directed towards the U.S. stock market, especially tech stocks led by AI. The bad news is that these funds are limited, and there is not enough capacity to continue pushing stock prices higher in the market.
This also suggests that it is very likely these fund managers will sell off a portion of their holdings in the short term. For some investors, this could be a frightening prospect, raising concerns about the onset of a bear market. However, historically, fund managers are not typically long-term holders. Even in 2025, there were two significant reductions in holdings: the first from January to February, where cash allocation dropped from 4.2% to 4%, and the second in April, where it fell from 4% to 3.8%. Currently, it remains at 3.8%, with a peak of 4.1% reached between June and July.
In other words, even if fund managers reduce their holdings, it does not necessarily mean the market will enter a bear market. A short-term pullback is possible, but in the long term, it may not be a sign of a shift in fund managers' positions. The previous increase from 3.8% in April to 4.1% in June could be a similar operational pattern, as the core focus this week remains on the Federal Reserve's monetary policy rather than a misguided approach.
Additionally, since last week, investors have purchased over $3.9 billion in U.S. stocks. The net inflow of funds into individual stocks reached $4.1 billion, marking the fifth highest point since 2008 and the highest record for the S&P 500 during a week when it fell by at least 1%. This indicates that current investors are buying the dip, and this buying is driven by institutions, with institutional inflows reaching $4.4 billion, the highest since November 2022.
Among them, retail investors bought over $1.1 billion, marking the second weekly net purchase in six weeks. Hedge funds sold off $1.6 billion, marking the fifth consecutive week of selling. It is worth noting that the statistics for "hedge funds" likely overlap with the sample of global fund managers, so their selling behavior can represent a certain degree of global fund positioning adjustments. Moreover, this is just regarding the U.S. stock market; there is also a set of data specific to tech stocks.
This week, investors in tech stocks saw a net inflow of $10.4 billion, compared to $9.2 billion the previous week, resulting in nearly $20 billion in net inflows over two consecutive weeks, the largest two-week net inflow ever recorded. This also marks the fourth consecutive week of significant inflows into tech stocks.
Therefore, even from these three data points alone, we can see that the market does carry some risks. The risk lies in the fact that fund managers, who no longer have sufficient funds, may reduce some of their positions, and they may have already begun doing so. However, institutions and retail investors continue to buy into the U.S. stock market, especially tech stocks, during downturns. The reduction in holdings by fund managers is likely to coincide with the buying by these investors. Next week, the earnings reports from the "seven sisters" will be released, and if there are no significant issues with the reports, it should provide a positive stimulus for the market.
Of course, aside from the issues within the U.S. stock market itself, there are also concerns about whether Trump and China's tariff issues will trigger a downturn similar to that in April. Especially early Wednesday morning, Trump mentioned that if no agreement is reached, tariffs on China could rise to 155%. However, from the perspective of the risk market, after the information was released, both the Nasdaq and the S&P experienced declines, but the drop was very limited, and overall stability remained. The reason is simple: the prevailing view in the market is that Trump will definitely TACO. The term TACO stands for "Trump Always Chickens Out."
What Trump is doing is making threats and creating panic on the surface, but ultimately he will back down or resolve things through negotiation. Why? Because Trump cares more about the ups and downs of the stock market compared to other U.S. presidents; he is more concerned about his achievements during his time in office and is more willing to portray himself as a savior. This is the core essence of MAGA (Make America Great Again). It might sound a bit convoluted, but to put it in a story we heard as children, it's like "The Boy Who Cried Wolf."
Initially, when Trump wielded the knife of tariffs, everyone was worried he would be very extreme. However, once the market began to decline in response to his actions, he would find ways to recover, the most common being to delay implementation. April 2025 is an example of this. When the tariff plan was announced in February, the market was shocked, but the actual tariff strategy implemented was not as aggressive. This might also be a common tactic: first scare the market with exaggerated numbers to create panic, and then alleviate anxiety with more moderate measures. After doing this repeatedly, people's attention wanes.
The fluctuations during the U.S.-China trade war in April seem trivial compared to now. At that time, mainstream media discussed TACO daily, but now, although the final outcome has not yet materialized, the next statement is often "meeting with Xi Jinping in a few weeks," which itself increases the probability of TACO.
However, if we set aside the emotional theatrics of Trump and look at it from a macro perspective, once the 155% tariff is actually implemented, the pain points for both China and the U.S. are completely different. The impact on China is more structural. The U.S. remains China's second-largest export market, accounting for about 15%. Industries such as electronics, home appliances, machinery, furniture, and clothing will be directly affected, especially export-oriented private enterprises. In the short term, employment pressure will increase, and the renminbi may also face temporary pressure. This is not what China wants to see.
On the other hand, the pain for the U.S. is political and inflationary. Although it seems that Trump is protecting American manufacturing, in reality, the U.S. is highly dependent on the Chinese supply chain, especially in areas like low- to mid-end consumer goods, electronic components, pharmaceutical raw materials, photovoltaics, and rare earth materials. If tariffs are fully enforced, inflation will quickly rise by 0.7% to 1.2%, and CPI data will reflect this within the next two months. Therefore, fundamentally, the trade war does not align with China's interests or the economic needs of the U.S., and the severity for China is somewhat higher.
Once inflation rises, the Federal Reserve may be forced to slow down its rate cuts, maintaining high interest rates for the dollar and increasing the debt burden. This means stock market pressure and electoral risks for Trump. This is why the market generally believes Trump will not truly escalate the situation. What he wants is a tough image, not a crashing market. I have been saying since February this year that Trump will absolutely not artificially create an economic recession; instead, he will work harder to delay its arrival.
Overall, the probability of the U.S.-China trade dispute turning into TACO is high. Trump fears not the trade deficit, but a stock market crash and declining public opinion. China also does not want to exacerbate burdens in the currently somewhat tense economic environment. For this reason, the market tends to believe that U.S.-China negotiations will either continue to be postponed or find a new balance. This tariff sounds frightening, but it is more likely a short-term deterrent operation. For the market, unless Trump truly follows through, every "tariff threat" followed by a pullback seems more like a passively created buying opportunity.
After all, behind the market turbulence stands not the hawkish Federal Reserve Chair Powell, but Trump himself, who always comes out to soothe the market after every drop in U.S. stocks.
Of course, there is not a complete lack of good news in the market. After months of discussion, there is a real possibility that the war between Russia and Ukraine may come to an end. European countries are collaborating with Ukraine to formulate a "12-point peace plan," proposing an immediate ceasefire, establishing post-war security mechanisms, and terms for reconstruction and sanctions. Some core points include:
Immediate ceasefire, using the current front lines as a basis for negotiations, halting further territorial advances by both sides.
Security guarantees, with Ukraine receiving security commitments and defense support from European and NATO allies.
Post-war reconstruction funding, with Ukraine receiving dedicated funds for rebuilding infrastructure, restoring the economy, and supporting livelihoods.
Return of children and prisoner exchanges, with all abducted children returning to Ukraine and a large-scale prisoner and war prisoner exchange mechanism.
Mechanism for Ukraine's EU membership, providing a fast track or clear path for Ukraine's integration into the EU system.
Use of frozen Russian assets, with some of the frozen assets of the Russian central bank being used for Ukraine, but only returned when Russia complies with the agreement and supports reconstruction.
Governance negotiations for occupied areas, with no legal recognition of Russian sovereignty over occupied areas, but both sides will negotiate the future governance of these regions.
Of course, this "12-point peace plan" has not yet received agreement from Russia, but if it can be smoothly advanced, it would mark a shift from pure military conflict to a political resolution framework and reconstruction phase. Although we may not yet see a final ceasefire date, this is the first time there is a real possibility of stopping the war. If it stops, the U.S. will gradually lift sanctions on Russia, and Ukraine's exports will become smoother. The inflow of goods from Russia and Ukraine will also alleviate inflationary pressures in the U.S., which would help guide the Federal Reserve's rate cuts.
Now, let's take a look at the on-chain data.
From the BTC supply on exchanges, the overall trend during the week is still decreasing, meaning that although the price of BTC is falling, there is no panic; instead, many investors are buying in. Even in the last 24 hours, the new chips transferred in have only increased by about 1,600 coins from the lowest point, which has little impact on the spot price.
Last week, we also saw from the annual data that BTC is currently at its lowest value of the year and also the lowest in the past six years. This indicates that even though some investors believe the bull market has ended, in reality, more BTC is being bought by investors. The more the price drops, the more they buy. Currently, it is becoming increasingly difficult for long-term investors to release their holdings at low prices, while short-term investors are holding less and less BTC.
Although the exchange data is relatively good, the spot ETF data representing traditional investors is very poor. From the data over the last seven consecutive working days, both BTC and ETH have seen net outflows, indicating that traditional investors' attention and investment in cryptocurrencies are very low. Especially on Monday, when BTC and ETH prices rebounded nicely, the ETF did not show the previous trend of chasing highs and cutting losses, remaining in a state of net outflow, further indicating that traditional users' influence on price movements is very limited.
Next, regarding the data of long-term holders who have held their positions for over a year, last week we saw this data trending towards accumulation, corresponding to the decline in BTC prices. This data has been studied for a long time; besides true holders, there is a significant possibility that the holdings on exchanges are decreasing. Currently, it still appears to be in a distribution state. Theoretically, this could correspond to further declines in BTC prices, but whether it will completely enter a distribution phase or reverse midway is still uncertain.
This data cannot be used as a long-term judgment of bull or bear markets. For example, from March to April last year, there was also a situation of long-term holders accumulating, but just two months later, with the end of the U.S.-China trade conflict, it returned to a distribution state. The biggest contention now is not just the U.S.-China trade, but more importantly, the Federal Reserve's monetary policy.
Looking at the data on open contracts, BTC and ETH have been deleveraging over the past week, but due to the significant losses on October 11, the recent week has seen less noticeable deleveraging. BTC has now reached the lowest point of open contracts in the past year, and leverage should be running low. ETH has also reduced leverage from one-third to half. From a historical data perspective, lower leverage often attracts more funds, which is generally helpful for price increases.
Next is the data on position distribution. This data shows that although high-net-worth investors and small-scale investors are showing signs of bottom-fishing as prices decrease, high-net-worth investors are doing so more smoothly and without showing aggressive behavior, while small-scale investors are more reactive to price changes, leaning towards short-term holdings.
Finally, regarding the URPD data, the support level remains very stable, with no signs of collapse. Over this long period, the support between $104,500 and $111,000 remains very solid. As long as there are no systemic issues, the resilience of this support can withstand tests. Although BTC's price trend is not good right now, it is clear that the changes in the chips held at a loss are not very intense, and there are no signs of panic or hasty exits. This indicates that even more loss-making investors have a positive expectation for BTC's long-term trend.
In summary, next week we will enter the October interest rate meeting. Currently, the CME predicts a 98.9% probability that the Federal Reserve will continue to cut rates in October. The likelihood of this expectation being realized is quite high, mainly because the government shutdown has increased the risks in employment data, and the Federal Reserve is unlikely to take risks. Even if the CPI data is not good, it can be interpreted as a one-time impact from tariffs. Then there is the impact of U.S.-China trade; although the trade war does not benefit either side in the big picture, Trump is clearly more aggressive, and we hope it does not provoke excessive retaliation from China.
Although the market has mixed opinions on the shutdown, it has already lasted 21 days, and there is still at most half the time before operations resume, so the actual impact should still be relatively limited. As for BTC, I still believe that BTC and tech stocks will maintain a high correlation. As long as there are no systemic risks and the AI bubble does not burst, I personally still think there is at least a chance for BTC to continue rising.
Related: Bitcoin (BTC) miner debt surges 500%, miners actively prepare for the computing power battle.
Original article: “AI Bubble Intact, U.S. Stocks and Bitcoin Resonating: Liquidity-Driven Bull Market Logic”
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
