TL;DR
· Bank of America warns that there are too many risk signals in the US stock market and advises investors to take profits and control risk.
· Leading AI companies still have revenue and capital expenditure support, but the market has already priced in much of the future growth.
· Related assets: SPY, QQQ, NVDA, MSFT, GOOGL, AMZN, META, AVGO, AMD, SOXX.
US stock market investors are now facing a situation that goes beyond a simple bullish or bearish view.
On one side is Bank of America's US Equity and Quant Strategy team. The team led by Savita Subramanian released a client report titled "Too many red flags. Take profits." on June 5. According to Axios' report on June 9, the report believes that there are too many risk signals in the US stock market and provides a more direct position recommendation: take profits.
On the other side, we have the resilient AI fundamentals. Microsoft, Google, Amazon, and Meta are still increasing AI and data center capital expenditures, while Nvidia's data center demand remains a core anchor of the semiconductor cycle. Unlike the 2000 internet bubble, this round of leaders has shifted to a group of giants with cash flow, profits, cloud revenue, and chip orders.
So the real question has shifted from "Is AI in a bubble" or "Is Bank of America calling the top" to another more difficult question: How should investors understand the current risk in the US stock market when historical top signals and real AI growth exist simultaneously?
The answer may be more uncomfortable than simply being bearish: The AI bull market may not have ended, but it has transitioned from the "buy growth" phase to the "test growth realization speed" phase.
Bank of America's Warning is about Worsening Odds
The value of Bank of America's report lies in placing the current market in a historical risk framework rather than providing an exact top timing.
According to multiple financial media citing the Bank of America report, of the 10 bear market warning signs it tracks, about 70% have been triggered. This ratio is close to the average level seen before multiple S&P 500 peaks since 1990. The Bank of America framework also shows that out of 20 valuation indicators, 17 show statistically high levels of overvaluation for the S&P 500, with 8 exceeding the peak of the 2000 internet bubble. The CAPE (cyclically adjusted price-to-earnings) or P/E10 is around 40, in a historically very high range.
Each of these metrics by itself is debatable. A high valuation does not necessarily mean a drop tomorrow. Historical signals being effective does not mean they are always accurate. AI companies are more profitable, indeed making today different from the year 2000. But when valuation, market breadth, style divergence, and momentum all show extreme readings simultaneously, Bank of America's key point is closer to: the market can still continue to rise, but the odds have worsened.
Market breadth is key here. Despite the index being at a high level, the uptrend is increasingly relying on a few AI and tech leaders. The current U.S. stock market is exhibiting narrow leadership characteristics similar to historical market top phases: a few stocks are driving the majority of the index's gains, the proportion of S&P component stocks above key moving averages is falling, and many individual stocks are far from their respective highs. The index-level strength is masking declining internal participation.
Style divergence is also reinforcing the same signal. Bank of America mentions that the median return difference between the top five percentile stocks in the tech sector and the bottom five percentile stocks is approximately 120 percentage points, the highest since February 2000, close to the 130 percentage points before the peak in March 2000. This is more like a concentrated bet by funds on a few certain narratives, and broad diffusion, typical in a regular bull market, has not appeared.
For investors holding SPY, QQQ, NVDA, or SOXX, the most dangerous aspect of this structure is the reduced margin of error. Of course, the index could still continue to rise, but as the gains are increasingly determined by a few stocks, any deviation in earnings, guidance, capital return on investment, or valuation assumptions of any leading stock could be amplified into a drawdown for the entire portfolio.
This Time, AI Can't Simply Follow the 2000 Playbook
If you only look at Bank of America's valuation and breadth signals, it is easy to directly compare the current market to the year 2000. However, this analogy is only half correct.
The typical characteristic of the 2000 Internet bubble was that many companies lacked a mature business model, and investors mainly traded on the imagination of the "Internet changing the world." The AI leaders of today are different. Microsoft, Google, Amazon, Meta's cloud and AI businesses are already reflected in real revenue, capital expenditure plans, and data center demand. NVIDIA is not just a narrative center but also a chip supplier highly concentrated in profit and cash flow.
NVIDIA's latest financial report provided the strongest support for the bulls. The FY2027 Q1 report released in May 2026 showed quarterly revenue of $81.6 billion, with data center revenue at $75.2 billion, a 92% year-on-year increase. Faced with these figures, simplifying the AI market as a "bubble without fundamentals" is not very convincing.
The AI optimists, including top management of large tech companies and growth investors, are using this to push back against the bubble theory. They believe that this current surge is more akin to an infrastructure cycle: the demand for training and inference is driving GPU, networking, storage, power, and data center construction. Cloud providers are increasing their capital expenditure in exchange for future AI service revenue, while enterprises are integrating AI into software, advertising, search, office, and development processes.
This framework is backed by facts. Over the past few earnings seasons, major cloud providers have consistently emphasized strong AI demand, keeping their cloud businesses growing. NVIDIA's data center revenue has become a key pillar of the U.S. stock market's earnings growth narrative. Companies like Broadcom, AMD, data center, and power infrastructure firms are all part of the same investment chain. The market is willing to assign higher valuations to these companies not just because of a compelling story, but because orders, revenue, and profits are indeed materializing.
This is also why Bank of America's signal should not be rudely interpreted as "the AI bull market is over." If the underlying fundamentals are still improving, the high valuation rally can last longer than historical experience would suggest. Especially in a market where passive funds, index weights, and institutional allocations reinforce the dominance of leading companies, the strong getting stronger is already part of the capital flow mechanism.
However, AI reality does not equate to valuation safety. Here is where a common misunderstanding can arise: that as long as a technological revolution is real, the price is not expensive. Many historical bubbles have actually been built on prematurely and overly pricing in real technology. The internet did indeed change the world, but investors who bought into many internet stocks in 2000 still underwent a prolonged valuation compression.
The core disagreement in the current AI market is shifting from "Is AI valuable?" to "How many years ahead has the market already priced in?" Bank of America's historical signal is important precisely because it reminds investors: even if the fundamentals are true, when the price has already factored in too much good news about the future, the risk will still rise.
Pressure Shifts to Revenue and Cash Flow
The AI bull market is entering its toughest phase, not because demand has suddenly disappeared. The real change is that the market is starting to demand more proof.
Over the past two years, investors were willing to pay a premium for AI leaders because the growth trajectory looked clear: cloud providers were increasing capital expenditure, chip companies were selling more high-end GPUs, data center and networking equipment firms were receiving orders, and future enterprise applications would unlock even greater revenue. As we move past 2026, the market needs to see not just continued investment, but also whether that investment can translate into sufficiently high revenue, profit margins, and free cash flow.
Capital expenditure is the focus of this issue. Microsoft, Google, Amazon, and Meta continue to increase their AI and data center investment, with the direction largely clear, but different institutions and media outlets have significantly different estimates of the specific scale. More importantly, investors have begun to worry about the higher capital expenditure putting pressure on free cash flow and return on investment. This cannot simply be stated as "AI investment cannot be recovered," but after the investment curve steepens, the market's expectations for the return curve will also increase.
For Microsoft, Google, Amazon, and Meta, further increasing AI investment is strategically necessary. Those who pause may fall behind in the cloud, search, advertising, office, models, and developer ecosystem. However, from a shareholder's perspective, the higher the capital expenditure, the more the future financial reports will need to demonstrate that these investments can bring income growth, stable profit margins, and cash flow resilience.
For the semiconductor chain represented by NVIDIA, Broadcom, AMD, and SOXX, the logic is slightly different. They are direct beneficiaries of the AI investment cycle, with orders and profits realized earlier. However, because the market already sees them as the core winners of the AI infrastructure cycle, once downstream cloud providers slow down capital expenditure, postpone purchases, or start emphasizing investment discipline, semiconductor valuations will react first.
This will form a more fragile feedback loop. Cloud giants increase capital expenditure, supporting chip company revenue. Chip companies with high growth support index gains. Index gains and profit upgrades further strengthen the market's confidence in the long AI cycle. If any part of this chain slows down, the market may not necessarily face "the end of AI," but rather a need to reevaluate valuations to match cash-out speed.
Second-Half Financial Reports Need to Demonstrate Growth Can Cover Risks
Bank of America's 70% bear market signal will not automatically turn into a top, and the strong earnings of AI leaders will not automatically eliminate valuation risks. What really needs to be verified next is whether sustained growth can cover these valuation and market structure risk signals.
The most direct observation window will be the second-half financial reports of 2026. Investors need to see the continued growth of AI revenue from large tech companies, while profit margins are not significantly eroded by capital expenditure and depreciation pressures. Cloud providers, while continuing to invest, also need to demonstrate that customer demand is strong enough. Orders and guidance from semiconductor companies like NVIDIA, Broadcom, and AMD will reflect whether downstream investment pace has slowed down.
Another variable is market breadth. If the S&P and Nasdaq continue to hit new highs but fewer stocks participate in the rise, high P/E ratio stocks continue to systematically outperform low P/E ratio stocks, Bank of America's mentioned historical top structure will be harder to ignore. Conversely, if profits spread to more industries and the index no longer relies solely on a few AI leaders, risk signals have the opportunity to be slowly digested over time through performance.
For the average investor, it is currently more suitable to conduct a position and concentration check. Simply saying "bullish on AI" or "bearish on the US stock market" will not solve the problem. AI may still be the most important investment theme in the coming years, but holding onto it amidst rising valuations, market breadth, and capital expenditure pressures has shifted from early trend discovery to a bet on the speed of realization.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
