The S&P 500 is still trading close to record highs, but the market underneath the index is telling a very different story.
That is what makes the current setup interesting for investors.
The headline numbers still look healthy. Corporate earnings are strong, AI optimism remains powerful, and some of the biggest technology stocks continue to push the broader market higher.
But market breadth has deteriorated sharply, retail investors are becoming more cautious, and a growing share of the index is being carried by a relatively small group of large companies.
None of this automatically means a market crash is coming.
It does mean investors may want to look beyond the S&P 500 headline and ask a more important question:
How healthy is the rally underneath the surface?
The S&P 500 is hiding a weaker market underneath
One of the clearest warning signs is market breadth.
Only around 25% of S&P 500 stocks are currently trading above their 50-day moving average, according to the material in the source. That is down sharply from roughly 70% in mid-August.
The picture is similar over the longer term.
Only about 47% of S&P 500 stocks are above their 200-day moving average, while another reading cited from Morgan Stanley puts the figure around 49%.
Either way, the message is clear.
The index may be holding up, but fewer stocks are participating.
A healthy bull market usually has broad participation. When more companies are moving higher alongside the index, it suggests that investors are buying across the market rather than concentrating their bets in a handful of names.
Right now, that participation has narrowed considerably.
The Magnificent Seven are doing a lot of the heavy lifting
The recent strength in the S&P 500 has been closely tied to the continued enthusiasm around artificial intelligence.
Investors have increasingly returned to large technology and AI-linked companies, including Nvidia, Meta, Microsoft, Alphabet, Amazon and other major players.
That concentration has become one of the biggest structural features of the current market.
According to the source material, the 10 largest S&P 500 companies account for around 39% of the index’s total value, or approximately $27.4 trillion out of a roughly $70.4 trillion market.
That is a remarkable level of concentration.
And it creates an important dynamic.
When those companies perform well, the entire index can look strong.
But if sentiment turns against them, the impact can spread far beyond individual stocks.
The stronger the concentration becomes, the more important those companies become to the health of the overall index.
AI is powering the rally, but investors are asking a bigger question
There is a good reason investors have been willing to keep paying up for technology and AI stocks.
The earnings story has been strong.
The source notes that analysts have actually raised their third-quarter S&P 500 earnings expectations by 1.3% since June 30.
That is unusual.
Historically, analysts have tended to cut earnings expectations as a quarter progresses. Over the past five years, quarterly earnings expectations have fallen by an average of 2.2%.
This time, expectations are moving in the opposite direction.
The S&P 500 is also expected to deliver approximately 29% year-over-year earnings growth for the third quarter, which would represent the third consecutive quarter with growth above 25%.
So this is not simply a story about investors buying stocks without any fundamental support.
Earnings are doing real work.
The bigger question is whether that earnings growth can continue at a pace that justifies today’s expectations.
That is where the AI investment cycle becomes particularly important.
The AI spending boom has become a market-wide bet
The biggest technology companies are spending enormous amounts of money on data centers and AI infrastructure.
Amazon, Meta, Microsoft and Alphabet alone reportedly spent more than $300 billion on data centers during the first half of 2026, according to the source.
That spending can create enormous opportunities for chipmakers, cloud companies, infrastructure providers and other parts of the technology ecosystem.
But eventually, investors will want to see a return on that spending.
The question is no longer simply whether AI will change the economy.
Most investors already believe it will.
The harder question is:
How much profit will companies actually generate from all this investment?
The source cites Goldman Sachs estimates suggesting hyperscalers could need roughly $300 billion in AI revenue over the next few years simply to break even, with around $1 trillion annually needed for healthier profit margins.
That creates a high bar.
If AI monetization keeps accelerating, today’s valuations could look much easier to justify.
If the returns take longer to arrive, the market could become much less forgiving.
Valuations are sending another warning
There is another reason investors are becoming cautious.
Stocks are not cheap by historical standards.
The source cites a Shiller CAPE ratio above 40, with another reading putting it at around 41.5. That is close to levels seen during the dot-com bubble.
The Buffett indicator is also close to 240%, another historically elevated reading.
But there is an important distinction here.
High valuations are a warning, not a timing signal.
A market can remain expensive for years.
Investors who sell simply because valuations look stretched can miss substantial gains if earnings continue to grow and the market keeps moving higher.
The S&P 500 itself provides a useful reminder of that.
Valuation concerns were already visible years before several major market downturns, yet investors who exited too early would have missed significant subsequent returns.
So the takeaway is not that expensive stocks must immediately fall.
It is that future returns may become harder to achieve if earnings fail to keep pace with elevated expectations.
Retail investors are starting to pull back
Another interesting signal is coming from individual investors.
According to Vanda Research data cited in the source, retail investors purchased only about $1 billion of individual stocks over the past 20 trading days.
That was the lowest 20-day total in at least two years.
Total retail equity purchases have also fallen to around $10 billion over the same period, close to the lowest level since October 2024.
That tells us sentiment among individual investors is changing.
Higher oil prices, rising Treasury yields and concerns about the economy appear to be making investors more cautious.
But institutional investors have not completely backed away.
The S&P 500 remains near record levels, supported by strong earnings expectations and continued enthusiasm around technology.
That creates an interesting split in the market.
Retail investors are getting more defensive while institutional confidence remains stronger.
New lows are another piece of the puzzle
Breadth problems are not limited to moving averages.
As of the Monday session referenced in the source, new 52-week lows on the NYSE had outnumbered new highs for 10 consecutive trading sessions, and in 14 of the previous 15 sessions.
That is not what you would normally expect to see when an index is sitting near record territory.
It suggests that underneath the headline index performance, a meaningful number of individual stocks are struggling.
The same pattern can be seen in the Russell 3000.
Around 51% of its stocks had experienced at least a 20% pullback since June, according to the cited Morgan Stanley analysis.
So while the major index may not look particularly nervous, a large number of individual companies are already experiencing significant declines.
This is not necessarily the start of a crash
This is where the story gets more complicated.
It would be easy to look at weak breadth, high valuations and concentration and conclude that a major market decline is inevitable.
The data does not support that certainty.
The source itself makes an important point: weak breadth is not an ultimate sell signal.
There is also a strong argument that the current market is not showing the kind of widespread euphoria that often accompanies major long-term market tops.
Investor sentiment remains divided.
Some investors are extremely optimistic about AI and technology.
Others are increasingly worried about valuations, interest rates, inflation, energy prices and geopolitical risks.
That lack of broad-based euphoria matters.
Major market tops are rarely created simply because a few stocks become expensive. They typically involve much wider participation and confidence.
Today’s market looks concentrated and expensive, but it does not necessarily look like everyone is blindly bullish.
The real risk may be concentration, not AI itself
It is easy to frame this as an argument against AI stocks.
That would miss the bigger point.
AI has delivered genuine business growth, and companies across the technology sector are seeing enormous demand for computing power, chips, cloud infrastructure and data centers.
The risk comes from how much of the broader market now depends on that story continuing.
If Nvidia, Meta, Microsoft, Alphabet and other major AI beneficiaries continue delivering strong earnings, the S&P 500 can remain resilient.
But if earnings disappoint, capital spending slows, AI monetization takes longer than expected, or investors simply decide valuations have run too far, the effect could be amplified by the concentration of the index.
When a small number of companies drive a large share of market returns, their problems become everyone’s problem.
What investors should watch next
The next phase of the market could come down to a few key indicators.
1. Earnings growth
Strong earnings are currently providing an important foundation for the market. Investors will want to see whether companies can continue delivering results that justify their valuations.
2. Market breadth
If more S&P 500 stocks begin moving above their 50-day and 200-day averages, it would suggest the rally is becoming healthier.
If breadth continues deteriorating while the index remains near highs, the divergence becomes harder to ignore.
3. AI spending and returns
Investors will increasingly want evidence that massive AI infrastructure spending is translating into sustainable revenue and profits.
4. Interest rates
Higher Treasury yields can put pressure on equity valuations, particularly when investors are already paying high prices for future growth.
5. The biggest technology stocks
The performance of the largest AI-linked companies could remain one of the most important drivers of the entire index.
If they continue to lead, the S&P 500 can remain strong.
If leadership breaks down, the market’s concentration could magnify the impact.
The bigger picture for investors
The current market is sending two messages at the same time.
One is optimistic.
Corporate earnings are strong, AI investment remains powerful, and analysts are raising expectations rather than cutting them.
The other is cautious.
Market breadth is weakening, valuations remain elevated, retail participation is cooling and the S&P 500 has become heavily dependent on a small group of companies.
Both can be true.
That is probably the most important takeaway.
This is not a simple “sell everything” setup. It is also not a market where investors should ignore the warning signs simply because the index is near a record high.
For long-term investors, the better question may be less about predicting the next correction and more about understanding what is actually driving portfolio returns.
If the S&P 500 continues climbing because earnings broaden across sectors, that would be a healthier signal.
If it keeps rising while fewer and fewer companies participate, the rally becomes increasingly dependent on a small group of winners.
The index may still be climbing. The question is how many stocks are climbing with it.
And right now, that number is getting smaller.