Everything About AI
NextFin News - The longest-running bull market of the 2020s turns four years old on October 12, and it arrives at a fork in the road: the artificial-intelligence engine that has powered a 117% gain in the S&P 500 since the October 2022 low is now being asked to carry a heavier load. The index is trading near record highs above 7,700, up about 14% this year, but the rally's next leg depends on something the market has not yet reliably delivered - broad participation beyond the megacap AI builders. After four years of easy money in the chips and cloud giants, the burden of proof is shifting from capital spending to profits, and from a handful of names to the other 493 stocks in the index.
The Rally at Four: Strong, But Narrow
By the standard definition - a gain of at least 20% following a decline of at least 20% from a peak - the current bull market began on October 12, 2022, when the S&P 500 printed its cycle low. As of early October 2026, the index has gained 117%, making this the sixth-best-performing bull market since World War Two, according to Ryan Detrick, chief market strategist at Carson Group. It also ranks as the eighth-longest.
"Four years is not, by any stretch of the imagination, scary with regards to a bull market," said Mark Hackett, chief market strategist for Nationwide. "They don't end of old age; they end from disease."
The question for investors is whether the current illness is merely a cold - a rotation that healthy markets undergo - or something that attacks the rally's central premise. That premise is simple: artificial intelligence is not a speculative theme but a capital-investment super-cycle that will lift corporate profits across the economy.
So far, the evidence for that premise is overwhelming at the top of the market and thin everywhere else. The market capitalization of Nvidia has soared to $5.8 trillion from $286 billion on October 12, 2022, making it the largest company by market value in the world. Thirteen US companies now boast market values of at least $1 trillion - all but two either in the technology sector or with significant AI exposure. The weight of the top 10 companies in the S&P 500 has grown to about 40% from about 28% in October 2022, according to J.P. Morgan Asset Management.
Of the 11 S&P 500 sectors, only technology and communication services - which includes megacap AI players Alphabet and Meta Platforms - have posted stronger gains than the index itself over the full bull run. In other words, the market has not broadened; it has concentrated.
Why Concentration Is Not (Yet) a Bubble
The bear case writes itself: a market where 40% of the index sits in 10 names, financed against a 10-year Treasury yield hovering around 5.2% - recently the highest level in 24 years - is a market begging for a mean-reversion event. History supports the worry. Narrow leadership has preceded most major drawdowns, and the last bear market that led to the October 2022 low coincided with a sharp rate-hiking cycle, the same medicine the Federal Reserve is administering again to bring down high inflation.
But this time the concentration rests on something rarer than multiple expansion: real earnings. S&P 500 earnings are expected to rise more than 35% in 2026, lifted by spending from the hyperscalers to build data centers. For the third quarter of 2026, the estimated year-over-year earnings growth rate for the S&P 500 is 29.5%, which would mark the third-straight quarter of growth above 25%, according to FactSet's October 2 Earnings Insight. And 62% of S&P 500 companies are issuing positive EPS guidance for the third quarter, well above the five-year average of 40% and the 10-year average of 41%.
This is the mechanism that separates the current cycle from the dot-com boom. In the late 1990s, the concentration was in companies with promises. Today, the concentration is in companies with cash flows - and with customers. The four largest hyperscalers - Amazon, Microsoft, Alphabet and Meta - plan to allocate about $725 billion to capital expenditures in 2026, up 77% from last year's already record $410 billion. Goldman Sachs now expects a combined $5.3 trillion of capex spending across the four from fiscal 2025 to fiscal 2030, up from $4.5 trillion before first-quarter earnings season began.
Oxford Economics estimates that about one-third of recent US economic growth stems from AI, counting both the direct investment in AI infrastructure and the wealth effect from stock-market gains lifting consumer spending. "You're seeing that AI theme show up in the economy and in corporate profits," said Anthony Saglimbene, chief market strategist at Ameriprise.
The transmission channel is concrete: hyperscaler capex becomes revenue for chipmakers, server manufacturers, data-center builders, utilities and power-equipment suppliers, which then hire, pay dividends and buy back shares. That is a structural shift in the composition of US corporate investment, not a cyclical sugar high. A structural claim requires evidence of a permanent regime change, and the AI buildout qualifies: general-purpose computing capacity is being repriced downward by orders of magnitude, and that does not reverse on its own. The cyclical leg is separate - it is the pace of spending, the interest rate that finances it, and the quarterly rotation between builders and users - and that leg will mean-revert even if the structural one does not.
"The AI theme is the defining feature of this bull market. What you're seeing in terms of the bull market four years in is, I think, the easy money around AI has been made ... as we get further into this bull market, there is just going to be more pressure on especially technology companies to prove that the spending that they're doing today is actually going to translate into the profits."
That quote from Saglimbene captures the pivot point. The first phase of an infrastructure buildout rewards the builders. The second phase rewards everyone who uses the infrastructure to make money. The market is now demanding that phase two begin.
The Building Blocks Are Starting to Appear
Here is the second-order fact that the concentration narrative misses: the broadening is already underway, just not at the index level. In the first half of 2026, the equal-weight S&P 500 returned 12.1% versus 10.2% for the traditional market-cap-weighted index, according to First Trust Advisors' analysis of index data through June 30. The Mag 7 - Apple, Nvidia, Microsoft, Amazon, Tesla, Alphabet and Meta - accounted for 33.4% of the index but were anything but magnificent in that stretch: three of the seven posted negative returns and collectively subtracted nearly 2 percentage points from the index's total return. Microsoft alone fell 22.5% and shaved 1.4 percentage points off the index. The other 493 companies, by contrast, contributed a combined 10.2 percentage points.
The participation rate tells the same story. In 2023, just 26.9% of S&P 500 stocks outperformed the index - the narrowest market since at least 1995. That pattern persisted in 2024 and 2025, when only 27.7% and 30.5% of stocks beat the index, one of the most concentrated stretches of leadership in nearly three decades. In the first half of 2026, 46.3% of stocks outperformed. That is not a breakout, but it is the clearest inflection in market internals since the bull market began.
The reason the headline index does not yet show the broadening is arithmetic. When the top 10 names command 40% of the weight, the other 490 stocks can outperform on average and the cap-weighted index can still lag. This is why the S&P 500 is a misleading thermometer for market health in 2026. The index is not the market; it is a weighted average of two markets - the AI builders and the AI users - and for most of the past three years only one of them has been allowed to speak.
The catalyst that could lock in the handoff is earnings season. Sam Stovall, chief investment strategist at CFRA Research, laid out a roadmap at the end of 2025 that anticipated exactly this transition. He forecast the S&P 500 would reach 7,400 by year-end 2026 - a target the index has already surpassed ahead of schedule - on the back of 13.5% profit growth, with double-digit increases coming from consumer discretionary, industrials, information technology and materials. His positives for the year were "AI infrastructure buildouts, strong EPS growth, improving capital expenditure visibility and lower rates."
Two of those four pillars are already visible. The third - broad EPS growth - is the missing building block. The fourth - lower rates - is actively working against the market, with the Fed hiking and the 10-year yield at 5.2%.
The Counter-Thesis: Breadth Is a Mirage
The strongest argument against the broadening story is that it confuses a rotation with a regime change. The equal-weight outperformance in the first half of 2026 came largely during a midyear pause in AI enthusiasm, when investors rotated out of the megacaps while waiting for AI spending to deliver returns. As soon as the hyperscalers reaffirmed their capex guidance, the megacaps resumed leadership. And 46.3% of stocks outperforming the index is hardly a broad-market breakout - in a healthy bull market, that number typically runs above 60%, and readings above 80% mark genuine broad strength. As of October 8, 2026, 47% of actively traded US stocks were above their 200-day moving average, with 30% above the 50-day and 43% above the 20-day - a market in transition, not in a sprint.
Angelo Kourkafas, senior global investment strategist at Edward Jones, put the risk plainly: "It is a reflection of fundamental strength and earnings outperformance, but also it introduces some risks. The risk of concentration is that if the prevailing theme goes out of favor, portfolios might feel it in an outsized way." Edward Jones remains overweight equities, he said, but the recommendation is "less aggressive than it was before," with the strategist noting the increasing attractiveness of fixed income. "We still think that the bull market is not about to end ... but it makes sense to us to take some of the risk off the table."
The rate backdrop sharpens the danger. A 5.2% risk-free yield is not the zero-rate environment that powered the 2023-2024 re-rating of AI stocks. Every dollar of future profit is discounted more heavily, which means the earnings conversion the market now demands must be larger and faster than today's valuations can comfortably absorb at this yield. If the 10-year yield pushes toward 5.5% while hyperscaler capex growth slows, the two forces would compound: higher discount rates and lower terminal growth assumptions hitting the same stocks at the same time.
There is also a political risk that has no clean hedge. Markets could be volatile heading into the November 2026 US midterm elections, and the AI buildout depends on policy continuity - power permits, chip subsidies, immigration for engineering talent - that a change in the balance of Congress could complicate. Lerner of Truist Advisory, citing the seasonal pattern, said the fourth quarter of midterm-election years has produced an average gain of 7% and has been positive 84% of the time since 1950 - but added that rising interest rates posed the biggest risk to the upward trend. "We do not expect a straight line higher for markets," Lerner said.
What Comes Next: Three Scenarios
Base case - measured broadening. The third-quarter earnings season confirms that profit growth is spreading beyond the megacaps, with S&P 500 earnings excluding the Magnificent Seven growing in the high single digits to low double digits for full-year 2026. The index grinds higher but with more volatility, and the equal-weight index continues to narrow the gap. The top-10 weight stabilizes near 40% rather than expanding toward 45%. This scenario keeps the bull market intact through year-end and into 2027, supported by the seasonal tailwind cited by Truist Advisory.
Upside case - AI profits go mainstream. The hyperscalers' $725 billion capex plan begins generating visible returns for the AI users - software companies, enterprises, industrial firms - and S&P 500 earnings growth for the ex-Mag-7 cohort accelerates above 15%. The 10-year yield stabilizes or falls on cooling inflation, removing the discount-rate headwind. In this scenario the rally broadens decisively, the top-10 weight rolls over toward 35%, and the bull market enters its fifth year with healthier internals than it has had since 2023.
Downside case - the conversion fails. Hyperscaler capex stays elevated but the profits do not follow, either because AI monetization lags or because the 5.2% yield environment forces a multiple compression that earnings growth cannot offset. The top-10 weight breaks toward 45% as investors retreat to the only names with visible cash flows, the equal-weight index rolls over, and the index becomes even more dependent on a shrinking group of leaders. That is the configuration that has preceded most major drawdowns, and it is the disease Hackett warned about.
The Signal to Watch
The single most important number for the next six months is not the S&P 500 level. It is the earnings growth rate of the S&P 500 excluding the Magnificent Seven. If that cohort fails to reach double-digit earnings growth for 2026 while hyperscaler capex remains above $700 billion, the broadening thesis is wrong: the AI buildout is enriching the builders but not the users, and the concentration at 40% is a fragility, not a strength. Conversely, if ex-Mag-7 earnings accelerate above 15% while the top-10 weight holds or declines, the market has built the additional blocks it needs, and the four-year-old bull market has room to run.
Short term, watch the breadth readings: a sustained move above 60% of stocks above the 200-day moving average would confirm broadening; a drop below 30% would signal the rotation is failing. Medium term, watch whether the Fed's rate path bends lower as inflation cools - the 5.2% 10-year yield is the tax on every growth stock in the index. Long term, the structural question is settled only when AI spending shows up in aggregate productivity data, not just in the earnings of chipmakers.
The bull market did not get to four years by accident. It got there because a genuine technological shift met a genuine earnings cycle. But a rally carried by 10 stocks is a narrow bridge, and the market is now asking whether the other 490 can carry their share of the weight. A four-year-old bull market does not die of old age; it dies when the builders stop earning and the users never start.
Explore more exclusive insights at nextfin.ai.
