Everything About AI
NextFin News - The market's AI reckoning has a price tag, and Morgan Stanley has put a number on it. The bank's strategists, led by chief U.S. equity strategist Mike Wilson, see the S&P 500 sliding to roughly 7,100-7,200 before finding material support - a further drawdown that would mark the deepest correction since the Iran war sell-off earlier this year. The trigger is not a recession scare or an earnings collapse. It is a quieter, more structural question: whether credit markets can fund the AI spending boom that has powered this cycle, and whether the companies promising that spending can earn an adequate return on it.
The SPDR S&P 500 ETF closed at 752.10 on Monday, down nearly 1 percent, after zigzagging lower from its early-October record as oil pushed past $100 a barrel and post-earnings declines hit several of the market's highest-profile AI names. The ETF now trades below its 50-day average of 767.49, having pulled back from a 52-week high of 781.62, yet it remains up sharply from its 52-week low of 629.28. Morgan Stanley is not turning bearish. Its base case keeps a year-end target of 8,000 for the S&P 500, implying a 12 percent rally from the trough. The path there, however, runs through what strategists call "clearing events" - a set of hurdles that must resolve before the next leg higher can begin. The first, and in many ways the most consequential, is AI.
The Situation: A Correction That Is Supposed to Look This Bad
Wilson frames the pullback not as the start of a bear market but as a "classic mid-cycle transition." In a late-October appearance, he laid out the mechanics plainly: the peak rate of change in earnings revisions arrived roughly a year after the recession troughed, right on schedule, and the market has responded the way it usually does at this point in the cycle.
"We've described this correction as a classic mid cycle transition. The peak rate of change in earnings revisions arrived right on schedule about a year after the recession troughed."
The damage underneath the index, however, is broader than the headline suggests. Wilson noted that about 60 percent of Russell 3000 stocks are already down more than 20 percent from their June peaks, with early-cycle groups - semiconductors, autos, consumer cyclicals and capital goods - hit hardest. On that measure, much of the correction has already happened stock-by-stock even as the index itself has held up. The remaining risk is one more index-level drawdown that flushes out the last pockets of crowded positioning.
That is why the bank's 7,100-7,200 support zone matters. It is not a random round number; it is the level where, in Morgan Stanley's view, the correction should find material support if the index moves lower.
"If the index does move lower, I believe 7,100 to 7,200 on the S and P 500 is where the correction should find material support."
The reset is being driven by three clearing events. First, AI: investors are questioning the return on capital, the cost of capital, and whether credit markets can fund all the spending that has been promised. Second, the midterm elections, which bring uncertainty but have historically been followed by strong six- to 12-month returns. Third, and most important, liquidity: financial conditions have tightened sharply, in a pattern that Wilson says is beginning to resemble 2018 and is not far from 2022, despite very modest tightening from the Fed so far.
The Analysis: Why AI Became the First Clearing Event
The Capex Boom Outran the Financing Story
To understand why AI moved to the front of the clearing queue, start with the scale of the spending. A year ago, Morgan Stanley projected combined capital expenditure for the largest hyperscalers at around $450 billion in both 2026 and 2027. That estimate has since moved to roughly $800 billion in 2026 and $1.16 trillion in 2027. In other words, the AI infrastructure buildout has roughly doubled in a year, and the 2027 figure alone exceeds the annual capital spending of large swaths of the corporate sector.
This is the structural heart of the story. AI has become a capital-spending story, and increasingly a credit-market story. Data centers, power infrastructure, chips and networks now shape equities, credit, rates and even commodities. The spending is real, it is accelerating, and it is not being funded entirely from cash.
Here lies the tension. Companies will borrow to finance part of the buildout, bringing a wave of new corporate bonds to market. Even issuers that look financially healthy may face investors demanding better terms when there is so much new supply to absorb. AI can support earnings, but it can also put pressure on financing costs and credit spreads - and that is exactly what the market is now pricing in. Credit spreads on data-center-related bonds have widened in recent months, reflecting higher borrowing costs, a greater perceived risk in holding AI-related debt, and a sharp increase in supply.
Wilson put the mechanism in plain terms:
"Investors are questioning the eventual return on capital, the cost of capital, and whether credit markets can fund all the spending that has been promised. ... Over the next few weeks, the market will need to absorb more supply... I view the supply as the event itself. Once it clears, the risk should begin to fade."
That framing - supply absorption as the event itself - is the key insight. The market is not waiting for AI demand to be disproven. It is waiting for the financing pipeline to clear, for the lowest-quality deals to be priced or rejected, and for the cost of capital to stabilize. Once that happens, the risk premium should begin to fade even if the fundamental AI thesis is unchanged.
Cyclical Correction, Structural Buildout: Keep the Two Separate
The most important analytical distinction in this reset is between the cyclical and the structural. The near-term correction is cyclical: a mid-cycle transition driven by positioning, liquidity and the digestion of a rapid rise in capital spending. Cyclical corrections mean-revert. They end when sentiment reaches an extreme, when positioning washes out, and when the clearing events resolve. Morgan Stanley's own indicators point toward that outcome: hedge fund positioning has come down meaningfully and is near the lows of the past twelve months, while the S&P 500 has already given back most of its post-summer gains and now sits roughly 8 percent below the early-October peak that would correspond to the 7,100-7,200 support zone.
But the AI infrastructure buildout underneath it is structural. A doubling of hyperscaler capex from $450 billion to $800 billion in a single year is not a cyclical fluctuation; it is a regime shift in how the economy allocates capital. Structural shifts do not revert on their own. They change the baseline. The error investors make is blending the two - either assuming the cyclical drawdown disproves the structural trend, or assuming the structural trend protects them from the cyclical drawdown. Both are wrong. The correction can be real and the buildout can be durable at the same time.
There is a second structural shift embedded in this reset: the rotation of leadership within AI itself. The first phase of the AI rally rewarded the enablers - the chipmakers, the power suppliers, the data-center builders. The next phase, Morgan Stanley argues, will favor the companies using the technology to lift revenue and profits. Wilson has said the market's next excitement will come from "the actual users of this technology leading to better revenue growth and profitability," rather than the builders of AI infrastructure. That is a rotation from promise to practice, from capex to cash flow, and it is a structural re-pricing within a structural trend. The bank's economists have already lifted expectations for S&P 500 margin expansion by 40 basis points in 2026 and 60 basis points in 2027, a sign that the productivity gains from AI adoption are expected to compound rather than disappear.
The Second-Order Question: What the Bond Market Is Really Pricing
The first-order read of this story is simple: AI spending concerns are pressuring technology stocks. The second-order question is more consequential. When credit spreads widen on data-center debt, the market is doing more than repricing a sector. It is tightening financial conditions for the entire economy - and doing the Fed's tightening for it.
Wilson's point on financial conditions is the hinge of the whole thesis. Rates are higher, balance-sheet growth is slower, demand for capital is stronger, and energy prices have risen. The result is a tightening in financial conditions that resembles 2018 and is not far from 2022, despite very modest tightening from the central bank.
"The tightening is beginning to resemble 2018, and it's not far from what we experienced in 2022 despite very modest tightening from the Fed so far."
Follow that chain one step further. If tighter financial conditions create funding stress or destabilize credit markets, the Fed and Treasury have tools to respond. The question is whether conditions must deteriorate further first. In Morgan Stanley's telling, that deterioration may be what clears the path for a policy pivot and the next leg higher for stocks. The paradox is deliberate: the very thing causing the correction - tightening financial conditions - is also what sets up the policy response that ends it.
That is the second-order implication most investors are missing. The AI capex debate is not just about whether AI earns its keep. It is about whether the financing of AI becomes the mechanism that forces a policy response, which then re-rates every duration-sensitive asset in the market. Semiconductors, in Wilson's view, are the canary: he sees the semiconductor index falling another 10 to 15 percent into mid-November, and he treats that move as a benchmark for the broader correction.
The Counter-Thesis: This Time the Capex Cycle Could Break
The strongest case against Morgan Stanley's constructive view is not that AI is a bubble. It is that the financing structure of this capex cycle is more fragile than the last technology buildout, and that a funding shock could turn a mid-cycle correction into something deeper.
The bear argument runs like this: hyperscaler capex of $800 billion in 2026 and more than $1 trillion in 2027 requires revenue growth that has not yet materialized. If AI monetization lags - if the next wave of AI applications fails to convert into enterprise spending at the pace the models assume - then the spending will have to be cut, not merely refinanced. A capex cut would hit earnings for the very companies carrying the index, credit spreads would widen further, and the "mid-cycle transition" would become a late-cycle slowdown. In that scenario, 7,100-7,200 is not support; it is a wayside on the way lower.
This counter-thesis has a named foundation. Skeptics point to the gap between capital committed and revenue recognized, to the concentration of AI spending in a handful of hyperscalers, and to the fact that about 60 percent of the Russell 3000 is already in bear-market territory - a sign that the damage is systemic, not confined to a narrow trade. If earnings revisions, which have already peaked, roll over into absolute declines rather than merely slowing, the mid-cycle script no longer applies.
The falsifying signal is specific and observable. Morgan Stanley's constructive view depends on the correction finding support at 7,100-7,200 on the S&P 500 and on the semiconductor index stabilizing after its expected 10-15 percent decline into mid-November. If the S&P 500 breaks decisively below 7,000 on a weekly closing basis while credit spreads on data-center debt continue to widen, the mid-cycle-transition thesis is wrong. At that point the move is no longer a clearing event; it is a regime break, and the year-end 8,000 target should be abandoned.
What Comes Next: Three Horizons and Three Scenarios
The forward look splits cleanly by time horizon, and the horizons point in different directions.
In the short term - the next few weeks into mid-November - sentiment and liquidity dominate. Semiconductors are the watch item, with a further 10 to 15 percent decline expected into mid-November. Sentiment has not yet reached the extreme that typically marks a durable low, even though hedge fund positioning is near 12-month lows. Volatility can persist through the midterm elections and the absorption of new AI-related bond supply.
In the medium term - six to twelve months - fundamentals take over, and the setup is constructive. Midterm elections have historically been followed by very strong six- to 12-month returns. If the clearing events resolve - if AI financing supply is absorbed, if election uncertainty passes, and if tighter financial conditions prompt a policy response - the path to 8,000 by year-end 2026 remains the base case. The recovery is expected to broaden beyond the market's biggest names, reversing the narrow rally that has left many stocks behind.
In the long term - beyond 2027 - the structural trend is intact. AI infrastructure spending is forecast to keep rising, and the rotation from enablers to adopters should support a broader earnings expansion. Morgan Stanley's economists have lifted expectations for S&P 500 margin expansion by 40 basis points in 2026 and 60 basis points in 2027, a sign that the productivity gains from AI adoption are expected to compound rather than disappear.
Three scenarios frame the path:
- Base case: The S&P 500 tests the 7,100-7,200 support zone, semiconductors fall another 10-15 percent into mid-November, the clearing events resolve, and the index rallies toward 8,000 by year-end. Trigger: stabilization in the semiconductor index and absorption of AI-related bond supply.
- Upside case: Financial conditions ease faster than expected, a policy pivot arrives, and the recovery broadens early. The index reaches 8,000 ahead of schedule, with adopters of AI leading alongside the enablers. Trigger: a decisive break above recent resistance on improving market breadth.
- Downside case: AI monetization disappoints, credit spreads widen further, and the S&P 500 breaks below 7,000 on a weekly close. The mid-cycle script fails and the correction deepens into a late-cycle slowdown. Trigger: an S&P 500 weekly close below 7,000 accompanied by widening data-center credit spreads.
The reset Morgan Stanley is describing is not a repudiation of AI. It is the market's way of separating the spenders from the earners, the promises from the cash flows, and the cyclical noise from the structural signal. The investors who get this right will not ask whether AI is real. They will ask who pays for it, who profits from it, and at what price the risk is finally cleared.
Explore more exclusive insights at nextfin.ai.
