U.S.-China Relations
NextFin News - US stock futures were little changed as Americans headed to the polls on November 3, with control of both chambers of Congress at stake and the Nasdaq 100 still recovering from its third technical correction of the year. The S&P 500 futures were muted while Nasdaq 100 contracts rose 0.3% as of 7:30 a.m. in New York, clawing back earlier losses, as technology shares attempted a tentative rebound a day after the index tumbled more than 10% from its October peak.
The twin overhangs are straightforward to name and harder to price. Midterm elections have been statistically kind to stocks over the following year, yet the market is trading as if the most probable outcome - a Democratic sweep, in the view of prediction markets - would be a net negative for the AI-heavy megacaps that have carried the index. The AI buildout, meanwhile, is being funded by capital expenditure commitments that nearly doubled in a single year, and the market is now demanding proof that the spending converts into profits rather than promises.
This is not a correction born of recession data or a policy shock. It is a valuation test: how much future AI earnings power can investors reasonably underwrite today, and how much political uncertainty can a market priced for perfection absorb before it asks for a higher risk premium? The answer will determine whether this dip is a buying opportunity or the first chapter of a deeper repricing.
The Election Overhang: Priced-In, But Not Priced Comfortably
On the surface, Election Day should be a non-event for equities. Since 1950, the S&P 500 has gained an average of 15.1%, excluding dividends, in the 12 months following midterm elections, according to investment manager Capital Group. Midterms also tend to punish the president's party, and history suggests the most likely result is a divided government - the gridlock that markets have traditionally rewarded with a stability premium.
But 2026 is not behaving like a typical midterm year. Prediction markets are pointing toward a Democratic sweep of both chambers as the base case: Polymarket traders assigned a 77% probability to that outcome on Election Day, with a Republican sweep at just 6%; Kalshi put the sweep probability at 62%. Decision Desk HQ's model gave Democrats a 63% chance to control the House and Senate, projecting roughly 241 House seats and 52 Senate seats - one more than needed for a majority. Voters will decide all 435 House seats and 35 Senate seats.
Here the market's message is genuinely conflicted, and the conflict is the story. Morgan Stanley strategists, led by chief US equity strategist Michael Wilson, said in a Monday note that a Democratic sweep is already priced in by investors - which should make the actual result a non-event. Yet other investors and analysts have argued the opposite: that the stock market is not yet pricing in a Democratic sweep of Congress, despite what the polls and prediction markets say, and that a blue wave could weigh on share prices because it would hand one party the agenda-setting power to scrutinize fiscal discipline and the AI buildout. Under a divided-Congress scenario, legislative stalemate would make major policy changes harder to enact, which could shift investor attention toward Washington's existing fiscal challenges and the deficit-driven spike in bond yields. Under a Democratic sweep, the housing outlook could improve through affordability measures, but the AI-spending headache would grow: Democrats could ramp up scrutiny of data-center electricity costs and the broader AI buildout, adding uncertainty for hyperscalers already facing questions about enormous capital bills. Under a Republican sweep, deregulation could support investment banks and dealmaking, but deficit-financed spending could keep upward pressure on Treasury yields, worsening the picture for housing and highly leveraged companies.
The asymmetry is clear: every plausible outcome carries a channel through which it could hurt the current market leadership. That is why futures were flat rather than rallying into the vote. Whether the sweep is priced in or not, the market has concluded that nothing on the ballot is a catalyst to buy.
The AI Trade: When Capex Doubles, the Burden of Proof Shifts
The correction that closed out October was not a broad-based panic; it was a concentrated interrogation of the AI thesis. The Nasdaq 100 fell 1.4% on Monday, November 2, to 27,989.82, putting it down more than 10% from its October 6 peak close of 31,224.47 and meeting the technical definition of a correction. The index remains up 11% for 2026, but this is its third correction of the year - it also dipped below the threshold on March 27 and July 29, and both times it later set new records. On October 30, the index reached an intraday low of 28,093.24 before climbing back above the line to close at 28,384.85, a 3.46% decline. A close below that October 30 intraday low would put the index in correction on a closing basis and break the pattern of quick recoveries seen earlier in 2026.
The trigger was earnings. Intel cratered 28.6% after guiding fourth-quarter revenue to a range of $15 billion to $16 billion, below what it had just reported. Amazon joined the list of megacap companies whose results failed to reassure investors, and Apple - reported to have surpassed NVIDIA as the most valuable company at roughly $5 trillion - slipped alongside Palantir as the vote began. The question that moved from the margins to the center of the trade was posed plainly by market commentators: can these companies keep spending, and what happens if they can't?
The spending numbers are staggering. Amazon, Alphabet, Meta, and Microsoft expect to record between $720 billion and $745 billion of capital expenditures in 2026, based on their latest guidance. Amazon raised its plan to approximately $220 billion after initially budgeting about $200 billion; Alphabet's outlook sits at $195 billion to $205 billion; Meta's range is $130 billion to $145 billion; Microsoft's accounting-adjusted estimate is approximately $175 billion. Combined 2025 capex across the same four companies was approximately $410 billion - a year-over-year jump of roughly 77%. Add Oracle, and the five largest US cloud and AI infrastructure providers have collectively committed to spending between $660 billion and $690 billion in 2026.
That is the mechanism behind the selloff. AI infrastructure spending is not funded out of current AI revenue; it is funded out of the core cash flows of cloud, advertising, and enterprise software businesses, underwritten by the assumption that AI will eventually generate returns large enough to justify the depreciation. When capex roughly doubles in a year while AI revenue remains a fraction of the infrastructure being deployed on its behalf, the margin for error collapses. Investors are no longer asking whether AI is real. They are asking whether the economics of the buildout can clear the hurdle rate that a 10%-down index now demands.
Concentration amplifies the risk. Technology stocks' share of the US stock market reached 37.5% as of May 2026, surpassing levels seen during the late-1990s internet bubble - and that figure does not count Alphabet and Meta, which are classified in communication services, or Amazon, which sits in consumer cyclical. All three are heavily involved in AI. When the index's gains are carried by the same companies whose capex plans are under scrutiny, a single disappointing guide can pull the whole complex lower.
Cyclical Dip or Structural Repricing? The Verdict Is Mixed - and That Is the Point
Is this a cyclical fluctuation that will revert, or a structural regime shift that will not? The honest answer splits by time horizon, and getting this wrong flips the conclusion.
On a cyclical read, the evidence for mean reversion is strong. The correction began in June, and Morgan Stanley strategists said in a Monday note that the US market correction that began in June is in its "final innings," adding that the November 3 midterm elections should help remove a major market overhang. The Nasdaq 100 has already fallen into correction twice this year and recovered to new highs both times. Midterm-year drawdowns are the norm, not the exception: during midterm election years, the S&P 500 has typically averaged a 14% drawdown at some point in the second half of the year, according to Jonathan Krinsky, chief market technician at BTIG. By that measure, the market is late to the drawdown, not early to a crash. Earnings outside the AI complex remain resilient, and the post-midterm one-year record - an average 15.1% gain since 1950 - favors buyers who can tolerate near-term volatility.
"We expect near-term clearing events to provide an attractive opportunity to add risk to quality cyclicals that have de-rated as we head into 2027," strategists led by Michael Wilson, Morgan Stanley's chief US equity strategist, said.
But the structural leg is real, and it cuts the other way. A cyclical correction reverts when the temporary driver - sentiment, positioning, a single bad earnings print - passes. A structural repricing happens when the discount rate applied to future cash flows changes permanently. Three forces argue that something structural is underway. First, the cost of capital has shifted: Treasury yields, while off their multidecade highs, remain uncomfortably high, and $100-plus oil prices keep inflationary pressure alive. Brent crude fell roughly 6% on Monday to $106.65 a barrel after four consecutive months of gains, but Goldman Sachs has warned that Brent could reach $140 in 2027 if Gulf production remains 3.5 million barrels a day below its 2025 average. Second, the AI capex cycle is a multi-year commitment, not a one-quarter decision - the spending will hit depreciation schedules for years regardless of whether returns materialize. Third, the market's concentration at late-1990s levels means the same valuation compression that hit the internet bubble can repeat even if the underlying companies are more profitable.
There is also a subtler signal that the regime has shifted. "Investors are no longer overlooking FCF shortfalls as they often did earlier in the cycle, consistent with a more typical mid- to late cycle environment where earnings durability and cash conversion matter more," Wilson said. That is the crux: the market is moving from a growth-at-any-price mindset to a cash-conversion mindset, and the AI buildout is being judged by the new standard.
The correct call, then, is a barbell: a cyclical bounce is likely once election uncertainty clears and the oversold technicals normalize, but the structural ceiling on multiple expansion has been lowered. The market can rally from here without ever reclaiming the valuation it lost. That distinction matters because it separates the trade - buy the dip - from the regime - expect lower forward returns from the AI leaders.
The Counter-Thesis: What If the Market Is Overthinking a Routine Midterm Drawdown?
The strongest case against this analysis is that the market is attaching regime-level significance to a routine midterm drawdown in an otherwise healthy economy. The bull case rests on three pillars. First, earnings breadth outside the megacaps has held up; the correction is a sector rotation, not an economic signal. Second, the AI spending is not speculative - it is backed by contracted cloud demand, and hyperscalers have repeatedly guided to multi-year customer commitments that justify the infrastructure. Third, a divided government - still a 29% outcome on Kalshi and a 19% outcome on Polymarket - would deliver the gridlock premium that historically supports equities, and even a Democratic sweep would face a presidential veto on anything truly market-hostile.
There is merit to each point. But the counter-thesis underestimates the second-order channel. The issue is not whether AI spending is justified in isolation; it is what happens to the rest of the market when the megacaps stop being the marginal buyer of growth. If hyperscalers slow capex to protect free cash flow, the semiconductor, data-center, and power-equipment suppliers that have ridden the buildout face an earnings downgrade cycle that spreads well beyond the Nasdaq 100. And if a Democratic sweep brings serious scrutiny of data-center power costs, the discount rate applied to AI earnings rises even as the revenue trajectory slows - a double compression that a 15.1% historical post-midterm average cannot offset in the near term.
The signal that would prove the cyclical-bounce thesis wrong is specific and observable: a Nasdaq 100 weekly close below the October 30 intraday low of 28,093.24, accompanied by a failure of the 10-year Treasury yield to break lower on the election result. That combination would confirm that both the political and the AI channels are transmitting into a higher risk premium, not a temporary sentiment dip. Conversely, a decisive Republican hold of either chamber that sends yields down and the Nasdaq back above its pre-correction trend would validate the bull case.
What Comes Next: Three Scenarios and the Signals to Watch
The base case is a short-term relief rally once the election outcome is known, followed by a resumption of the AI earnings test when Nvidia reports fiscal third-quarter results on November 17, after the market close. The company has guided to revenue of $108.0 billion, plus or minus 2%, for the quarter. Investors should also watch the October jobs report on November 6 and the actual balance of power in Congress. Each will either confirm the cyclical-bounce read or harden the structural ceiling.
Upside scenario: a divided Congress or a narrow Republican hold, combined with a softer-than-expected jobs print that keeps rate-cut expectations alive. In that world, the Nasdaq 100 reclaims the 28,093 level quickly, the post-midterm seasonal tailwind takes over, and the correction joins March and July as a buying opportunity. Beneficiaries would include rate-sensitive housing, small caps, and the AI infrastructure suppliers that get a second wind from continued spending.
Downside scenario: a Democratic sweep that triggers a bond-market selloff on deficit fears, paired with any hint from Nvidia or the hyperscalers that capex growth is peaking. In that world, the correction deepens into a bear market for the AI complex, and the 37.5% technology concentration becomes the mechanism through which losses transmit to the broad index. The exposed are the semiconductor equipment makers, data-center REITs, and power-equipment suppliers whose valuations assume uninterrupted capex growth.
Short-term, sentiment and liquidity will dominate: election results, the jobs report, and positioning around the Nvidia print. Medium-term, fundamentals will decide: whether AI revenue growth accelerates fast enough to meet the doubled capex. Long-term, the structural question is whether the AI buildout proves to be a durable productivity regime shift or a capital-intensive boom that leaves excess capacity - and history suggests the answer usually arrives later than the market wants to wait.
The takeaway: this is not the market pricing a recession. It is the market repricing the AI trade's margin of safety at the exact moment political uncertainty removes the policy cushion - and a 77%-probable Democratic sweep is the outcome the current leadership is least equipped to absorb.
Explore more exclusive insights at nextfin.ai.
