US Stock Market Crash: $2.5 Trillion Wipeout, Nasdaq Sinks 5.5%, Dow Drops 1,833 Points

On Wednesday, Wall Street celebrated the US election result. On Thursday, it started worrying about what that result might actually change. In between those two reactions, the market’s biggest growth story ran into a very physical problem: building the infrastructure to deliver it.

The Nasdaq Composite plunged 5.5%. The S&P 500 fell 4.4%, and the Dow Jones Industrial Average dropped 1,833 points. Reuters reported that companies in the Nasdaq-100 lost about $2.5 trillion in market value as the selling tore through technology shares.

The immediate concern was that the Democratic sweep in the midterm elections could strengthen resistance to AI data centres. Add weak spots in chip earnings, expensive growth expectations and pressure on companies financing the AI buildout, and Wednesday’s relief rally became Thursday’s rout.

For investors, the difficult question is what to make of that reversal. Has the market become unnecessarily frightened, or are share prices finally catching up with risks that were easy to overlook while everything was rising?

How Far Did the Dow, Nasdaq and S&P 500 Fall?

Thursday’s decline more than erased Wednesday’s gains across all three major benchmarks. Here is the reversal side by side:

Sources: Associated Press market-close reports for November 4 and November 5. Percentage changes are rounded to one decimal place.

The damage spread beyond technology: all 11 S&P 500 sectors finished lower. The major benchmarks recorded their biggest percentage declines since April 4, 2025, during the tariff turmoil.

The Nasdaq Composite also closed approximately 11.9% below its October 6 record closing high. That has put it in a correction, the market term for a decline of at least 10% from a recent peak.

Why Did the Post Mid Term Election Rally Turn Into a Sell-Off?

Let’s start with what investors liked on Wednesday.

Democrats had won both chambers of Congress, setting up a Democratic Congress alongside President Donald Trump’s Republican White House. Investors initially welcomed the removal of election uncertainty and the prospect of a government in which sweeping policy changes would be harder to pass. Sharply lower oil prices and easing Treasury yields helped the mood.

That combination can support a wide range of businesses. Cheaper fuel can ease costs and inflation pressure, while lower bond yields can make future corporate profits more attractive. Those were reasons for the broader market to rally.

But there was already a warning inside the celebration. Chip stocks were falling even while the major indexes were rising. Nvidia ended lower on Wednesday, while Micron, Marvell and Sandisk each fell 7% or more.

As we explained in our breakdown of the post-election rally, investors were weighing the benefits of cheaper energy against the possibility of a slower AI infrastructure rollout. Different companies stood on different sides of that calculation.

Thursday brought the second concern to the centre of the market. A result that looked reassuring for broad economic policy also raised uncomfortable questions about the speed at which data centres could be approved, powered and built.

What Changed for AI Data Centres After the Midterms?

An AI service may appear as a small box on a phone screen. Behind it sit buildings filled with processors, memory, storage and networking equipment. Those buildings need land, electricity, cooling, water where required, and permission to operate.

The election sharpened the political risks around that last part.

In Texas, governor-elect Gina Hinojosa’s campaign platform called for a moratorium on new data centres until lawmakers establish protections for residents. Her proposals and public statements focused on community participation, electricity costs, water and tax subsidies. That gave investors a concrete policy position to assess in a major data-centre state.

The resistance also extended beyond one governor’s race. Duane Morris Government Strategies’ updated election tracker counted outcomes limiting development in 40 of 41 local data-centre ballot measures voted on November 3. The tracker described these as unofficial results and included several kinds of measures, including non-binding advisory votes. Forty restrictive outcomes therefore does not mean forty identical, immediately effective bans.

For investors, the financial issue is the effect on each project. An advisory vote, a temporary permitting pause and a binding prohibition can produce very different outcomes. Some projects may proceed, some may require more expensive safeguards, and others may move or face substantial delays.

The new congressional terms begin on January 3, 2027. Thursday’s prices were reacting to expectations about future decisions, alongside the existing local pushback. An election result by itself does not cancel every construction project or establish a nationwide ban.

That distinction matters because the damage depends on how much capacity is affected, for how long, and whether customers can build elsewhere. A delayed project and a permanently abandoned investment deserve different earnings assumptions.

Why Did Nvidia, AMD, Micron and Sandisk Fall So Much?

The companies supplying the AI expansion had the most immediate reason to worry about its construction schedule. The semiconductor index fell about 12%, while individual losses were larger in several memory and storage names.

Source: MarketWatch’s November 5 closing report on chip and storage stocks. Figures are deliberately rounded.

Consider what happens when a customer postpones a data centre. The processor supplier may have to wait longer for deliveries, the memory supplier may face a different demand schedule, and the networking company may see equipment installation shift into another quarter.

For memory and storage businesses, the concern can extend to pricing. If expected demand weakens while supply keeps expanding, assumptions about selling prices and profit margins may also need to change. That makes the earnings effect potentially larger than a simple shift in shipment dates.

This explains the sensitivity of semiconductor stocks to construction news. Their share prices reflect expectations about future orders and profits. Investors can reduce those valuations well before a company announces a cancelled contract.

Qualcomm added an earnings warning

There was also fresh company news for investors to digest. Reuters reported that Qualcomm’s smartphone chip sales missed expectations, adding pressure to the sector.

Qualcomm’s own filing showed fourth-quarter handset revenue of $5.137 billion, down 26% from a year earlier, and adjusted earnings per share of $2.17, down 28%. It also described broad increases in semiconductor input costs and expected the full margin benefit from its pricing actions to emerge in the second half of fiscal 2027.

The same release reiterated a $5 billion fiscal 2027 data-centre revenue target. That contrast is useful: a company can have a promising future business while its existing earnings remain under pressure. The smartphone figures concern a different end market from AI data centres, but they reinforced the need to examine each chip company’s actual sources of profit.

Selling pressure can amplify the original concern

A large price move can also set off a second round of decisions. Funds using borrowed money may need to shrink positions, while managers facing risk limits may reduce exposure across related holdings.

Reuters quoted Nationwide strategist Mark Hackett identifying aggressive institutional selling and deleveraging as major forces behind the rout. That was a market participant’s assessment of the flows, rather than a measured breakdown of every transaction. It helps explain how concerns about future earnings can become a much larger move in the share price over a single session.

Why Did Stocks Fall Even as Treasury Yields Dropped?

There is another detail that changes how this sell-off should be understood. US government bond yields fell during Thursday’s equity decline.

The Treasury’s official daily data show the 10-year yield falling from 4.90% on November 4 to 4.84% on November 5. The two-year yield fell from 4.58% to 4.46%. Bond prices generally rise when their yields fall.

Lower yields can help share valuations, but that benefit can be overwhelmed by a deterioration in expected profits. The combination of falling stocks and rising government bond prices is also consistent with investors seeking safety.

And a lower government borrowing rate does not guarantee easier financing for every business. Lenders can demand a bigger additional return from a risky corporate borrower. A financially stretched data-centre operator can therefore face pressure even while Treasury yields are falling.

The relevant concern here is the amount of cash a business can generate, when that cash arrives, and whether it can finance the wait.

Does This Mean the AI Boom Is Over?

Thursday’s prices show a loss of confidence in the pace and profitability of the buildout. Establishing whether the underlying investment cycle is breaking requires evidence about customer demand, project execution and returns on spending.

The distinction between companies is especially important because the same equipment purchase appears differently in two businesses’ accounts. For the supplier, it is revenue. For the cloud company placing the order, it is capital spending.

A large cloud operator could preserve cash by moderating construction, provided it can still serve its customers and grow revenue. If capacity constraints instead prevent it from meeting demand, that same spending restraint becomes a problem. Slower spending has to be judged alongside the revenue it supports.

A smaller operator carrying debt faces another calculation. Imagine, purely as an illustration, that it has financed equipment and hired staff but cannot open its new facility for six months. Interest and some operating costs continue, while the expected revenue starts later. A business with ample cash can absorb that wait more easily than one depending on another round of funding.

This is why “AI stocks” is too broad a category for a single investment conclusion. The financial consequences depend on whether a company supplies equipment, operates infrastructure or sells services; and on the strength of its balance sheet.

Is the US Stock Market Crash an Opportunity?

Lower prices improve the starting point only to the extent that the business remains capable of delivering the profits being assumed. A large fall from an earlier high says little about that high’s relationship to sustainable earnings.

Here is a simple illustration. Suppose a share costs $200 and is expected to earn $5 per share over a year. Its price-to-earnings ratio, or P/E, is 40: investors are paying $40 for each $1 of expected annual profit.

Now let the share price fall 20% to $160.

Illustrative calculations: P/E equals share price divided by expected annual earnings per share. These are assumed figures, not forecasts for a particular stock.

The same 20% price decline produces three different valuation outcomes. The stock becomes cheaper relative to earnings in the first case, equally expensive in the second and more expensive in the third.

That is the work hidden inside the phrase “opportunity after a crash.” Investors need to reconsider earnings, margins, financing costs and delivery schedules before treating the lower quote as a bargain.

Thursday deserves to be treated as a serious reassessment of AI growth assumptions. 

The more credible opportunity would be a business whose ability to generate cash remains resilient after allowing for realistic project delays. The weaker case would be a company whose apparent discount depends on every optimistic assumption surviving intact.

The fall creates a reason to revisit the numbers. The size of the fall, by itself, cannot settle the valuation question.

What Could Help US Stocks Stabilise?

The next useful evidence needs to address the concerns behind the decline. Four developments would make the situation easier to assess:

• Clearer project rules. Investors need to know which proposals affect existing permits, new construction, grid connections or particular locations. A workable timetable would allow companies to give more credible delivery and revenue estimates.
• Orders that translate into cash. Customer spending plans matter most when projects progress, equipment is accepted and suppliers collect payment. Guidance backed by execution would carry more weight than another large investment announcement.
• Financing that remains available. Businesses needing fresh capital must show that they can fund construction through delays at an affordable cost. The terms of that funding matter as much as its headline size.
• Broader market resilience. An improvement extending beyond a handful of large technology shares would suggest that confidence is recovering across more businesses. One strong rebound would provide less evidence about lasting stability.

There is also an immediate economic test. The Bureau of Labor Statistics scheduled the October employment report for Friday, November 6, at 8:30 AM US Eastern time, or 7:00 PM IST. Employment and wage data could change expectations about growth and interest rates, adding another source of volatility.

Author’s Take: What Should Investors Make of This Crash?

The most revealing part of this episode is that the weakness was visible during Wednesday’s rally. The companies supplying the AI expansion were already reacting to a less comfortable construction outlook; Thursday spread that concern across a market whose expectations had become heavily tied to AI.

The strongest businesses in this situation will be those able to serve paying customers, generate cash and finance their commitments through a slower rollout. Establishing which companies meet that standard requires more than comparing their current prices with last month’s highs.

For years, the AI story has encouraged investors to think about an enormous digital future. November 5 brought the physical requirements back into focus: that future still needs buildings, a power connection and permission to build.