Markets 6 November 2026 8 min read

Nasdaq −5.52%, Chips −12% (5 November 2026): Why the Midterm Result Hit AI Stocks — and Why the Bottleneck Is the Site, Not the Chip

The Nasdaq fell 5.52% on 5 November, its worst day since April 2025, as data-center regulation fears hit chips (SOX −12%). The mechanism, step by step.

Nasdaq −5.52%, Chips −12% (5 November 2026): Why the Midterm Result Hit AI Stocks — and Why the Bottleneck Is the Site, Not the Chip
Photo by Radiotrefoil, CC BY-SA 4.0, via Wikimedia Commons.

Nasdaq −5.52%, Chips −12% (5 November 2026): Why the Midterm Result Hit AI Stocks — and Why the Bottleneck Is the Site, Not the Chip

The Nasdaq Composite fell 5.52% to 24,315.00 on Thursday 5 November, its deepest one-day drop since April 2025, and confirmed a correction more than 10% below its 6 October record. The S&P 500 lost 4.36% and roughly $3 trillion of value, per Reuters. The trigger was not a chip-demand number but the midterm result: voters backed local limits on data-center construction almost everywhere they were asked, and the market re-priced how fast the AI build-out can be physically installed.

Key takeaways
  • The damage: Nasdaq −5.52%, S&P 500 −4.36% to 7,353.73, Dow −3.44% to 51,523.06; semiconductor index down more than 12%, Nvidia down more than 9%, per Reuters.
  • The trigger: 41 local data-center ballot measures across 10 states, development constrained or banned in all but one, per the Christian Science Monitor.
  • The mechanism: chip revenue depends on data-center capacity coming online, and capacity depends on zoning, power and permits — the levers held by the offices that were on the ballot.
  • Not a demand story: two days earlier AMD reported data-center revenue up 97%. The doubt is about pace, not appetite.
  • For how risk sentiment is scoring across the eight majors, see the live meter.

What happened on 5 November

The closing figures and company moves below are from Reuters' market report.

Market Move on 5 November Note
Nasdaq Composite −5.52% to 24,315.00 deepest daily drop since April 2025; correction confirmed
S&P 500 −4.36% to 7,353.73 about $3 trillion of value lost; still +7.4% in 2026
Dow Jones Industrial Average −3.44% to 51,523.06
Philadelphia SE Semiconductor Index more than −12%
Nvidia more than −9%
Qualcomm −9.5% missed estimates for Q4 smartphone-chip sales
IREN / Applied Digital each more than −14% data-center operators
CoreWeave almost −10% AI cloud capacity
GE Vernova / Constellation Energy −7.5% / −2.5% power equipment and generation
Meta / Amazon / Alphabet −4% / −2.2% / −2.2% the builders fell less than the suppliers

The last comparable day was the April 2025 tariff announcement. Even after the fall, the Nasdaq remains up 4.6% for 2026.

Two quotes in the Reuters report frame the two readings of the day. Melissa Brown of SimCorp: "Investors appear to be recognizing that some newly elected members of Congress and governors are looking to slow the pace of AI infrastructure development." Jed Ellerbroek of Argent Capital Management: "This feels like the capitulation phase for the data center capex beneficiary stocks, especially the semiconductors... But I think the magnitude of the moves is outsized." The first is a statement about the mechanism; the second is a view on how far the price went. This post deals only with the first.

What the ballot actually said

Headlines framed the result as a change in which party controls Congress. For data centers, the more specific signal was further down the ballot.

According to the Christian Science Monitor's count, there were 41 referendums, initiatives and referrals on data centers in localities across 10 states, and development was constrained or banned in all but one. The tally includes at least 18 bans — outright, on specific projects, or above a certain size — four moratoriums, and ten measures adding new restrictions through zoning or otherwise. The Monitor describes the opposition as bipartisan, running from the Los Angeles suburbs to the industrial Midwest, and driven by land use, local environmental impact and electricity prices.

The same report notes that investors were particularly focused on Texas, a state it describes as on track to overtake Virginia as the world's largest data-center market, where the governor-elect has called for a statewide moratorium.

Politics as an input, not a verdictWhether more or fewer data centers is good policy is not a market question, and this post takes no view on it. The market question is narrower: who holds the approvals that a data center needs, and did the people holding them change on 3 November? The answer to the second part was yes, in several of the largest markets at once — and that is the channel to prices.

The mechanism: the bottleneck is the site, not the chip

A chip only becomes revenue for its maker when a buyer has a place to put it. That place needs three things, and none of them is made in a chip factory.

Land and zoninglocal councils and ballot measures decide where a facility may be built
→
Power and gridstate utility regulators decide connection terms and who pays for new lines and plants
→
Capacity onlinemegawatts energised determine how many servers can be installed
→
Chip orderspurchases follow installable capacity, so a slower site pipeline pushes orders out

Most of those levers are local and state powers. Zoning is local. Electricity tariffs — including whether very large new users must pay the full cost of the grid upgrades they trigger — are set by state utility commissions. Permitting for power plants and transmission is largely a state matter. Federal legislation matters at the margin, through energy policy, tax credits and oversight, but the approvals that decide whether a particular campus breaks ground this year sit closer to home. That is why a ballot dominated by local measures and governors' races could move a semiconductor index.

The data-center and power debate has been running all year; this site's midterm elections preview set out how Congress reaches stocks, yields and the dollar. What was new on 5 November was evidence that the issue had moved from campaign rhetoric to votes, across many jurisdictions at once.

Why the suppliers fell harder than the builders

The pattern in the table follows from how each company's value relates to the pace of the build-out.

  • Data-center operators such as IREN, Applied Digital and CoreWeave (down 10%–14%+) sell computing capacity built on secured sites and power. If new sites become harder to approve, their growth pipeline is the thing that slows. Existing permitted sites may become more valuable, but the market priced growth first.
  • Chipmakers and power-equipment makers such as Nvidia (−9%+), the semiconductor index (−12%+) and GE Vernova (−7.5%) are valued on years of rising orders. Their revenue depends on how fast budgets grow, not merely whether spending continues — the same reason chips fell hardest in the 22 October sell-off.
  • The builders themselves — Amazon and Alphabet (−2.2% each) — are buyers of that capacity. A slower build-out delays their AI projects but also slows their capital spending, which supports free cash flow. Their other businesses carry most of their value. Meta, whose capex recently exceeded its operating cash flow, fell 4%.

Valuation duration explains the size of the moves. A company whose value rests on cash flows several years out is more sensitive to anything that pushes those cash flows later. A permitting slowdown does exactly that: it does not cancel the expected demand, but it moves it further into the future and adds uncertainty to when it arrives.

Why AMD's beat two days earlier didn't protect the sector

On 3 November AMD reported third-quarter revenue of $13.38 billion, data-center revenue up 97% year on year to $8.55 billion, and guidance for about $17 billion in fourth-quarter revenue, plus or minus 3%, per its filing with the SEC. This site's AMD earnings preview had identified the fourth-quarter guide as the number that would decide the read, and it came in strong.

That is evidence about current demand. The 5 November sell-off was about something an earnings report cannot settle: whether the physical capacity those chips need will be approved and powered on the schedule the market had assumed. Orders already placed reflect sites already secured. The question the election raised concerns the next round of sites — the ones that turn this year's growth into next year's.

What it means for indices and currencies

For index traders, the day was a reminder of concentration. The Nasdaq Composite fell 5.52% against 3.44% for the Dow because the same small group of AI-linked companies dominates the market-value-weighted indices. A shock that lands on that one theme moves the NAS100 and US500 more than an equal-weighted reading of the economy would suggest.

For currencies, this was a risk-sentiment event. Risk sentiment is one of the five factors the meter scores, and a sharp equity drawdown tends to show first in the defensive currencies, the yen and the franc. The dollar is pulled two ways: safe-haven demand on one side, and on the other the chance that a sustained equity fall and slower capital spending soften the US growth outlook and the case for further rate rises. The higher-beta Australian and New Zealand dollars usually feel risk-off days most among the majors.

See how risk sentiment, rates and growth are scoring across USD, EUR, GBP, JPY, CHF, CAD, AUD and NZD right now.Open the live meter →

What would change the picture

Election results are intentions; the price effect depends on what becomes binding. The checkpoints:

Development What it would mean for the build-out
State moratoriums or new permitting rules actually enacted lowers the build rate in those states; a direct hit to installable capacity
Large-load electricity tariffs requiring data centers to pay full grid costs raises the cost per site rather than banning it; slows marginal projects
Developers relocating projects to faster-approving states delay rather than lost demand; capacity arrives later, not never
Largest buyers cutting or deferring capex guidance confirms the slowdown in the suppliers' order books
Buyers holding capex guidance and naming secured sites suggests the pipeline is already locked in for the near term

Two features of the calendar matter. New officeholders take their seats in January, so policy changes come with a lag. And ballot measures are local, so their effect depends on how much planned capacity sits in the jurisdictions that voted. Neither is visible in a single day's price, which is why the next information that counts is company disclosure about where projects are and whether they are moving. More on how the site approaches these stories is on the about page.

Educational macro context only — not investment advice.

Advertisement

Frequently asked

Why did the Nasdaq fall 5.52% on 5 November 2026?
Because investors re-priced how fast the AI data-center build-out can proceed. The Nasdaq Composite fell 5.52% to 24,315.00, its deepest one-day drop since April 2025, and closed more than 10% below its 6 October record, confirming a technical correction, per Reuters. The trigger was the midterm result: Reuters reported that wins in states where opposition to new data centers has mounted fuelled concern that tighter regulation could slow the AI infrastructure build-out. On local ballots the signal was even clearer — the Christian Science Monitor counted 41 data-center measures across 10 states, with development constrained or banned in all but one. Chipmakers took the brunt: the Philadelphia SE Semiconductor index fell more than 12% and Nvidia more than 9%. Qualcomm added a company-specific drag, falling 9.5% on a smartphone-chip forecast miss.
How can an election affect chip stocks?
Through the site, not the chip. A GPU only becomes revenue for a chipmaker when a buyer has somewhere to plug it in: land with zoning approval, a grid connection, and a power price someone has agreed to pay. Nearly all of those permissions sit with local governments, state utility regulators and state permitting bodies — exactly the offices that were on the ballot. If approvals slow, data-center capacity comes online later, chip orders get pushed out, and the growth rate the market had priced for suppliers falls. Nothing about chip demand changed on 5 November; what changed was the market's estimate of how quickly that demand can be physically installed.
Why did data-center stocks fall more than Amazon or Alphabet?
Because exposure to the build-out is concentrated in some companies and diluted in others. Data-center operators IREN and Applied Digital fell more than 14% and CoreWeave almost 10%, per Reuters — their whole business is turning sites and power into rented computing capacity. Amazon and Alphabet each fell 2.2%: they are among the biggest builders, but advertising, retail, cloud software and other businesses carry most of their value, and a slower build-out also means less capital spending, which supports their free cash flow. Power-equipment supplier GE Vernova fell 7.5%. The rule of thumb is that the more a company's value rests on the build-out continuing to accelerate, the harder a threat to its pace lands.
Did AMD's strong earnings not help chip stocks?
They had, two days earlier, and that contrast is the point. On 3 November AMD reported third-quarter revenue of $13.38 billion, with data-center revenue up 97% to $8.55 billion, and guided fourth-quarter revenue to about $17 billion, plus or minus 3%. That is evidence that demand today is strong. The 5 November sell-off was about something an earnings report cannot settle: whether the physical capacity those chips need will be permitted and powered on the schedule the market had assumed. Strong current orders and a doubt about the pace of future installation can both be true at once.
What would change the picture after the 5 November sell-off?
Concrete policy rather than campaign positions. The things to watch are whether state legislatures or utility commissions actually adopt moratoriums, new large-load electricity tariffs or tougher permitting rules once new officeholders take their seats; whether developers report delayed or relocated projects; and whether the largest buyers change their capital-spending guidance. A shift of projects to states with faster approvals would mean a delay rather than lost demand, while binding restrictions in the biggest markets would lower the build-out rate itself. Until there is evidence either way, the market is pricing uncertainty about the pace.
PT
Pip Theory desk

We build the tools we write about. Educational macro context only — never investment advice.

About the desk