AMD, Nvidia, Intel Crash: Why Chip Stocks Sold Off on Nov. 5

Chip stocks had their worst trading session in months on Thursday, November 5, 2026 — and it had almost nothing to do with earnings. AMD closed down 17.3%, its sharpest single-day drop in years, while Nvidia, Intel, Broadcom and Arista Networks all fell more than 9%. The trigger, according to multiple market reports, wasn't a disappointing quarter or a surprise product delay. It was Tuesday's U.S. midterm election results, which handed Democrats the governorships of Texas and Georgia on platforms that included curbing the AI data-center boom those chipmakers depend on.

What happened on November 5

Five of the market's most-watched semiconductor and AI-infrastructure names posted double-digit or near-double-digit losses in a single session. The moves were large enough, and broad enough across the sector, that they stood out even in a year that has already seen several sharp AI-stock pullbacks. Unlike the late-October selloff that followed Intel's weak guidance, or the brief dip tied to a Broadcom earnings miss earlier in the year, Thursday's drop hit chipmakers with otherwise-strong recent results (AMD, Nvidia) just as hard as those without upcoming catalysts (Arista, Broadcom).

That pattern — a sector-wide rout untethered from any single company's numbers — is the clearest sign the move was driven by something macro, not fundamental. And the timing lines up almost exactly with the fallout from Tuesday's elections.

The numbers: who fell, and by how much

The table below shows closing-price moves for November 5, 2026, compared with the prior session's close.

All five stocks bounced the next day, Friday, November 6 — AMD alone rebounded 11% — which is typical of a sentiment-driven move rather than a reassessment of any company's underlying business. A one-day round trip like that is harder to explain if the sell-off were about a structural change to a company's actual order book.

This wasn't an earnings story

It's worth being explicit about what didn't happen on November 5, because the date sits right in the middle of this quarter's earnings season. AMD had already reported record Q3 2026 results two sessions earlier, on November 3, after the market closed: $13.4 billion in revenue and Data Center segment revenue up 97% year over year, according to the company's SEC filing. The stock actually rose on November 3 and again on November 4 as investors digested that beat — AMD closed November 4 at $488.78, up from its pre-earnings level. It was only on November 5, two full trading days after the results came out, that the stock reversed hard.

Intel had already delivered its own update the prior week — Intel's Q3 2026 revenue was up 21% on AI-chip demand — and that news was already fully public. None of the five companies that fell hardest on November 5 had new financial news that day. Nvidia doesn't report again until November 17, 2026 — nearly two weeks after this sell-off. Whatever moved these stocks, it wasn't a quarterly number.

The reported trigger: Tuesday's election results

The explanation that multiple outlets converged on points to the November 3, 2026 U.S. midterm elections. Democrats won governorships in several states, including two where the incumbent party had held the office for decades: Texas, where Democrat Gina Hinojosa defeated Republican Greg Abbott, and Georgia, where Democrat Keisha Lance Bottoms won the open seat left by the term-limited Brian Kemp. Both races were run, in part, on platforms that included limiting or pausing new AI data-center construction — a departure from the largely hands-off approach state governments have taken toward the buildout over the past two years.

That matters to chipmakers because AI data centers are the single biggest demand driver behind the current GPU and accelerator cycle. A political environment that makes it harder, slower, or more expensive to permit and power new data centers translates fairly directly into a slower runway for chip demand — even if nothing about current chip orders or contracts has actually changed yet. Markets tend to price in that kind of forward-looking risk well before any specific permit gets denied or any specific bill gets introduced, which is consistent with a sell-off that hit richly-valued AI suppliers broadly rather than any single company with disclosed exposure.

It's important to flag what is and isn't confirmed here. The election results themselves are a matter of public record. The link between those results and Thursday's stock move is the market's own interpretation — reported widely in financial media as the proximate cause — rather than a causal chain anyone can point to in a press release or regulatory filing. No federal agency announced a new data-center rule on November 5, and no company disclosed a change to its guidance that day.

Why data centers became a ballot issue

The underlying friction isn't new, even if its arrival at the ballot box is. Large AI data centers draw heavily on local electricity grids and water supplies, and residents near proposed or existing sites in Texas, Georgia, Virginia and elsewhere have complained for more than a year about rising utility bills, noise, light pollution and relatively few permanent local jobs relative to the tax incentives offered to build them. Those complaints had mostly played out at the city-council and state-legislature level. The midterms nationalized them, at least in a couple of marquee governor's races, by putting candidates who ran against unchecked data-center growth into statewide executive offices with real authority over permitting, zoning and utility regulation.

For investors, the risk isn't that any single data center gets shut down. It's that the regulatory and political tailwind the AI buildout has enjoyed — fast permitting, generous tax incentives, limited local pushback — may be turning into even a modest headwind in a growing number of states. Given how much of chipmakers' near-term revenue growth is tied to hyperscaler capital spending on exactly this kind of infrastructure, even a marginal slowdown in the pace of new construction shows up quickly in how the market prices AI-exposed stocks.

A separate, older fight: AI chip export rules

Running in parallel — but on a much longer and distinct timeline — is Washington's ongoing tightening of export rules for advanced AI chips headed to China. The Bureau of Industry and Security (BIS) finalized a rule, effective January 15, 2026, that changed how the government reviews export license applications for advanced computing chips bound for China and Macau, moving from a near-automatic presumption of denial to case-by-case review under added conditions, including third-party testing. Chips reported to fall into this review band include Nvidia's H200 and AMD's MI325X.

BIS followed with guidance reported on May 31, 2026 clarifying that license requirements also apply to AI-chip buyers headquartered in China — or whose ultimate parent company is — even when those buyers are legally based outside the country, closing what had been described as a loophole. Industry coverage of that guidance named Nvidia's Blackwell and upcoming Rubin chip families and AMD's MI350X as products potentially affected, though neither company has detailed exact shipment impacts publicly.

This export-control track is real, confirmed by the Federal Register rule itself, and relevant background for why "AI chip regulation" is a standing risk factor investors now watch closely. But it is not the same story as November 5's sell-off: it's a months-old, China-specific, federal licensing regime, not a domestic data-center political fight that only became acute after Tuesday's state elections. Readers should treat the two as related but separate pressures on the same group of stocks, not one continuous news event.

What Wall Street is saying

Sell-side reaction has been mixed rather than uniformly bearish. Some analysts have characterized the drop as a sentiment-driven overreaction to a political risk that is unlikely to materially change hyperscaler capital budgets in the near term, noting that most major data-center projects already under construction have permits, power contracts and financing locked in well before any new state administration could meaningfully intervene. Others have been more cautious, pointing out that the Nasdaq Composite had already pulled back roughly 8% from its early-October high heading into the election, meaning AI-adjacent stocks entered this week with less cushion — and more valuation sensitivity to bad news — than they had over the summer.

Neither camp disputes the basic mechanism: political risk around data-center permitting is now something equity analysts are explicitly modeling, in a way they weren't a year ago. That's a meaningful shift in how the sector gets covered, independent of how this particular week's price action nets out.

How this fits the bigger AI-trade story

November 5 didn't happen in isolation. Chip and AI-infrastructure stocks have been unusually volatile for weeks, with earlier pullbacks tied to Intel's cautious Q4 guidance, broader worries about whether hyperscaler AI capital spending is outrunning actual revenue, and periodic flare-ups over export policy. What made this week's move distinct is the source: not a company disclosure, not a Federal Reserve decision, not a new BIS rule, but a set of state election results whose connection to chip demand is indirect and, by its nature, open to interpretation. That's a new kind of catalyst for this sector, and one that's harder for any single company to manage or guide around in a quarterly earnings call.

It's also a reminder that "AI regulation risk" is no longer a single, trackable thing. Investors are now pricing in federal export-control actions, state-level data-center politics, utility and permitting fights, and antitrust or AI-safety legislation as separate, overlapping risk factors — each capable of moving chip stocks on its own, on its own schedule, for its own reasons.

What happens next

There's no single event that will resolve this. A few things to watch:

• Nvidia's Q3 FY2027 report, November 17, 2026. If demand commentary from hyperscaler customers stays strong, it would support the "sentiment overreaction" read on November 5.
• How Texas and Georgia's incoming governors actually act on data centers. Campaign promises and enacted policy — permitting changes, moratoriums, utility-rate interventions — are different things, and the gap between them will take months to show up.
• Whether BIS issues further guidance on the China-headquartered-entity rule. That track has moved roughly every few months since January 2026 and remains a live, separate risk alongside the domestic data-center story.
• Follow-through (or reversal) in the stocks themselves. All five names bounced on November 6; whether that bounce holds or November 5's losses resume will say a lot about how durable the market's fear actually is.

For now, the most defensible reading is that this was a real, sector-wide, sentiment-driven sell-off tied to a specific and identifiable event (Tuesday's election results), layered on top of an already jumpy AI trade — and not a sign that any chipmaker's actual near-term business changed on November 5.

Written byMara Lindqvist

Mara Lindqvist edits the hardware desk. She covers graphics cards, processors, memory and storage, the foundries and chip designers behind them, and what the numbers on a spec sheet mean for people choosing a PC. Specifications in her stories come from manufacturer spec pages and datasheets.

More from Mara LindqvistAll Hardware coverage