AI sector outlook darkens as Democrat wins threaten $900B capex cycle

$5 trillion in market value has been erased from 47 AI and chip stocks since October 6, and Wednesday’s premarket session suggests the Democratic midterm sweep is set to extend the damage before the opening bell.

NVIDIA (NASDAQ:NVDA), the stock most exposed to the AI data center build-out thesis, closed Tuesday at $206.69 and was pointing up 1.48% to $209.75 in premarket, per Investing.com data. But infrastructure names with the heaviest policy exposure were moving the other way: CoreWeave (NASDAQ:CRWV) was down 1.1% in premarket to $53.16, and Oracle (NYSE:ORCL) fell 1.4% to $124.47 despite Tuesday’s 5.85% rebound, according to MT Newswires as reported by Fidelity.com. SpaceX, which is not publicly traded, reportedly fell around 3.7% in private markets, MT Newswires reported.

The election result lands on a sector already under severe stress. Democrats made gains across Tuesday’s midterms, including a projected victory in Texas — a state that, alongside Georgia, accounts for roughly a third of all U.S. AI data center investment, according to Trending Topics EU. For investors pricing the uninterrupted rollout of a multi-year infrastructure cycle, that outcome carries meaningful regulatory tail risk.

The Texas problem

"If Texas imposes a data center moratorium like New York, projects could be delayed further," Kim Seok-hwan of Mirae Asset Securities said, as quoted by Trending Topics EU. TD Securities analyst Prashant Newnaha was equally blunt in the same report: "Greater regulatory scrutiny of tech and AI companies and a higher likelihood of political gridlock beckon with a Democratic sweep."

The concern is not abstract. New York’s moratorium precedent showed how state-level environmental and energy-grid reviews can freeze siting approvals for months or years. Texas has historically been the most permissive jurisdiction for large-scale power infrastructure, and a shift in the state’s political leadership could change that calculus for projects already in the planning pipeline.

The stakes are enormous. S&P Global has estimated the top six U.S. hyperscalers — Amazon, Microsoft, Alphabet, Oracle, Meta, and SpaceX, are expected to spend more than $7 trillion on data centers and AI-related capital expenditure between 2025 and 2030, while running negative free cash flow until at least next year. J.P. Morgan projects global AI capex will reach $5.5 trillion by 2030, with $900 billion being spent in 2026 alone, according to Reuters. Any slowdown in permitting, particularly in Texas, would pressure timelines across that entire build-out.

Hedge fund unwind accelerates

The election result is hitting a sector that was already reeling from an institutional deleveraging event of historic proportions. Goldman Sachs prime brokerage data show hedge funds sold global AI stocks for five consecutive sessions in October, reversing every net AI purchase made since the start of 2025, according to 247 Wall St. On the heaviest selling day, long-position reductions outpaced new short bets by roughly 3.5 to 1, indicating funds were de-risking rather than betting outright on further declines.

Morgan Stanley prime brokerage data, as reported by BeInCrypto, show hedge funds are now underweight the broader AI trade for the first time since early 2023. Total stock exposure has been cut by more than 8% since June, the biggest retreat since at least 2017. Goldman warned the unwind looks closer to its start than its end, with funds’ AI holdings still ranking in the top 17% of the past three years, leaving substantial room for further reduction.

The scale of October’s damage illustrates what that unwinding looks like in practice. The Nasdaq Composite fell over 5.8% through October, while the S&P 500 tech sector dropped 6.8% for the month, its worst performance since March 2025, per Reuters. NVIDIA and Alphabet each shed roughly $835 billion in market capitalization in the period, per Investing.com data. NVDA, despite Tuesday’s 2.29% session gain, remains more than 15% below its October 6 closing high of $239.24, according to Investing.com data.

Capex conviction under pressure

The political and positioning headwinds arrive at a moment when earnings from the hyperscalers themselves had already begun to erode confidence. Disappointing guidance from Amazon and Alphabet in late October, combined with an Intel collapse of nearly 29%, provided the initial catalyst for the selloff that gathered pace through the month.

An analyst cited in Reuters coverage under the name Kourkafas framed the systemic risk clearly: "If big tech companies start pulling back on capex, which has supported economic growth, it has boosted earnings across many different areas of the economy, that can be something that may be more impactful in terms of driving a broader risk-off sentiment." The concern is that capex cuts would ripple through suppliers, cloud customers, and the broader market in a way that a simple sector rotation would not.

Among the hyperscalers, Microsoft (NASDAQ:MSFT) and Meta (NASDAQ:META) are pointing higher in premarket, MSFT up 1.38% to $528.17 and META up 2.03% to $633.63, per Investing.com, suggesting investors are differentiating between names perceived as closer to monetizing their AI spend and those more exposed to pure infrastructure buildout risk. Both remain well off their 52-week highs, however, with MSFT’s high at $551.15 and META’s at $779.82.

Among the infrastructure plays, CoreWeave is the name most directly in the crosshairs. The neocloud is off 61% from its 52-week high of $138.25 and, as Reuters noted, has "effectively lost access to viable debt financing" amid the capex concerns that have driven the broader sector selloff. A data center permitting slowdown in Texas would hit operators like CoreWeave before it reaches the hyperscalers that anchor their customer base.

The Invesco QQQ Trust (NASDAQ:QQQ), which carries heavy weighting toward the mega-cap tech names driving the AI trade, is the most direct ETF expression of the combined policy and positioning risk described here, its year-to-date performance has tracked closely with the fortunes of NVDA, MSFT, GOOGL, and AMZN collectively.

NVIDIA’s Q3 fiscal 2027 earnings call, scheduled for November 17, 2026, is the nearest hard catalyst for the sector. Analysts broadly regard the report as the definitive near-term test of whether data center demand is genuinely decelerating or whether October’s selloff overshot fundamentals. Goldman Sachs’s warning that hedge fund AI holdings remain in the top 17% of their three-year range means a weak print or cautious guidance from NVIDIA could trigger another wave of institutional liquidation. A strong beat, by contrast, would be the most credible argument for calling a floor. Beyond that, the 119th Congress convenes on January 3, 2027, and the committee chairmanships Democrats assign in the opening days will signal how aggressively they intend to pursue the AI oversight agenda that TD Securities and Mirae Asset have already flagged as a primary market risk.