The Midterm "Bloodbath": The Best Stocks to Buy Are the Ones That Fell

Wall Street had two very different reactions to the midterms. On Wednesday, the day after Democrats won both the House and the Senate, the Dow had its best post-midterm day since 1982 according to CFRA. On Thursday the Nasdaq fell more than 5% and slid into correction territory, more than 10% below its October record. Chip and data-center stocks took the worst of it. Some called it a bloodbath.

Our view: the selloff is pricing a political threat to AI spending that Congress has little power to deliver, while Wednesday’s winners are pricing policy relief that still has to get past a presidential veto. That gap is the opportunity. The best buys are good businesses the market marked down for the wrong reason. Wednesday’s winners are not on the list.

Two days in numbers (Thursday as of 3:09 p.m. ET)

What happened

Wednesday was a relief rally. President Trump told The New York Times and The Wall Street Journal that he accepted the results and would work with Democrats, and stock futures turned higher on the news. Brent crude fell 8.7%, to below $95 a barrel, and the 10-year Treasury yield eased to 4.88% from 4.95%. Hospital owners and dollar stores led: HCA closed up 10.3%, Universal Health Services 10.9% and Centene 12.6%, and Dollar Tree gained about 9%. Defense stocks fell, with Lockheed Martin down about 7%, on the view that a Democratic Congress could push the president to pull back from the Iran war. Exxon fell about 5% with oil. Tesla, SpaceX and Trump Media dropped on the prospect of congressional investigations.

Thursday was the AI trade unwinding. Tech analyst Dan Ives wrote that Democratic control, especially of the Senate, creates “major hurdles” for AI and tech legislation, and he said 800 to 1,000 planned US data centers could stall over two years. In Texas, the Democrat Gina Hinojosa won the governorship in an upset. Wall Street had expected Governor Greg Abbott to lift his moratorium on new data-center construction after winning re-election. Now the moratorium is expected to stay. Memory makers fell hardest: Sandisk dropped 19% and Micron 15%. AMD lost 15%, Broadcom 9% and TSMC 9.5%. By mid-afternoon the selling had spread well beyond tech: the Dow was down 3.3% and the small-cap Russell 2000 about 4.5%.

How much of this is policy?

Less than the price action implies, in both directions. The new Congress takes over in January, and President Trump keeps his veto. Anything Democrats pass on AI, drug prices or energy credits still needs his signature or a two-thirds override. Both are unlikely on contested issues. In practice a split government governs through hearings, investigations and the budget, and the next budget fight, over December’s funding deadline, comes before the new Congress is even seated.

That cuts against both of the week’s big trades. A House committee cannot stop a private data center from being built. Most of the permitting power sits with states, counties and local utilities, which is why the Texas result matters more for data centers than the Senate does. And it limits the healthcare relief too. Restoring the Medicaid and food-assistance cuts in the One Big Beautiful Bill Act would need a law the president signs. As 24/7 Wall St. put it on Thursday, hospital stocks are pricing in relief “that needs Congress, a presidential signature, and a December funding fight to survive.”

History favors patience. By most counts, the S&P 500 has been higher 12 months after every midterm election since 1950, with an average gain of roughly 15–18% depending on the start date. That is about 19 cases, and a pattern is not a promise. But it argues for buying weakness rather than selling into it.

Our three buys

The buy list, as of 3:09 p.m. ET, November 5

The P/E uses 2027 consensus EPS. Targets are the consensus average. Many were set before October’s slide, so treat the upside column as a ceiling, not a forecast.

1. TSMC (TSM): the selloff’s best bargain

TSMC makes nearly every advanced AI chip, whoever designs it. It fell 9.5% on Thursday and is now 26% below its October 5 close. The data-center worry is about permits in the United States, and in one state above all. TSMC’s customers build data centers around the world, and the biggest buyers are still spending more. Alphabet guided $195–205B of capex for 2026 and said 2027 will be “significantly” higher. Amazon expects about $220B this year and more in 2027.

Analysts expect earnings per ADR to rise about 65% this year and another 35% in 2027, to roughly $22.23 at today’s exchange rate. At $358, that is 16.1x 2027 earnings for the most important manufacturer in the AI supply chain. TSMC’s own risks are Taiwan, export controls (on Thursday the administration was reported to be at odds with SK Hynix over equipment licenses for its China plant), and any real slowdown in AI orders. The US midterm result is not one of them.

2. Alphabet (GOOGL): the reset already happened

Alphabet did its falling before the election. It dropped 12.6% the day after its Q3 report, on a double miss and heavier spending, and it actually rose through the midterms: it is up 1.4% since Tuesday’s close. At $284 it sits near its 52-week low of $270.70 and 30% below its high.

The business underneath is accelerating where it matters. Google Cloud revenue doubled to $30.7B in Q3, its operating margin rose to 36.2%, and the backlog stands at $561B. At 18.9x 2027 earnings, Alphabet is the cheapest of the big AI spenders except Meta. The risks are real: free cash flow has been negative for two quarters, Search monetization is lagging usage while AI Mode ads roll out, and a Democratic Congress means more antitrust and AI-safety scrutiny. We think 19x earnings already prices most of that in.

3. First Solar (FSLR): the one election winner that is still cheap

Of the stocks the market tied to a Democratic win, First Solar is the only one still trading well below its Street targets. It is down 29% this year and 39% from its 52-week high after cutting its 2026 outlook in October, when customers delayed decisions. The stock rose only 4.6% on Wednesday, and it is at 8.5x 2027 earnings.

The election helps in two ways. A Democratic Congress makes it much harder to repeal the Section 45X manufacturing credits, which provide most of First Solar’s current profit, so a key tail risk is gone. And as William Blair noted on Thursday, Democratic gains at the state level should support utility-scale solar demand, though new bookings will take time. First Solar is the largest US-based panel maker and does not depend on a Chinese supply chain, and it reported US bookings near $0.36 per watt last quarter. The bear case is that demand stays soft into 2027, and that the credits keep carrying a business that would earn far less without them.

If you want the hospital trade, buy it cheaply

Universal Health Services (UHS) rose 10.9% on Wednesday and held the gain on Thursday. At $198, it trades at 8.5x 2027 earnings, against 16.2x for HCA. It is cheap for a reason: Q3 adjusted EPS fell 8% to $5.22 and missed estimates, and analysts expect only 3–4% earnings growth next year. But its behavioral-health business is heavily exposed to Medicaid, so it would gain the most from any rollback of the cuts, and its multiple leaves room for disappointment. Treat it as a small, optional position, not a core holding.

Three we would not chase

Wednesday’s winners against the Street

These are good companies, and the policy logic behind Wednesday’s move is sound. More Medicaid and exchange coverage means fewer uninsured patients for hospitals and more members for Centene, and restored food assistance means more spending at dollar stores. But all three now trade at or above the average analyst target, and the policy changes that would justify going higher still need a law the president signs. Centene has more than doubled from its 52-week low of $31.63. HCA, as 24/7 Wall St. noted, had its biggest one-day jump in four years before a single law changed. If the December funding fight goes badly for Democrats, these are the stocks with the most to give back.

What about the hardest-hit chips?

Micron, Sandisk and AMD fell the most on Thursday, but they had also run up the most. Sandisk is still up more than fourfold from its 52-week low, even after losing 64% from its high. Memory prices are cyclical, and we would wait for the selling to stop before buying. TSMC gives you the same AI demand with less cycle risk and a lower multiple.

What would change our view

What would make us more positive

• Hyperscalers confirm or raise 2027 capex in their Q4 reports (Alphabet reports February 3, 2027). That would show Thursday’s data-center scare was sentiment, not orders.
• A December funding deal that restores some Medicaid funding. That would make the hospital trade more than a hope, and UHS the clearest winner.
• The new Texas governor lifts or narrows the data-center moratorium.

What would make us more cautious

• The hyperscalers cut 2027 capex, or several states follow Texas with data-center moratoriums. Both would hit TSMC directly.
• The president reverses his acceptance of the results, or legal challenges drag on. In the CNBC Fed Survey, 77% of respondents saw at least some market risk from extended uncertainty.
• Oil spikes again if the Iran war escalates, pushing yields back toward 5%. That would pressure every multiple on this list.

---

Prices are from 3:09 p.m. ET on Thursday, November 5, 2026, before the close, and may differ from closing prices. P/E ratios use 2027 consensus EPS (Dollar General’s fiscal year ends January 2028). TSMC’s per-ADR estimate is converted from NT$734.78 at NT$33.05 per dollar. Street targets are consensus averages and may predate October’s decline. This article is general research, not personal investment advice. It does not consider your goals, timeline or tax situation.

Sources: market and company data, consensus estimates and price targets from Maz Vantage’s data feed as of November 5, 2026; CNBC (November 4) on the president’s acceptance of the results and the CNBC Fed Survey; the Associated Press (November 4) on Wednesday’s session, oil and Treasury yields; FXStreet (November 5) on the memory selloff, Dan Ives’s note and the Texas result; 24/7 Wall St. (November 5) on hospital stocks; Barron’s (November 5) on William Blair’s First Solar view; Alphabet’s Q3 2026 report; and published post-midterm return studies from LPL, BNY and Pepperstone.