Tech Stocks Split After the Midterms as AI Spending and Chip Demand Collide

Stocks rose sharply in New York on November 4. The Dow Jones Industrial Average closed up 3.03 percent at 53,356.12, the S&P 500 rose 1.58 percent to 7,688.62, and the Nasdaq Composite gained 1.44 percent to 25,736.51. Yet inside the indexes, technology stocks moved in opposite directions. Data center operators fell sharply while large technology companies rose. It was the same artificial intelligence story, but the stocks split. Behind the split are three questions. Will data center construction continue, can the money pouring into AI be afforded, and is semiconductor demand real?

The market was frightened first. While votes were being counted overnight, Nasdaq 100 futures fell as much as 2.4 percent and S&P 500 futures fell 1.4 percent, on concern that Democratic control of Congress would bring tougher regulation of AI and data centers. Then President Donald Trump accepted the results, and the market turned. The rally began in the morning session and carried into the afternoon.

The split continued at the open on November 5. Around 9:55 a.m., AMD had fallen 11.7 percent from its previous close of $488.78 to $431.50. Nvidia dropped 4.5 percent, from $200.13 to $191.12. Microsoft gave back part of its 6.7 percent jump from the day before, falling 1.5 percent. The S&P 500 was down more than 1 percent in early trading, and Nasdaq 100 futures had fallen more than 0.5 percent in premarket trading while Dow futures rose more than 0.5 percent.

Qualcomm beat earnings expectations at $2.17 a share but fell more than 10 percent in premarket trading after smartphone chip revenue dropped 26 percent to $5.14 billion and missed estimates. In Asia, the Kospi fell 3.9 percent, with Samsung Electronics and SK Hynix down sharply, according to AJP. For a second straight session, stocks fell even after companies reported results.

New claims for unemployment benefits, released at 8:30 a.m., came in at 189,000, the lowest since 1969 and below the 196,000 economists had expected. The Bank of England raised its benchmark rate by a quarter point to 4.00 percent in an 8 to 1 vote, its first increase since August 2023. The 10-year Treasury yield fell to 4.85 percent, its lowest since September 10. Brent crude, after falling more than 8 percent the day before, rose more than 2 percent to regain the $96 level as reports of tankers on fire in the Strait of Hormuz circulated.

Data Centers Fall While Big Tech Rises

Democrats took both the Senate and the House. The result is divided government in Washington. Democrats do not hold enough Senate seats to overcome a filibuster, the rule that lets a minority stretch out debate to block a bill, which takes 60 votes to end. The president holds a veto, the power to send back a bill Congress has passed. That has led to the view that sweeping policy change is now less likely.

Yet the message from voters was clear. Voters approved 40 ballot measures to block or limit AI infrastructure, SiliconAngle reported. Democrats also won in data center hubs such as Texas and Georgia, Reuters reported. Opposition to data center construction has now been confirmed at the ballot box.

By the November 4 close, data center operators had fallen across the board. CoreWeave dropped 9.9 percent, Hut 8 fell 23 percent, and IREN lost 11 percent. Cipher Digital fell 21 percent, Reuters reported. Solar stocks moved the other way. First Solar rose 4.6 percent and Enphase Energy gained 3.7 percent, reflecting expectations that a Democratic Congress will keep supporting clean energy.

Large technology stocks moved in the opposite direction. Microsoft rose 6.7 percent, Amazon 6.1 percent, Alphabet 3.2 percent, and Meta 1.4 percent. The companies building data centers fell, and the companies using them rose.

There is a calculation behind the paradox. If data center construction slows, Big Tech’s spending burden eases and the value of data centers already built rises. Investors made that calculation. Robert Pavlik, a portfolio manager at Dakota Wealth Management, told Bloomberg that a slowdown in data center construction could improve Big Tech’s finances. Analyst Ben Thompson argued that tighter regulation would make computing infrastructure scarcer, lifting the value of existing facilities.

That does not mean construction stops. Analysts believe it would be hard to block data center construction outright by law, Reuters reported. Instead, more hearings and investigations are expected. Mark Malek, chief investment officer at Siebert Financial, said the AI infrastructure buildout will become more expensive, not disappear. Smaller tax benefits and heavier power and water costs will squeeze returns, he explained.

The $740 Billion Question

The second question is money. Capital spending, the money spent on assets such as factories, equipment, and data centers, by Amazon, Alphabet, Microsoft, and Meta is expected to reach $720 billion to $740 billion in 2026. Including SpaceX, the five companies reach $828 billion, according to Fortune. In the third quarter alone, the four companies spent about $207 billion. Gartner expects worldwide AI spending to reach $2.67 trillion this year, up 49.5 percent.

The problem is that the money is going out faster than it is coming in. According to an analysis by the Financial Times, Google, Meta, and Amazon generated $36.4 billion in free cash flow, the money left after investment is subtracted from earnings, a year ago, but swung to a $45.2 billion deficit in the third quarter of this year. Reuters reported that Microsoft, Alphabet, Amazon, Meta, and Oracle earned $542.9 billion from operations in the first nine months of the year but spent $556.9 billion on capital expenditure, leaving a $14 billion shortfall.

Ways to raise the money are visible. Meta said it plans to sell up to $95 billion in shares to fund new data centers. Oracle filed to issue $20 billion in bonds and shares. Borrowed money and sold shares are paying for AI construction.

S&P Global expects the test to last. Capital spending by the six largest technology companies will exceed $1.3 trillion in 2027, free operating cash flow will stay negative in 2026 and 2027, and recovery will not come until 2029, the firm said.

There is also the opposite argument. Pershing Square, run by Bill Ackman, sees the spending not as a defensive cost but as an offensive investment. Capital spending by the four companies rose from $140 billion in 2023 to $740 billion in 2026, and the market expects it to pass $1 trillion next year, the firm said. Building a 1-gigawatt AI data center costs $40 billion to $50 billion and can be expected to earn a 15 to 25 percent pre-tax return, by its calculation. Pershing Square holds shares of Microsoft, Meta, and Amazon.

The issue is not the size of the spending but when the money comes back. If revenue beats expectations, the spending is justified. If not, spending will shrink. Neither outcome has been confirmed yet.

Strong Chip Demand, Shaky Chip Stocks

The third question is semiconductors. The numbers showing demand are strong. AMD reported third-quarter revenue of $13.377 billion, a record, up 45 percent from a year earlier. Data center chip revenue rose 97 percent to $8.55 billion, accounting for 64 percent of total revenue. Its fourth-quarter revenue outlook of about $17 billion also beat market expectations.

Yet the stocks moved on their own. On October 30, the Philadelphia Semiconductor Index fell 7.2 percent in a single day. Nvidia dropped 4.2 percent and AMD 10.9 percent. Intel plunged 28.6 percent after guiding fourth-quarter revenue of $15 billion to $16 billion, below the roughly $17.1 billion the market expected, and warning of weak personal computer demand.

Even a record quarter did not stop AMD’s slide the next morning. In early trading on November 5, AMD fell 11.7 percent from its previous close to $431.50. The drop over five trading days passed 25 percent. The Philadelphia Semiconductor Index had fallen 3.13 percent on November 4 as well.

The memory market is hotter still. Contract prices for DRAM and NAND memory are expected to rise again in the fourth quarter. TrendForce expects DRAM prices to rise 10 to 15 percent and NAND 15 to 20 percent from the previous quarter. SK Hynix has signed long-term supply contracts with 14 major customers. High-bandwidth memory, or HBM, the fast memory attached to AI chips, is sold out through 2027 and is being contracted at about 2.5 times the price of standard DRAM. Samsung raised its share of HBM revenue from 21 percent in the first quarter to 33 percent in the second, while SK Hynix fell from 58 percent to 50 percent over the same period, Counterpoint said. Micron reported fiscal 2026 revenue of $133 billion, up 256 percent, with a gross margin of 81 percent.

The bottleneck is also clear. Orders for TSMC’s advanced packaging equipment, the process that binds chips and memory together, are backed up for 52 weeks. Even if chips are designed, AI accelerators cannot be produced without the equipment to package them.

Counter-signals are growing at the same time. Demand for consumer memory is weakening, and supply from China is rising. Big Tech’s cash burden raises the question of how long the demand supporting memory prices will last. Memory makers’ earnings and demand are strong, but the signals pointing to a peak have become clearer than before, this week’s industry check found.

The next checkpoint is Nvidia’s earnings on November 17. Nvidia sits at the center of the AI chip market. Until those results arrive, there is no yardstick to judge whether demand is strong or expectations have run ahead of it.

What the Search Box Is Asking

Search shows what people are actually wondering about. On Google Trends, average relative interest in the United States over the past 30 days was 62 for “Nvidia stock” and 45 for “AMD stock.” “AI stocks” was only 9. People search for stock names, not topics. Trend figures are a relative index from 0 to 100, a basis for comparison rather than absolute search volume.

Searches for AMD stock hit 100, the period’s peak, at 4 p.m. on November 3, just after the earnings release. It stayed at 98 on November 4. Searches for Nvidia stayed high but flat all month, dipping on weekends and rising on weekdays. Unlike AMD, which jumps only when an event such as earnings arrives, Nvidia is the stock people keep checking.

The fast-rising searches are on the fear side. “ai stocks crashing” and “ai growth stocks sell off” both received Google’s Breakout label, which means search volume surged more than 5,000 percent. On the semiconductor side, “why are semiconductor stocks down” and “why is semiconductor stocks down today” surged. What people are asking is why stocks fell.

Interest in the term “data center” itself is also striking. Over the past seven days, relative interest in “data center” averaged 33, dwarfing the 0 for “data center stocks.” People are more curious about what is happening with data centers themselves than about data center stocks. That means stories about ballot measures, opposition, and power and water issues are being searched before stock charts.

By region, relative interest in “AI stocks” was highest in Kansas at 100, followed by Washington at 52 and Hawaii at 51. The interest is nationwide, not confined to coastal technology centers. As a relative index, it does not mean absolute scale.

Why did data center stocks fall while Big Tech rose, where is the $740 billion coming from, and is semiconductor demand real? The answers will not arrive all at once. The October jobs report comes on November 6, the consumer price index on November 10, Nvidia earnings on November 17, and the Federal Open Market Committee meeting in December. The new Congress takes office on January 3, 2027. Until then, the market will hold on to these three questions.