Semiconductor Stocks Sink as Oracle Revives AI Bubble Fears — Nvidia, AMD, Micron in the Spotlight on December 11, 2025

U.S. chip stocks came under heavy pressure on Thursday, December 11, 2025, as investors digested a cocktail of AI bubble worries, fresh Federal Reserve policy signals, and new geopolitical twists in the semiconductor trade.

By late morning in New York, the tech‑heavy Nasdaq was down about 1%, even as the Dow Jones Industrial Average hit a new intraday record, with the weakness centered squarely on AI‑exposed technology and semiconductor names. ^[1] The Philadelphia Semiconductor Index (SOX) fell roughly 2%, with Nvidia, AMD, Micron, Broadcom and other chipmakers sliding between about 3% and 4%. ^[2]

At the heart of the sell‑off: Oracle’s earnings and spending plans, which have reignited fears that AI infrastructure spending is racing ahead of its near‑term returns.

Market snapshot: chip stocks lead tech lower

As of around midday Eastern time, major U.S. semiconductor names were broadly lower:

Nvidia (NVDA) hovered near $178, down roughly 3% from the prior close.
Advanced Micro Devices (AMD) traded around $215, off about 2.8%.
Broadcom (AVGO) slipped to about $402, a decline of roughly 2.7%.
Intel (INTC) fell just over 3% to around $39.50.
Qualcomm (QCOM), Taiwan Semiconductor’s U.S. ADR (TSM), and leading chip ETFs like SOXX and SMHwere down around 2% apiece.

Investopedia’s midday market wrap noted that Oracle’s results “dragged AI‑tied shares,” with Nvidia, AMD and Micron Technology all down between roughly 2.5% and 4%, while Broadcom fell ahead of its earnings report. ^[3]

In short: after a powerful multi‑year run, AI chip stocks are once again showing just how volatile the trade can be.

Oracle’s AI spending shock and fresh “bubble” worries

The immediate trigger for Thursday’s rout was Oracle’s quarterly update. The cloud and AI infrastructure heavyweight missed Wall Street’s revenue expectations while dramatically increasing its long‑term capital spending plans.

Key points from today’s coverage:

Oracle lifted its fiscal 2026 capex plan by about $15 billion, taking expected spending to roughly $50 billion, largely for AI data centers linked to its massive deal with OpenAI. ^[4]
At the same time, its revenue growth and cloud contract metrics came in below forecasts, raising doubts about how quickly that spending will translate into profits. ^[5]
Oracle’s stock fell about 13–15%, erasing tens of billions in market value and prompting at least 13 brokerages to cut price targets. ^[6]

Reuters reported that Oracle’s weak guidance and big capex ramp “reignited worries” that AI companies may not be able to monetize their investments fast enough, stoking fears of a bubble reminiscent of the dot‑com era. ^[7]

Those worries didn’t stop at Oracle. The same Reuters piece and other reports noted that Nvidia, AMD, Micron, Broadcom and Arm all fell between 3.1% and 4.2%, dragging the semiconductor index lower. ^[8]

For investors, the message from today’s trading is clear: any sign that AI capex might be too aggressive—or that profits may be slower than hoped—can prompt a swift re‑rating across the chip complex.

Nvidia and AMD: AI leaders caught between China, policy and sentiment

Nvidia: China export green light, new tracking tech… and higher scrutiny

Earlier this week, the U.S. government dramatically shifted its stance on advanced AI chip exports to China:

President Trump announced that the U.S. will allow Nvidia’s H200 AI accelerators to be exported to “approved customers” in China, with Washington taking a 25% cut of related sales. Similar rules are expected to apply to AMD and Intel. ^[9]
Policy experts warn this marks a major change in U.S.–China tech policy, loosening controls that, since 2022, aimed to keep China at a significant disadvantage in AI hardware. ^[10]

At the same time, Nvidia confirmed it has built a location‑verification software layer for its GPUs. The tool uses secure computing features and network latency to estimate where chips are operating, helping prevent smuggling into restricted markets like China. ^[11]

The software will debut on Nvidia’s latest Blackwell chips and may later extend to Hopper and Ampere architectures. ^[12]
The company stresses it cannot remotely shut down GPUs—no “kill switch”—and says telemetry is read‑only, targeting compliance rather than control. ^[13]

This combination—new export permissions plus monitoring tools—could expand Nvidia’s addressable market while also adding regulatory and geopolitical complexity.

Yet none of that spared the stock today. Midday reports put Nvidia down about 3–3.5%, as investors focused more on Oracle‑driven AI bubble fears than on policy wins in China. ^[14]

AMD: collateral damage from Oracle and China policy

AMD has been one of 2025’s strongest large‑cap performers, but it, too, was hit hard today:

A StockStory/Finviz write‑up noted that AMD shares fell around 3.8% in the morning session, as Oracle’s weak results raised questions about enterprise AI spending and a recent U.S. policy shift appeared to benefit Nvidia at AMD’s expense. ^[15]
The article highlighted that by allowing Nvidia to ship certain advanced chips into China again, Washington shrinks the market opening AMD had hoped to exploit with its China‑compliant MI series accelerators. ^[16]

In other words: both Nvidia and AMD are now wrestling with whiplash from policy changes and shifting customer budgets, even as secular AI demand remains robust.

Micron and the memory supercycle: bullish forecasts behind today’s dip

While Micron (MU) traded lower alongside peers today, the bigger story for 2025 has been its extraordinary rally and the emergence of an AI‑driven memory supercycle.

Recent coverage underscores just how central Micron has become to the AI narrative:

Investopedia reports that Micron’s stock has gained more than 200% year‑to‑date, hitting record highs just below $264 per share. ^[17]
Citi, Deutsche Bank and Morgan Stanley have all raised their price targets, with some now looking as high as $300–$325 on expectations of continued strength in high‑bandwidth memory (HBM) and AI‑focused DRAM. ^[18]
Recent investor commentary notes that Micron’s HBM supply for 2025 is effectively sold out, signaling tight capacity and strong pricing power. ^[19]

Fundamentally, Micron is in a very different place than just a few years ago:

Fiscal 2025 revenue reached about $37.4 billion, up nearly 50% year‑over‑year, with record quarters driven by AI‑related demand. ^[20]
The company is exiting its Crucial consumer memory business by early 2026 to redeploy capacity toward enterprise and AI data‑center products such as HBM and advanced NAND. ^[21]

Analysts quoted across multiple outlets argue that today’s pullback looks more like short‑term profit‑taking in a name that has already tripled this year, rather than a fundamental break in the AI memory story. ^[22]

Test equipment and second‑derivative winners: Teradyne’s AI tailwind

Not all semiconductor‑related stocks are pure chip designers. Test equipment makers like Teradyne (TER) live one layer deeper in the supply chain—and recent analysis suggests they may be among the more resilient beneficiaries of AI capex.

A fresh Zacks/Nasdaq report today highlighted that:

Teradyne’s Semiconductor Test revenue rose 7% year‑over‑year and 23% sequentially in Q3 2025, reaching nearly 79% of company sales. ^[23]
Memory test revenue more than doubled sequentially, driven largely by DRAM and HBM devices used in AI workloads. ^[24]
Management expects Q4 2025 revenue of $920 million to $1 billion, again underpinned by robust demand from AI accelerators, networking silicon, and memory suppliers. ^[25]

Teradyne shares have surged over 130% in the last six months, significantly outperforming broader tech indexes, signaling that investors are increasingly looking beyond the obvious AI GPU names for sustainable growth. ^[26]

Not all chips are equal: analog and industrial lag behind

The semiconductor universe is broader than AI accelerators and memory. Analog and industrial chips—used in cars, factories and power systems—have been in a much slower recovery.

Back in October, Texas Instruments (TI) warned of a prolonged slump:

The company issued a downbeat Q4 outlook that deepened worries about a “drawn‑out recovery in the analog chip market” as customers delayed factory and capex spending. ^[27]
TI shares fell about 6% on the day, and analysts suggested that the rest of the analog group—peers like ON Semiconductor, NXP and Analog Devices—would likely see similar softness. ^[28]

Thursday’s move did little to change that picture: while AI‑linked names are swinging wildly on sentiment, cyclical industrial and auto‑exposed chipmakers remain in a slower grind, still digesting inventories and bracing for tariff volatility.

Macro backdrop: Fed cuts rates, but valuations face a reality check

All of this is playing out against a shifting monetary policy backdrop that should, in theory, favor growth and tech stocks:

On December 10, 2025, the Federal Reserve cut its benchmark rate by 25 bps to 3.50%–3.75%, the third consecutive cut since September 2025. ^[29]
However, both the Fed’s dot plot and subsequent commentary point to a likely pause, with policymakers projecting only one additional cut in 2026 and several officials arguing against further easing. ^[30]

Markets initially cheered the move—major indices rallied on Wednesday—but Thursday’s trading suggests valuation worries in AI and big tech now outweigh the tailwind from lower rates, at least in the short term. ^[31]

The bigger picture: structural demand still booming

Despite today’s sell‑off, the fundamental demand story for semiconductors remains very strong.

Fresh research from Omdia released today shows:

Global semiconductor revenue hit $216.3 billion in Q3 2025, up 14.5% quarter‑over‑quarter, marking the first time the industry has ever surpassed $200 billion in a single quarter. ^[32]
At this pace, the market is on track to exceed $800 billion in total revenue for 2025, nearly 20% growth over 2024. ^[33]
Crucially, growth is broadening beyond Nvidia and memory: excluding those segments, the rest of the market is still growing around 9% annually, suggesting a healthier, more diversified upcycle. ^[34]

A separate Omdia “2026 Trends to Watch” excerpt notes that data‑processing chips now represent roughly 46% of total semiconductor revenue, with that segment having doubled in just two years and adding about $200 billion in sales between 2023 and 2025. ^[35]

Taken together, these figures are hardly what you’d expect from a sector on the verge of structural collapse. If anything, they underline that today’s AI‑driven chip cycle is real, even if stock prices periodically run ahead of fundamentals.

What today’s move could mean for semiconductor investors

For investors following chip and semiconductor stocks after December 11’s volatility, a few themes stand out:

AI capex is entering a “show‑me” phase.
Oracle’s results showed that investors are no longer willing to blindly reward huge AI spending without clear profitability timelines. That sentiment can hit AI hardware providers—even if their own fundamentals remain strong. ^[36]
Policy risk is now core to the thesis for Nvidia, AMD and others.
The new U.S. decision to allow H200 exports to China, combined with chip‑tracking requirements and ongoing export controls, makes regulation and geopolitics as important as product roadmaps for high‑end AI chipmakers. ^[37]
Memory and test equipment may provide more diversified exposure.
Micron’s explosive year, along with Teradyne’s surge in semiconductor test revenue, highlights that the AI build‑out lifts multiple parts of the value chain, not just GPU designers. ^[38]
Analog and industrial chips are in a different cycle.
Texas Instruments’ cautious outlook and continued softness in industrial demand show that not every sub‑sector is enjoying an AI boom. This can cut both ways: it adds cyclicality risk but also offers value opportunities if and when industrial demand revives. ^[39]
Valuations remain a double‑edged sword.
Analysts still publish optimistic long‑term forecasts—for example, some 24/7 Wall St. pieces debate whether Nvidia stock can reach $500 by 2030—but recent commentary from prominent voices like Bill Gates warns that not all AI valuations will survive intact. ^[40]

Bottom line

Today’s drop in semiconductor stocks is less a verdict on the long‑term future of chips and more a reminder that AI is moving from a simple “buy anything with AI in the ticker” trade into a more complex, selective phase.

The fundamentals—record industry revenue, tight AI memory supply, and broad‑based demand across data centers and devices—remain powerful.
The risks—bubble concerns, policy whiplash, tariff threats, and rich valuations—are just as real.

For anyone following chip and semiconductor stocks, December 11, 2025 is a textbook example of how macro, policy and single‑company earnings can collide to create sharp moves in one of the market’s most important sectors.

This article is for information and education only and is not investment advice. Before making any investment decisions, consider your own financial situation, time horizon and risk tolerance, and consult a qualified professional if needed.