Best Semiconductor Stocks to Buy Now (December 7, 2025) – 10 Chip Leaders for the AI Supercycle

Updated December 7, 2025 – informational, not investment advice.

The semiconductor industry is in one of the most powerful upcycles in its history.

The World Semiconductor Trade Statistics (WSTS) group just raised its outlook again: it now expects global chip sales to jump about 22–23% in 2025 to roughly $772 billion, and then grow more than 26% in 2026 to around $975 billion – putting the sector within striking distance of a $1 trillion market by 2026. ^[1] Logic and memory chips – the core of AI and data centers – are forecast to grow by more than 30% in 2026 alone. ^[2]

At the same time, the Semiconductor Industry Association and WSTS highlight broad-based strength across regions, with Asia-Pacific and the Americas leading growth. ^[3] Research from PwC and Omdia suggests that AI could sustain a multi‑year growth cycle, with 2024–2030 fab investments topping $1.5 trillion and AI-optimized devices reviving demand in smartphones and PCs. ^[4]

Chip stocks have responded. Major indexes have recently pushed higher as semiconductor names rebounded on expectations of a 2026 AI infrastructure build‑out and potential rate cuts. ^[5] Yet advisors are increasingly warning clients about the risk of an AI-driven bubble in parts of the market. ^[6]

Against that backdrop, here’s a news‑driven, up‑to‑date look (as of December 7, 2025) at some of the best semiconductor stocks to watch now, plus a couple of ETFs for investors who prefer a diversified approach.

Note: Nothing below is a personalized recommendation. Semiconductor stocks are volatile and can fall sharply. Always consider your own goals, time horizon, and risk tolerance, or speak with a licensed advisor.

1. Nvidia (NVDA): The AI compute king, priced for perfection

If AI is the new digital infrastructure, Nvidia is still the default pick-and-shovel provider.

Nvidia controls well over 90% of the data center GPU market, thanks to its CUDA software ecosystem and tightly integrated systems, according to recent analyst commentary comparing Nvidia and AMD’s AI offerings. ^[7]
For its fiscal 2026 Q3 (ended Oct. 26), Nvidia reported $57 billion in revenue and issued robust Q4 guidance, underscoring relentless demand for its AI chips. ^[8]
Zacks now expects Nvidia’s earnings to grow roughly 55% in fiscal 2026 and 51% in 2027, after a wave of upward estimate revisions in the last month. ^[9]
One detailed bull thesis from IO Fund argues that Nvidia is on track to generate around $186 billion in data center revenue in fiscal 2026 as part of a $500 billion+ medium‑term opportunity, and speculates about a path to a multi‑trillion‑dollar market cap if growth and valuation hold. ^[10]

Why it’s a top semiconductor stock to watch

Moat: CUDA, software tools, and developer mindshare make switching costly for customers.
Demand visibility: Cloud giants and AI startups continue to sign multi‑year deals for Nvidia-based systems.
Optionality: Beyond data centers, Nvidia is pushing into automotive, robotics and digital twins.

Key risks

Valuation: Many of the most aggressive growth scenarios are already “in the price”. A slowdown in AI capex or regulatory shocks could hit the stock hard.
Competition: AMD and custom AI chips from hyperscalers are pushing for share, even if Nvidia is ahead today. ^[11]

2. Advanced Micro Devices (AMD): The fast‑growing challenger in AI GPUs

AMD has become the credible No. 2 in AI compute, and 2025’s news flow has been increasingly bullish.

Recent headlines:

Cloud provider Vultr announced a $1 billion AI cluster in Ohio built around 24,000 AMD Instinct MI355X GPUs in a 50 MW data center, with the site expected online in early 2026. ^[12]
Hewlett Packard Enterprise just adopted AMD’s Helios rack‑scale AI architecture, which can pack up to 72 next‑gen Instinct MI455X GPUs per rack, into systems launching in 2026. The platform emphasizes open, Ethernet-based interconnects as an alternative to Nvidia’s proprietary NVLink. ^[13]
AMD’s own investor day laid out a plan to lead a $1 trillion compute market by 2030, targeting >80% CAGR in data center AI revenue over the next 3–5 years, and aiming for more than 50% server CPU revenue market share with its EPYC lineup. ^[14]
For Q4 2025, AMD guided to about $6.1 billion in revenue, driven primarily by AI data center products. ^[15]

Why it’s a top semiconductor stock to watch

Under‑penetrated, high‑growth niche: AMD’s AI GPU market share is still <10%, but growing quickly as cloud providers seek a second source to Nvidia. ^[16]
Open ecosystem story: Partnerships like Helios with HPE and Ultra Ethernet initiatives make AMD central to the “open AI infrastructure” narrative. ^[17]
Valuation vs. Nvidia: Many investors view AMD as a way to play AI with somewhat less multiple risk than Nvidia (though it’s still far from “cheap”).

Key risks

Execution on software (ROCm and ecosystem tools), supply chain scaling and continued heavy capex from customers all need to go right. If AI spending slows, second‑source suppliers often feel the pain first.

3. Taiwan Semiconductor Manufacturing (TSMC, TSM): The foundry backbone of AI

TSMC is the factory behind much of the AI boom, manufacturing chips for Nvidia, Apple, AMD and dozens more.

Recent developments:

Industry research indicates TSMC’s 2‑nanometer (N2) process will reach mass production in late 2025, using next‑generation gate‑all‑around (GAA) transistors. ^[18]
TrendForce reports that 2nm capacity could hit 40,000 wafers per month by end‑2025, expanding about 50% by mid‑2026 and potentially doubling to 80,000–90,000 wafers per month by late 2026. ^[19]
Analysts expect TSMC’s 2026 capex to approach $50 billion, driven by 2nm expansion and global build‑outs. ^[20]
A recent Yahoo Finance piece notes TSMC shares trading around $278 in November, up about 38% year‑to‑date, with an average one‑year Wall Street price target in the low‑to‑mid $300s. ^[21]
Needham recently reiterated a “buy” rating with a $360 target, while other firms remain broadly positive, even as one outlet downgraded to “hold” after big gains. ^[22]
An Investors Business Daily screen of AI stocks near buy points today flags TSMC as trading close to a breakout zone after consolidating gains from the early‑2025 AI rally. ^[23]

Why it’s a top semiconductor stock to watch

Structural role: As long as advanced AI chips keep shrinking, TSMC’s cutting‑edge nodes remain essential.
Geographic & customer diversification: TSMC is investing heavily in the U.S., including a long‑term plan to spend around $165 billion on Arizona fabs and related infrastructure, according to recent tech stock research. ^[24]
Pricing power: Leading‑edge capacity is scarce; TSMC’s 2nm and beyond should command premium pricing.

Key risks

Geopolitical tensions, capex intensity and potential over‑build are the core concerns. If AI demand normalizes just as massive new capacity comes online, foundry margins can compress.

4. ASML (ASML): The EUV monopoly tied to advanced nodes

If TSMC is the AI factory, ASML is the company that makes the factory’s most critical tools.

ASML expects sales of its extreme ultraviolet (EUV) systems to rise roughly 30% in 2025, driven by demand for advanced logic and memory for AI and high‑performance computing. ^[25]
The firm reported about $6.3 billion in Q3 bookings, beating expectations as AI‑driven megadeals helped offset earlier worries about trade restrictions. ^[26]
Management has been candid that 2026 revenue visibility is less clear, citing softness at customers like Intel and Samsung, even as AI demand remains strong. ^[27]
Recent earnings updates suggest ASML still expects mid‑teens sales growth in 2025 with gross margins around 52%, and an improved outlook for 2026 vs comments earlier in the year. ^[28]

Why it’s a top semiconductor stock to watch

Near‑monopoly position: No EUV, no 2nm. ASML’s technology is critical for Nvidia, TSMC, Samsung and Intel to keep shrinking transistors.
Long‑term AI lever: Every AI chip at advanced nodes pulls EUV and High‑NA systems through the supply chain.

Key risks

The stock is cyclical and sensitive to capex pauses. A big slowdown in foundry or memory capex, or tougher export controls, would hit orders.

5. Broadcom (AVGO): Networking and custom AI chips at hyperscaler scale

Broadcom is becoming one of the biggest winners in custom AI silicon and networking.

Recent AI‑driven catalysts:

In its September earnings update, Broadcom disclosed roughly $10 billion in orders for custom AI chips from a major new customer, widely believed to be a top cloud provider. ^[29]
Analysts at Mizuho recently reiterated an “Outperform” rating and raised their target price, arguing that Broadcom is well‑positioned to benefit from the AI infrastructure boom. ^[30]
A compilation of recent opinions shows UBS, Bank of America, Morgan Stanley, Goldman Sachs, Citigroup and Barclays all rating Broadcom as “buy” or “overweight” as of late November/early December. ^[31]

Why it’s a top semiconductor stock to watch

Multiple AI revenue streams: Custom accelerators, networking (switches, NICs), and traditional infrastructure chips.
Diversified business: Beyond semis, Broadcom has major software assets, which can help smooth the cycle.

Key risks

Customer concentration is real – a shift in spending plans at just one hyperscaler could move the needle. The stock also carries regulatory and integration risks after years of big acquisitions.

6. Marvell Technology (MRVL): A potential “dark horse” of AI chips

Once seen as a laggard in 2025, Marvell has roared back into the AI conversation.

Big recent stories:

Marvell announced a $3.25 billion acquisition of Celestial AI, a startup specializing in photonic fabric technology that uses light for ultra‑fast, energy‑efficient data center communication. ^[32]
Management believes Celestial’s technology could unlock a $10 billion market opportunity and generate up to $1 billion in annualized revenue by fiscal 2029, with Amazon receiving a warrant to buy Marvell shares tied to future purchases. ^[33]
In Q3 fiscal 2026, Marvell delivered $2.07 billion in revenue (up ~37%) and adjusted EPS of $0.76, topping expectations, and guided to further growth next quarter. ^[34]
The company forecasts data center revenue growth of more than 25% next year and at least 10% overall revenue growth in fiscal 2027, driven by AI networking and custom chips. ^[35]
Oppenheimer recently raised its price target to $150, and one Zacks article dubbed Marvell a “potential dark horse of AI chips in 2026.” ^[36]

Why it’s a top semiconductor stock to watch

Sweet spot: AI networking, co‑packaged optics and custom accelerators – all areas with strong secular growth.
Re‑rating potential: After underperforming earlier in the year, good execution plus AI‑centric M&A has reopened the bull case.

Key risks

Integration risk from Celestial AI, high expectations for data center growth, and intense competition from established networking players and in‑house cloud designs.

7. Micron Technology (MU): High‑beta bet on the AI memory supercycle

Micron is one of the highest‑beta ways to bet on AI memory and storage.

Recent developments:

For the fiscal year ended August 2025, Micron reported $37.4 billion in revenue, up about 49% year‑over‑year, powered by higher DRAM/NAND prices and booming demand for high‑bandwidth memory (HBM) in AI servers. ^[37]
Analysts at Mizuho recently lifted their price target on Micron to $270, citing the AI memory “supercycle.” ^[38]
At the same time, the stock has been extremely volatile. Micron shares dropped more than 20% in November as rumors circulated that it would exit its Crucial consumer memory business, raising concerns about strategy and execution. ^[39]
Those rumors have now been confirmed: Micron will shut down its Crucial consumer brand by February 2026to focus resources on HBM and data‑center products. ^[40]
Micron also plans to invest about $9.6 billion in a new HBM fab in Hiroshima, Japan, with production slated around 2028, positioning it to compete with market leader SK hynix in next‑gen HBM4/4E. ^[41]

Why it’s a top semiconductor stock to watch

Direct exposure to AI memory shortages: HBM demand is so strong that some PC and device makers are already warning of higher RAM prices and constraints. ^[42]
Operating leverage: When DRAM/NAND pricing is favorable, Micron’s earnings can swing dramatically higher.

Key risks

This is one of the most cyclical names on the list. Memory markets can go from shortage to oversupply quickly. Micron’s pivot away from consumers may pay off, but it also concentrates risk in data center demand.

8. SK hynix (000660.KS): HBM powerhouse with record profitability

For investors able to buy Korean stocks or ADRs, SK hynix is arguably the purest play on HBM – the “gold rush” memory behind AI GPUs.

Recent highlights:

SK hynix reported record Q3 2025 results, with revenue of 24.45 trillion won, operating profit of 11.38 trillion won (a massive ~47% margin), and net profit of 12.6 trillion won, driven by surging HBM and high‑performance server memory demand. ^[43]
Management says it has already secured full customer demand for its entire DRAM and NAND output for next year, and is accelerating HBM capacity expansion. ^[44]
SK hynix recently showcased 12‑layer HBM4, claiming double the I/O channels of the prior generation and over 40% better power efficiency, with shipments set to begin in Q4 2025. ^[45]
A senior executive told Reuters that SK hynix expects the AI memory market to grow about 30% annually through 2030, with custom HBM alone reaching “tens of billions” of dollars. ^[46]
Shares have roughly tripled in 2025 on the back of insatiable HBM demand. ^[47]
On December 7, 2025, SK hynix was named “Best Financially Managed Semiconductor Company” and “Outstanding Asia‑Pacific Semiconductor Company” at the Global Semiconductor Alliance Awards, highlighting its strong profitability and execution. ^[48]

Why it’s a top semiconductor stock to watch

Category leader: SK hynix is widely viewed as the leader in HBM for AI accelerators, currently supplying a large share of Nvidia’s HBM requirements. ^[49]
Pricing power: Tight supply and long‑term AI build‑out give SK hynix significant leverage in contract negotiations.

Key risks

Exposure to memory cycles, geopolitical factors, and trade policies – some of which are specific to Korea–US–China relations – can add volatility.

9. Qualcomm (QCOM): Edge AI and automotive growth

Not all AI semiconductors live in hyperscale data centers. Qualcomm is one of the best‑positioned names at the “edge” – smartphones, laptops, cars and IoT.

Qualcomm’s Q4 FY 2025 results showed record performance in its QCT segment, with especially strong growth in automotive and IoT, and management highlighted the ramp of AI PCs and on‑device AI features. ^[50]
At Snapdragon Summit 2025, Qualcomm laid out an “AI‑first” product roadmap, emphasizing generative AI that runs locally on phones, PCs, XR headsets and vehicles, rather than solely in the cloud. ^[51]
A recent valuation commentary suggests that in scenarios where edge AI and automotive momentum stay robust, Qualcomm could warrant a fair value near $300 per share, though that’s only one of several modeled outcomes. ^[52]

Why it’s a top semiconductor stock to watch

Diversified AI exposure: Smartphones, PCs, automotive and IoT – all areas where AI is being embedded directly into devices.
Licensing model: Royalty streams from 3G/4G/5G/IP add resilience versus pure hardware names.

Key risks

Smartphone cycles can still be lumpy, Apple’s long‑term modem strategy is a wild card, and Qualcomm must prove it can maintain premium pricing as AI commoditizes at the edge.

10. Texas Instruments (TXN): “Picks and shovels” in analog and power chips

For investors seeking a more measured way into semiconductors, Texas Instruments offers exposure to the analog and power chips that underpin industrial automation, autos and energy transition.

In June 2025, TI announced plans to invest more than $60 billion across seven U.S. fabs, calling it the largest investment in foundational semiconductor manufacturing in U.S. history. ^[53]
Earlier this year, TI gave a better‑than‑expected forecast on improving demand in industrial and automotive markets, sending shares up about 8.5% on the day. ^[54]
More recently, TI’s Q3 2025 earnings showed $4.7 billion in revenue, up 14% year‑over‑year, but management flagged a “slower” semiconductor recovery and weaker demand for some analog chips heading into Q4 amid tariff uncertainty. ^[55]

Why it’s a top semiconductor stock to watch

Less AI hype, more real‑world demand: TI chips go into industrial drives, cars, grid equipment, and “boring” electronics that are harder to displace.
Onshoring beneficiary: Scale U.S. fab investments could align well with CHIPS Act incentives and reshoring trends.

Key risks

Analog is still cyclical, and heavy capex is pressuring free cash flow in the near term. If industrial demand slows or stays weak longer than expected, the stock could underperform more AI‑levered names during this phase of the cycle.

Prefer diversification? Two semiconductor ETFs to consider

If you don’t want to bet on a single winner, semiconductor ETFs bundle many of these names together.

iShares Semiconductor ETF (SOXX)

Holds about 30–35 semiconductor stocks, including AMD, Broadcom, Nvidia, Micron and Applied Materials as its top positions. ^[56]
As of December 5, 2025, SOXX’s NAV was about $309, with a year‑to‑date total return near 43%. ^[57]
A recent Motley Fool article even called SOXX “one of the best semiconductor stocks to hold for the next 10 years”, emphasizing its long‑term average annual return in excess of 20%. ^[58]

VanEck Semiconductor ETF (SMH)

Tracks roughly 25 of the largest U.S.-listed semiconductor companies, with Nvidia (~17%), TSMC (~9–10%), Broadcom, Applied Materials and Micron among its top holdings. ^[59]
Concentrates more heavily in a handful of mega‑caps than SOXX, which can amplify both gains and drawdowns.

Why ETFs can make sense

Instant diversification: One purchase spreads risk across many companies and subsectors (compute, memory, analog, equipment).
Reduced single‑stock risk: You still ride the AI and semiconductor trend without needing to pick exact winners.

Big picture: What the 2026–2030 semiconductor forecast means for investors

Putting it all together:

Structural growth: WSTS now projects chip sales to reach about $975 billion in 2026, with memory and logic leading and the industry on track to top $1 trillion by 2030. ^[60]
Capex supercycle: PwC estimates that 2024–2030 fab investments will exceed $1.5 trillion, largely to support AI, high‑performance computing and AI‑optimized devices. ^[61]
Shifting cycle dynamics: Omdia and other research houses argue that the AI boom could extend the traditionally short semiconductor cycle into a longer multi‑year uptrend – but they also warn that pockets of oversupply and digestion phases are inevitable. ^[62]
AI bubble concerns: Financial advisors are increasingly focused on whether AI‑exposed sectors, including semis, are forming a bubble, even as they acknowledge that clients want (and often demand) exposure. ^[63]

So what can you actually do with that?

Practical ideas (not advice)

Decide your “AI risk budget.” If you want aggressive upside, you might tilt more toward Nvidia, AMD, Micron or Marvell. For more balance, TSMC, ASML, Qualcomm, Texas Instruments and broad ETFs like SOXX/SMHcan smooth the ride.
Use time, not timing. Because semiconductors are volatile, many investors choose dollar‑cost averaging – spreading purchases over months or quarters instead of going all‑in on one date.
Diversify by role in the value chain. Combining compute (NVDA/AMD), foundry & equipment (TSM/ASML), memory (MU/SK hynix) and edge/analog (QCOM/TXN) helps reduce the risk that any single segment suffers an extended downturn.
Have an exit (or “don’t look”) plan. Decide in advance how you’ll respond to a 30–50% drawdown, which is common in this sector even in long‑term bull markets.

Quick FAQ

Are semiconductor stocks still a buy after the 2025 rally?

Fundamentals – especially for AI‑linked names – remain strong, and long‑term forecasts have improved. But valuations are elevated in many cases, and regulators and advisors are openly discussing AI bubble risks. That means upside could be substantial but volatility will likely remain high. ^[64]

Which semiconductor stocks are most directly tied to AI data centers?

Primarily Nvidia, AMD, Broadcom and Marvell on the compute and networking side, plus Micron and SK hynix on HBM memory. TSMC and ASML are critical suppliers that indirectly benefit from all of them.

What if I don’t want to pick individual stocks?

Broad ETFs like SOXX and SMH give you diversified exposure to many of the companies above and have delivered strong returns so far in 2025, though future performance is never guaranteed. ^[65]