Semiconductors

A broader theme than AI infrastructure: every public company meaningfully exposed to semiconductor manufacturing and design, from foundries to fabless designers to the materials and equipment that feed them. Cyclical, but structurally growing as compute demand expands.

What is the semiconductor supply chain?

The semiconductor supply chain is a relay of specialists, and the semiconductors theme is built to span all of it rather than betting on one link. Fabless designers like NVDA, AMD, AVGO, QCOM-class peers, and MRVL define what a chip does but own no factories. Foundries like TSM turn those designs into physical silicon. Equipment makers like ASML and AMAT build the machines the foundries run, while materials and consumables suppliers like ENTG, MKSI, MTRN, ROG, and ESI feed the fabs the chemistries, metals, and process sub-systems they consume on every wafer.

Memory and packaging close the loop. MU supplies the DRAM, NAND, and high-bandwidth memory that sits beside AI accelerators, RMBS supplies the interface chips that move that memory, and AMKR handles the packaging and test that turns finished die into usable parts. Because the semiconductors theme holds names across design, foundry, equipment, memory, materials, and packaging, it tracks the full chip economy instead of a single segment.

How do semiconductor companies make money?

Semiconductor companies monetize different rungs of the same ladder, which is why the semiconductors theme mixes business models on purpose. Fabless designers such as NVDA, AMD, and AVGO earn high gross margins by selling proprietary architectures and software, while outsourcing the costly manufacturing. TSM earns its return on capital by running the most advanced fabs in the world at high utilization, charging for leading-edge capacity that almost no one else can supply.

Further up the chain, ASML and AMAT sell capital equipment and the service contracts that follow it, so their revenue tracks fab build-outs rather than end-device sales. Analog and embedded specialists like TXN sell huge catalogs of long-lived parts into industrial and automotive markets, and IP licensors like ATOM and RMBS collect royalties on technology that ships inside other companies' chips. Holding the semiconductors theme exposes you to all of these revenue engines at once.

Why are semiconductors cyclical?

Semiconductors are cyclical because chip demand, inventory, and fab capacity rarely move in step. When end markets like phones, PCs, autos, and servers overstock, orders contract across the whole chain at once; when they understock, the next demand surge overwhelms lead times. Memory names like MU feel this first and hardest, since DRAM and NAND pricing swings sharply, while analog and embedded names like TXN tend to lag the turn by a quarter or two.

The semiconductors theme spreads exposure across these different cyclical rhythms. Equipment makers like ASML and AMAT ride multi-year fab investment cycles that can run on a different clock from the design and memory segments, and materials suppliers like ENTG and MKSI track wafer volumes more than chip prices. Owning a basket across the stack does not remove the cycle, but it keeps the semiconductors theme from depending on any single segment's timing.

What gets a stock into the Semiconductors theme?

Majority of revenue derived from designing, manufacturing, packaging, or supplying the semiconductor industry. Includes both AI-specific names and broader analog, memory, and specialty silicon.

Semiconductors stocks

Every public name that fits the Semiconductors thesis, with the rationale for inclusion. Click any ticker for the full stock guide. The basket above starts equal-weighted; you set your own target weights inside Walnut.

How to invest in Semiconductors

There are a few common ways to get exposure to semiconductors. You can buy individual stocks across the stack: a designer like NVDA, AMD, or AVGO, the leading foundry TSM, equipment leaders ASML and AMAT, memory through MU, and specialty materials or packaging names like ENTG, MTRN, or AMKR. Picking single names lets you weight the exact links of the semiconductors supply chain you find most compelling, at the cost of more research and more concentration risk in any one company. The ETF route is simpler: SMH (VanEck Semiconductor) and SOXX (iShares Semiconductor) both package the semiconductors industry into one ticker. The tradeoff is concentration versus breadth. SMH is more top-heavy, letting the largest names like NVDA and TSM drive most of the return, while SOXX caps individual weights more aggressively for broader coverage of the semiconductors theme.

The third approach is building a semiconductors basket in Walnut. You describe the exposure you want, for example a global semiconductor supply chain spanning designers, foundry, equipment, memory, and materials, and Walnut's AI assistant proposes constituents and target weights drawn from the semiconductors theme. You fund the basket through your own connected broker, and you approve every order before it is placed. Walnut never trades for you. From there the semiconductors basket tracks as a single performance line you can compare cleanly against SMH, SOXX, or the broader market, so you can see how your specific slice of the semiconductors theme is doing over time.

ETFs used as passive proxies for Semiconductors

If you want the theme as a single ticker rather than as a basket, these are the ETFs people most commonly use. Each has trade-offs (concentration, expense ratio, sector overlap) covered in the individual ETF guides.

FAQ

What is the semiconductors theme?

+

Semiconductors covers every public company whose revenue materially depends on designing, manufacturing, packaging, or supplying the chip industry. That's fabless designers (NVDA, AMD, AVGO, QCOM, MRVL), foundries (TSM), equipment makers (ASML, AMAT, LRCX), memory (MU), analog and embedded (TXN), specialty consumables (ENTG, MKSI), advanced packaging (AMKR), and IP licensing (RMBS, ATOM, IDCC). Broader than AI infrastructure on purpose: the theme captures the full cycle, not just the AI window.

How many semiconductor stocks does Walnut cover?

+

18 names as of early 2026: NVDA, AMD, AVGO, TSM, MU, ASML, AMAT, MRVL, TXN, ENTG, AMKR, MKSI, RMBS, ATOM, MTRN, AXTI, ROG, ESI. The list spans designers, foundries, equipment, materials, packaging, and IP.

What's the biggest semiconductor stock?

+

NVIDIA (NVDA) by market cap, having passed Apple and Microsoft into 2025-2026. Before NVIDIA's run, TSMC was the largest by revenue; it still is on most operating metrics. Broadcom (AVGO) has been the third major contender thanks to its custom AI silicon business plus VMware acquisition.

What ETFs cover semiconductors?

+

SMH (VanEck Semiconductor) and SOXX (iShares Semiconductor) are the two standards. SMH is more top-heavy (NVDA and TSM combined are 25-30% of the fund), SOXX is more diversified. Both charge 0.35%. VGT and XLK include semiconductors as a subset of broader tech sector exposure at 0.09% expense ratios.

SMH vs SOXX: which is better?

+

Functionally similar exposure, different concentration. SMH lets the biggest names run; if you believe in the NVDA-and-TSMC-led top of the cycle, SMH gives you more of it. SOXX caps individual weights more aggressively; if you want broader diversification within semis (including equipment and analog), SOXX delivers it. Returns over multi-year windows have been close; SMH has typically outperformed in the strongest AI-driven legs because of concentration.

Is the semiconductor industry cyclical?

+

Famously yes. Industry revenue has historically swung 20-40% peak-to-trough, and individual segments (memory especially) move 50%+. The current AI cycle has compressed some of that volatility because AI training capex has been more durable than past handset or PC cycles, but pretending the cycle has been repealed has been wrong every prior time. Analog (TXN, ADI) typically lags the cycle by 1-2 quarters; memory leads it.

How do I invest in semiconductors?

+

Three common approaches. (1) Buy SMH or SOXX for diversified passive exposure. (2) Buy NVDA + TSM + AVGO directly for concentrated leaders. (3) Build a Walnut basket spanning the supply chain (designers + foundry + equipment + packaging + materials), which most users do when they want differentiated exposure to specific links rather than market-cap-weighted concentration.

What drives semiconductor stocks?

+

Three structural drivers in 2026: AI training and inference capex (NVDA, AVGO, TSM), end-market demand cycles (smartphone, PC, automotive, industrial), and fab capacity expansion (drives equipment makers AMAT, ASML, LRCX). The cyclical overlay is inventory: when end customers overstock, revenue contracts across the chain; when they understock, the next demand surge buries lead times.

What's the difference between semiconductors and AI infrastructure?

+

Semiconductors is broader. It includes analog (TXN), specialty materials (MTRN), and IP licensing (IDCC, RMBS) that aren't AI-specific. AI infrastructure is tighter on names whose revenue meaningfully tracks AI training capex plus the hyperscalers buying that capex. NVDA, AVGO, TSM are in both themes. ASML is in both because every leading-edge AI chip is patterned on its EUV machines. TXN, ATOM, AXTI are in semis only.

Why is TSMC so important?

+

TSMC manufactures essentially every leading-edge logic chip: NVIDIA H100/B100, AMD MI300X, Apple Silicon, every hyperscaler custom silicon design. At the 5nm/3nm/2nm nodes, TSMC has 95%+ market share. Its capacity scheduling, yield ramps, and Arizona fab buildout are inputs to almost every other AI infrastructure investment thesis. Geopolitical risk (Taiwan Strait) is the structural tail risk every long semi thesis carries.

What is ASML's monopoly?

+

ASML is the only company in the world that makes extreme ultraviolet (EUV) lithography systems, the machines required to pattern transistors at 7nm and below. Every leading-edge fab (TSMC, Samsung, Intel, Micron HBM) buys ASML EUV systems. The technology took 20+ years to develop; competitive replication isn't a near-term risk. The risk to ASML is delayed capex from its customers, not new competition.

Are semiconductors a good investment right now?

+

Walnut isn't an investment adviser. Factually, semi industry revenue is recovering off the 2023 inventory trough, AI-driven capex has continued to scale, and equipment companies are guiding to multi-year backlog. Valuations on the most-loved names have rerated significantly; the cycle is closer to mid-to-late than early. Time horizon and what else you own matter more than any single moment in the cycle.

Can I build a semiconductor basket in Walnut?

+

Yes. Tell Walnut's AI assistant something like 'global semiconductor supply chain' and it proposes a 5-8 stock basket spanning designers, foundry, equipment, memory, and materials. You set the weights and the basket tracks as one performance line you can compare cleanly to SMH or SOXX.

Build the Semiconductors basket in Walnut

Walnut's AI assistant takes the thesis above, proposes 5 to 6 constituents with target weights, and lets you fund the basket through your existing broker. You approve every order; we never trade on your behalf.

Other themes

  • AI infrastructure. Picks and shovels of the AI buildout: GPUs, networking, foundries, and the software platforms training the largest models.
  • Data center power and cooling. The grid, switchgear, liquid cooling, and electrical contracting that AI data centers can't run without.
  • Defense and modernization. Software, sensors, and specialty materials at the center of US and allied defense buildouts.
  • Critical materials. Rare earths, specialty metals, and strategic materials at the center of supply chain reshoring.
  • Dividend growth. Companies that compound a growing dividend through cycles. The boring core of many long-term portfolios.

Walnut is informational, not investment advice. Theme membership is descriptive, not prescriptive; nothing on this page should be read as a recommendation. Always verify current financials and your own circumstances before investing.

    Semiconductors: How to Invest, Stocks & ETFs (2026), Walnut