Semis growth valuations
Query: US semiconductor stocks with market cap above 10 billion dollars and revenue growth above 20%, compared by forward P/E · data as of 2026-10-06
Screened 15 US-listed semiconductor / semicap stocks (mkt cap >$10B, revenue growth >20% YoY) and ranked them by forward P/E ascending. All 15 names had forward estimates available (no exclusions needed; Intel, ON, STM, etc. dropped out on the growth filter). Table includes growth, both P/E measures, 1M/6M/1Y returns, distance from 52W high and 150D SMA, and momentum potential, plus a valuation-vs-growth heatmap and a takeaway on the ~15x forward-multiple dispersion (Micron ~7x to Arm ~99x) mapped to supply-chain tier.
15 US-listed semiconductor and semicap names above $10B market cap growing revenue >20% YoY, as of October 6, 2026. Sorted by forward P/E (cheapest first). Valuation dispersion is extreme: cheapest names trade under 20x forward earnings while AI-levered compounders clear 40-100x.
High-Growth Semiconductors, Ranked by Forward P/E
| Symbol | Security Name | Industry / Theme | Market Cap | Revenue Growth % | Forward P/E | Trailing P/E | 1M Return | 6M Return | 1Y Return | From 52W High | % above 150D SMA | Momentum Potential | Comments |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MU | Micron Technology, Inc. | Semiconductors | 1.181e+12 | 48.85% | 6.75 | 23.64 | 2.85% | 176.95% | 457.20% | -13.84% | 31.18% | 99.1 | The cheapest name in the group and the single biggest anomaly: a memory upcycle is collapsing its forward multiple even as the AI-driven HBM demand story stays intact. Pricing power in DRAM/NAND is the swing factor — this is a cyclical earnings-power trade, not a multiple trade, so the low forward P/E can compress further or evaporate fast if memory pricing rolls over. |
| NVDA | NVIDIA Corporation | Semiconductors / AI | 5.777e+12 | 65.47% | 18.63 | 36.64 | 3.97% | 34.64% | 27.82% | 0.00% | 16.01% | 100 | Unusual to see the AI bellwether this far down the valuation stack — a forward multiple below the trailing figure implies the market is already capitalizing a steep forward EPS ramp. The key debate is whether hyperscaler capex growth can keep outrunning the Blackwell/Rubin cadence; the break-above signal suggests momentum has re-engaged. |
| AVGO | Broadcom Inc. | Semiconductors / AI / Data Center | 1.794e+12 | 23.87% | 19.24 | 62.63 | 5.20% | 12.91% | 11.89% | -21.96% | -0.53% | 97.7 | Custom AI ASIC/networking demand is re-rating the physical supply chain, but the stock is the notable laggard in trend terms. Its forward multiple sitting far below trailing reflects a big expected EPS step-up; the divergence between that growth outlook and a downtrending chart is the core tension. |
| TSM | Taiwan Semiconductor Manufactur | Semiconductors | 2.501e+12 | 25.03% | 22.14 | 36.13 | 12.75% | 40.37% | 66.81% | -0.72% | 19.05% | 95.6 | Purest way to own leading-edge foundry share — sits in the sweet spot between cheap memory and expensive AI design names. Sitting near its 52-week high while the group digests gains reflects how much of the AI capex is physically landing on its fabs. Geographic concentration remains the standing risk. |
| CRDO | Credo Technology Group Holding | Semiconductors / Data Center | 4.151e+10 | 205.68% | 24.21 | 87.98 | 29.47% | 106.79% | 53.49% | -27.00% | 14.18% | 98.9 | Fastest grower in the screen by a wide margin, driven by active electrical cables and optical connectivity for AI clusters. At 200%+ growth, the multiple is arguably the most remarkable part — it looks modest precisely because estimates are still catching up to hypergrowth. Execution and customer concentration are the key risks. |
| NVMI | Nova Ltd. | Semicap/Equipment | 1.255e+10 | 30.96% | 28.88 | 49.25 | 6.01% | -11.23% | 22.50% | -34.87% | -12.77% | 68.2 | Metrology/inspection supplier levered to leading-edge and HBM capacity. One of only two names here in a downtrend and its only trend tag is a downward channel — a rare instance where a semis-equipment name is being de-rated even as the broader capital-equipment theme leads. Watch for a channel-resolution break before committing. |
| LRCX | Lam Research Corporation | Semicap/Equipment | 4.178e+11 | 23.68% | 28.89 | 57.87 | 8.64% | 49.09% | 130.05% | -22.95% | 12.56% | 96.5 | Deposition/etch leader with direct leverage to every new fab build, including HBM capacity. Sits high in the valuation stack versus peers, but its growth is being pulled by AI-driven capital intensity rather than a general equipment cycle — the durability of that spend is what justifies the premium. |
| RMBS | Rambus, Inc. | Semiconductors | 1.237e+10 | 27.13% | 30.86 | 52.33 | 33.50% | 24.19% | 14.10% | -33.15% | 3.69% | 84.8 | IP/royalty model means incremental revenue drops through at high margin, and its memory-interface and security IP sit squarely in the AI data-path. The stock's drawdown from its high reflects how sensitive a royalty stream is to a single end-market cycle — high torque both ways. |
| MTSI | MACOM Technology Solutions Hold | Semiconductors | 2.678e+10 | 32.58% | 41.35 | 112.04 | 30.38% | 47.97% | 175.23% | -14.40% | 18.79% | 87.6 | RF, optical and now defense/space exposure gives it more diversified end-markets than most peers, which is why it carries a growth premium without single-customer dependency. Its participation in government/space programs adds a defense-budget tailwind that pure commercial semis lack. |
| AMD | Advanced Micro Devices, Inc. | Semiconductors / AI | 1.06e+12 | 34.34% | 41.99 | 165.67 | 35.98% | 193.15% | 294.38% | 0.00% | 50.09% | 99.8 | The clearest 'show-me' story in the group: a trailing multiple that dwarfs forward implies the market is paying today for an MI-series accelerator ramp not yet in the numbers. At the 52-week high with the strongest momentum reading in the screen, sentiment is already leaning bullish — leaving little room for execution slippage. |
| MPWR | Monolithic Power Systems, Inc. | Semiconductors | 7.242e+10 | 26.43% | 42.35 | 89.86 | 20.59% | 24.08% | 61.46% | -12.79% | 8.60% | 85.2 | Power-management content grows with every accelerator and every data-center rack, making it an underappreciated AI-levered name. Uniquely, it's tagged in a fully-extended downward channel despite the group's strength — a technical divergence worth monitoring against a still-solid fundamental story. |
| MRVL | Marvell Technology, Inc. | Semiconductors / Data Center | 2.579e+11 | 42.09% | 45.92 | 98.29 | 28.39% | 162.59% | 233.59% | -9.30% | 45.06% | 97.9 | Custom silicon and optical DSP exposure makes it one of the most direct derivatives of hyperscaler AI networking buildouts. Carries a premium multiple for a reason, but that also makes it highly sensitive to any sign of a pause in data-center orders — a high-beta expression of the AI capex theme. |
| SITM | SiTime Corporation | Semiconductors | 2.13e+10 | 61.16% | 47.64 | — | 16.17% | 100.68% | 135.99% | -21.12% | 22.46% | 79.9 | Precision timing (MEMS oscillators) supplier with AI datacenter timing content, but its operating margin is negative — growth is being bought with investment. A high-torque, early-innings name where the watch item is whether scale turns margins positive before the multiple becomes a problem. |
| ALAB | Astera Labs, Inc. | Semiconductors / Data Center | 6.762e+10 | 115.13% | 60.97 | 192.02 | 25.58% | 227.59% | 94.18% | -19.30% | 43.02% | 99.3 | Connectivity/retimer play that is essentially a pure AI-cluster attach rate. Its growth is remarkable but so is its multiple; the bull case rests on every accelerator generation adding more connectivity content, the risk on hyperscaler in-sourcing or a networking transition. |
| ARM | Arm Holdings plc | Semiconductors / AI | 3.231e+11 | 22.79% | 98.93 | 305.62 | 20.02% | 110.32% | 98.22% | -31.15% | 20.31% | 91.8 | Highest multiple in the screen on the slowest growth metric — the market is paying for the architecture-license annuity across smartphones, autos and increasingly data-center CPUs, not for near-term revenue. That makes it the most duration-sensitive name here: the cleanest beneficiary if rates fall, the most exposed if the 10-year stays elevated. |
Valuation vs. Growth Dispersion
[heatmap: Valuation vs. Growth Dispersion]
Takeaway
Valuation dispersion is the story
Filtering the semiconductor complex for >$10B market cap and >20% revenue growth leaves 15 names — and the forward-P/E spread across them is roughly 15x wide, from Micron near 7x to Arm near 99x. That gap is not a data artifact; it maps cleanly onto where in the AI supply chain each name sits and how much of its growth the market already believes.
Three tiers emerge:
- Deep-value cyclicals (<20x): MU, NVDA, AVGO — memory pricing power, the AI accelerator incumbent, and custom ASIC/networking. Both MU and AVGO trade below their trailing multiples, i.e. the market is already capitalizing a large forward EPS step-up.
- Mid-teens to ~30x (the balanced zone): TSM, CRDO, NVMI, LRCX, RMBS — foundry and capital equipment, where growth is being pulled by physical fab and HBM buildout rather than multiple expansion.
- Premium growth (>40x): MTSI, AMD, MPWR, MRVL, SITM, ALAB, ARM — direct AI-networking attach rates, custom silicon, and one architecture-licensing royalty stream, all priced on multi-year content growth rather than next-quarter earnings.
Read-through: the cheapest names are cheap for cyclical or transition reasons (memory cycle, ASIC ramp timing), while the most expensive are expensive because their growth is structural and multi-year. The most uncomfortable position is the middle-high zone — AMD and MRVL carry premium multiples and premium expectations, so any AI capex wobble hits them hardest. ARM is the purest duration trade in the list: its multiple is a bet on rates as much as on chip content. Notably, all but two names (NVMI, AVGO) sit in uptrends, with AMD and MRVL showing the strongest trend-and-momentum alignment — the group's risk is valuation, not participation.
Run a version of this yourself
Describe the idea in plain English and Relaxfolio returns a researchable shortlist with the charts, fundamentals, and risks attached.
Educational research tool — not investment advice. A screen is a starting point, not a recommendation.