The Intel Business Growing 50% That Nobody's Talking About

Intel's custom ASIC business accelerated sharply in 2025, rising more than 50% year-over-year and 26% sequentially in Q4 to exit the year at an annualized revenue run rate above $1 billion, driven largely by networking and AI-infrastructure demand; management pegs the total addressable market at roughly $100 billion. Strategic deals (notably with Ericsson and AWS), CEO Lip-Bu Tan's Cadence background, and Intel's ability to bundle design services, x86 IP, and in-house manufacturing/packaging position the company to funnel custom-design customers into its foundry as Intel 18A scales and 14A is targeted by end-2027, though near-term Q4 results and guidance missed expectations and competition from Broadcom and Marvell remains significant.

Analysis

Market structure: Intel's $1B ASIC run rate signals an emergent vertically integrated competitor in a ~$100B custom-ASIC TAM where Broadcom (AVGO) and Marvell (MRVL) currently dominate. Customers with fast time-to-market needs (clouds, telco) benefit from bundled design+IP+fabs; pure-play fabless vendors and independent foundries (TSMC peers) face pressure in pockets where customers prioritize integration over node-leading performance. Expect pricing heterogeneity: premium for turnkey integration but downward pressure on margins where Intel uses fabs to win volume, shifting some client spend away from third-party foundries over 2–5 years.

Risk assessment: Tail risks include manufacturing execution failure (18A/14A yield shortfalls), major customer pullback (AWS/Ericsson concentration), and geopolitics (export controls) that could erase multi-year GTM plans. Immediate risk (days-weeks) is sentiment-driven volatility after earnings; short-term (3–12 months) depends on design-to-manufacturing conversion rates; long-term (2–5 years) hinges on sustained capex and yield trajectory. Hidden dependencies: advanced packaging suppliers, EDA/toolchain integration (Cadence/CDNS exposure), and substrate/material supply — failures here are second-order revenue killers.

Trade implications: Tactical longs: Intel offers skewed asymmetric upside if you believe integration wins share — size positions to 2–4% of risk capital with staged adds tied to process ramp milestones (18A yield targets, 14A availability). Relative ideas: long INTC vs short AVGO (0.5x) to express share gains in integrated design-fab segments over 12–18 months, and small long CDNS (1–2%) to capture EDA demand. Options: construct 12-month call spreads on INTC (buy 25% OTM, sell 60% OTM) to cap cost; buy protection (6–12 month puts) on material longs.

Contrarian angles: Consensus underestimates customer concentration and execution difficulty — $1B run rate is meaningful but small versus TAM and capex needs, so downside is underappreciated if yields slip. Conversely, market may be underpricing Intel's bundling advantage in networking/AI where speed-to-market matters more than absolute node leadership; historical parallels include early foundry-share shifts when integrated players offered turnkey services. Watch for unintended consequence: design wins that remain paper wins (design but no fab capacity) which inflate bookings without durable foundry revenue.

Key Decisions for Investors

• Establish a 2.5% long position in INTC sized to portfolio risk (12–24 month horizon); add another 1.5% only if Intel reports consecutive quarterly improvement in 18A yields or announces multi-customer tape-outs converting to production within 6 months.
• Put on a pair trade: long INTC (2% portfolio) / short AVGO (1% portfolio) for 12–18 months to express view that Intel's turnkey stack can win share in networking/AI ASICs; trim if INTC share of foundry external revenue exceeds $2B ARR or AVGO gross margin expands by >200bps sequentially.
• Buy CDNS (1.5% portfolio) as a hedge on increased EDA/design services demand; accumulate on any >8% pullback or after the next quarterly beat in design-tools revenue.
• Use options to define risk: buy a 12-month INTC call spread (long 25% OTM, short 60% OTM) sized to target ~3x upside vs cash outlay, and buy 6–12 month protective puts equal to 30% of your INTC position not covered by the spread if Intel misses two consecutive process-yield milestones by end-2027.