# Advanced Packaging Investment Thesis 2026 ## Core Argument I'm investing in the "glue" that makes AI chips possible. [[What Is Advanced Packaging|Advanced Packaging]] is now the bottleneck for AI accelerators, not transistor production. As [[Moore's Law]] slows, performance gains shift from shrinking transistors to packaging architectures. Three reasons this is structural, not hype-driven: 1. [[Chiplets and Heterogeneous Integration]] are the new scaling path. Reticle limits force multiple-die designs that depend on dense interconnects. 2. [[HBM and Package Integration]] is mandatory for AI workloads. You can't deliver bandwidth without advanced packaging. 3. [[Packaging Capacity Bottleneck]] is real and takes years to relieve. Building qualified advanced packaging lines requires 18-24 months plus customer validation. ## Why Value Exists Here The market prices [[OSATs - Outsourced Semiconductor Assembly and Test]] like cyclical industrials (forward PE in 20-30x range) despite exposure to multi-year structural demand. [[Packaging vs Foundry Economics]] means OSATs will never achieve foundry-like margins, but they don't need to. Mid-cycle cash flows at current valuations offer downside protection. Where margin of safety shows up: - **OSATs during downturns:** When consumer electronics weakness pressures the multiple, but AI packaging demand keeps utilization high - **Equipment during order weakness:** When bookings drop and market extrapolates tough conditions, missing the technology adoption curve ## What I'm Buying (Concentrated) Following [[Concentrated Packaging Portfolio]] and [[Bill Ackman]] style conviction: **Core position: [[Amkor - AMKR]]** (45-55% of portfolio) - Direct advanced packaging exposure with customer-backed capacity expansion - U.S. Arizona campus aligns with strategic reshoring - Diversified customer base limits concentration risk - Forward PE ~27-30x prices this like a cyclical, not a structural growth story - Free cash flow positive across cycles with manageable leverage **Optional second: [[ASE Technology - ASX]]** (35-45% if used) - Scale leader with similar characteristics to Amkor - Forward PE ~22-28x, comparable valuation - Only add if I want redundancy, not different exposure - Taiwan operations and global footprint complement Amkor **Small kicker: [[Kulicke & Soffa - KLIC]]** (5-10% max) - [[Advanced Packaging Equipment]] exposure to tool adoption - Net cash ~$472M provides balance sheet support - Extreme [[Equipment Tools Cyclicality]] means this stays small - Value entry only when orders weak but tech milestones advancing **What I'm avoiding:** - [[BESI - Hybrid Bonding Leader]]: Great technology, but EV/EBITDA 50-60x prices perfection. Not value-first. - Intel (INTC): Packaging is a positive but doesn't drive equity outcome. Too many execution dependencies. - Synopsys (SNPS): Exceptional business at forward PE ~50s. Pay-up quality, not mispriced durability. ## The Value Framework Using [[OSAT Value Investing Framework]]: **Normalized earnings matter.** OSATs swing with cycles. I underwrite mid-cycle margins and FCF, not peak or trough. Current multiples assume cyclical weakness while advanced packaging mix rises. **Balance sheet first.** Net leverage must stay manageable. Heavy capex can turn FCF negative in downturns. Strong balance sheets survive to capture the next upcycle. **Utilization is the tell.** If advanced packaging fabs stay loaded despite consumer weakness, that's structural demand winning. I watch quarterly commentary on capacity loading. **Customer concentration risk.** One or two lost programs can crush the equity. Diversification across customers and end markets provides downside protection. **Capex discipline.** OSATs adding capacity without committed customer contracts destroy value. I look for multi-year agreements or joint development programs backing new investments. ## What Has to Be True **For permanent capital impairment:** - Structural collapse in AI packaging demand (not a pause, but trend reversal) - Customers vertically integrate packaging at scale (takes years, requires massive capex) - Pricing power permanently destroyed by overcapacity **None are happening quietly.** I would see these signals in: - Customer announcements and capex guidance - OSAT utilization trends - Industry-wide capacity additions vs demand forecasts ## Risks and Monitoring **Primary risks:** 1. **Overcapacity:** Everyone builds at once, margins collapse. Watch divergence between orders and actual shipments. 2. **Customer concentration:** Loss of key programs. Track customer mix and contract terms. 3. **Technology transition:** Hybrid bonding or 3D integration displaces 2.5D faster than expected. Monitor technology roadmaps. 4. **Cycle timing:** Tools and OSATs can underperform for quarters even if thesis is right. Requires conviction to hold through volatility. **What I watch:** - **OSAT utilization rates** (quarterly calls) - **Advanced packaging revenue mix** trending up - **Customer capacity awards** backing capex - **Equipment bookings** (leading indicator) - **CoWoS and HBM supply tightness** (confirms bottleneck) ## Position Sizing If truly concentrated (Ackman style): - **Single name:** 100% in [[Amkor - AMKR]] - **Two name:** 50-55% AMKR, 35-45% [[ASE Technology - ASX]] - **With kicker:** Reduce OSAT positions by 5-10% to add [[Kulicke & Soffa - KLIC]] I avoid over-diversifying. More than two OSAT positions defeats the concentration thesis. Equipment stays small due to cyclicality. ## Why Now Three catalysts converge: 1. [[AI Capex Super-Cycle]] drives sustained packaging demand through [[The infrastructure layer and AI capex]] 2. [[Packaging Capacity Bottleneck]] keeps utilization high and supports pricing 3. Market still prices OSATs like cyclical industrials despite structural demand shift Valuation offers downside protection if growth pauses while providing upside if packaging mix improves and utilization stays firm. ## Exit Conditions **When to sell:** 1. **Valuation extreme:** Forward PE stretches 50%+ above historical range without clear earnings justification 2. **Thesis broken:** Customers announce vertical integration or demand reverses structurally 3. **Overcapacity confirmed:** Industry-wide utilization drops below 70% and stays there 4. **Better opportunity:** Compelling alternative with clearer margin of safety and similar or better risk/return I don't sell on quarterly noise or temporary weakness. Concentration requires conviction to hold through volatility. ## Summary Starting a position because: 1. **Advanced packaging is the bottleneck** for AI chips, not transistor production 2. **OSAT earnings look mid-cycle, not peak**, creating entry point 3. **Valuations price this like old semiconductor business**, not structural growth This is a value-first, concentrated bet on capacity scarcity with observable downside protection and multi-year structural tailwind. --- **Related frameworks:** [[Advanced Packaging MOC]] | [[OSAT Value Investing Framework]] | [[Concentrated Packaging Portfolio]] | [[Bill Ackman]] | [[Investing Principles]] | [[Moore's Law]] | [[Wright’s Law]] --- #investing #semiconductors #deeptech #kp