ASML HOLDINGS N.V. — EXECUTIVE MEMO

"Managing the Indispensable: Strategic Choices as the Bottleneck of Global AI"

To: Board of Management From: Office of the CEO Date: June 2030

SITUATION REPORT

We have become something none of us anticipated when we joined ASML: the central nervous system of global AI infrastructure. Every major compute cluster, every new foundry, every geopolitical strategic reserve of semiconductor capacity—all constrained by our production rate.

This memo addresses the strategic choices we face as supply constraint transforms from advantage to complexity.

THE CURRENT REALITY

ASML's June 2030 position: - Revenue annualization: €61 billion (up 42% YoY) - EUV scanner orders backlog: €312 billion (5.1 years of revenue visibility) - Operating margins: 35% after strategic investments - Production rate: 42 EUV scanners in 2030 (up from 28 in 2028) - Geopolitical restrictions: Dutch/EU export licensing controls 100% of advanced equipment

This is not a business problem. It is a resource allocation problem of unprecedented scale.

DECISION 1: CAPACITY EXPANSION TEMPO

We face a fundamental choice on manufacturing scale-up. Our current path targets 60-70 EUV scanners annually by 2032, achieved through: - €8 billion in capex investment (2030-2032) - Hiring 2,500+ engineers and technicians - Construction of new fab facilities in Veldhoven and potential satellite locations - Supply chain diversification for 2,000+ critical components

The Risk: Aggressive scaling into uncertain demand. If AI investment cycles, we've built excess capacity that consumes €2-3 billion in annual operating costs.

The Opportunity: Every month of undersupply costs our customers $500M+ in delayed foundry revenues. With pricing power at maximum, we could sustain margins even at lower utilization.

Recommendation: Proceed with full capacity expansion. The asymmetry favors scaling. A one-year delay in hitting 70-scanner production costs the industry an estimated $300B in deferred AI development. Our customers will shoulder this cost burden themselves—through higher foundry pricing—if we cannot supply. Better we capture this value through volume at improved margins.

We should target 80 scanners by 2035, which requires commitment to capex in 2030-2031.

DECISION 2: GEOGRAPHIC HEDGING & GEOPOLITICAL EXPOSURE

The Dutch government controls our destiny more directly than any shareholder. Export licensing on EUV systems is now law. The U.S. has pressured us to restrict sales to specific geographies. Taiwan lobbies our board directly on allocation priorities.

Our customers increasingly demand supply from multiple geographies: - TSMC wants Taiwan-based production capacity (impossible—we cannot operate there) - Samsung wants South Korean-based production (possible but politically fraught) - Intel wants U.S.-based production (feasible with government subsidy)

The Dilemma: Decentralized production offers geopolitical resilience but sacrifices manufacturing excellence. ASML's competitive moat depends on concentrated expertise in Veldhoven. Spreading operations across multiple countries increases cost by 15-25% while reducing quality upside.

Strategic Choice: We should maintain 85-90% of EUV production in the Netherlands and selective European presence, while establishing service centers and refurbishment facilities in Taiwan, U.S., and potentially South Korea. This preserves manufacturing excellence while mitigating single-geography risk.

The U.S. government has indicated CHIPS Act funding for advanced manufacturing. We should selectively accept government co-investment in U.S. assembly and testing capacity for systems, provided manufacturing remains in the Netherlands.

Action: Negotiate with the Dutch government for supply guarantee (committed domestic capex support in exchange for production priority allocation to European fabs).

DECISION 3: CUSTOMER PRIORITY ALLOCATION

This is the hardest decision. We could sell three times our current production to the highest bidder. Instead, we face a customer fairness problem with geopolitical, economic, and moral dimensions.

Current allocation principles: - TSMC: 35% (largest advanced fab operator, foundry for Western AI chips) - Samsung: 18% (geopolitical diversification, advanced node production) - Intel: 15% (domestic production, strategic partner) - SMIC/Chinese customers: 8% (legacy relationships, now restricted by government) - Other (Micron, others): 24%

Within this allocation, which customers get priority when delivering systems?

Scenario A (Status Quo): Continue TSMC-first allocation. This maximizes global AI compute production (TSMC is most efficient converter of ASML capacity to compute output) but concentrates supply in a single geopolitical entity (Taiwan).

Scenario B (Diversification): Reallocate 10% from TSMC to Samsung and Intel to reduce Taiwan concentration. This costs the global AI ecosystem approximately $50-100B in deferred compute capacity but reduces geopolitical risk.

Scenario C (Geopolitical Negotiation): Condition allocation on customer commitment to long-term purchasing agreements and geopolitical neutrality. Customers that demand exclusionary supply or resist government interaction face reduced allocation.

Recommendation: Hybrid approach. Maintain TSMC at ~30% (down from current 35%) while slightly increasing Samsung and Intel shares. Tie additional allocations to long-term purchase commitments (10+ year agreements) that lock in pricing discipline. This reduces our optionality in boom cycles but provides revenue certainty into 2040.

We should also establish a "strategic reserve" of 5-10 scanners annually that we allocate based on geopolitical stability and customer cooperation metrics. This gives us negotiating leverage with customers who demand supply in the next 12-24 months.

DECISION 4: PRICING STRATEGY & MARGIN PRESERVATION

We currently operate with 65-68% gross margins and 35% operating margins—extraordinary by any standard. The question is whether to maintain, expand, or deliberately constrain this pricing power.

The Economic Case for Higher Pricing: Our customers face a binary choice—accept our prices or accept production shutdown. A 10% price increase (~$20M per scanner) would expand gross margins to 72-75% while reducing customer demand by approximately 0-3% (inelastic below a threshold).

The Strategic Case for Price Moderation: Aggressive pricing accelerates customer efforts to: - Develop alternative lithography approaches (faster R&D spending) - Build redundancy into foundry capacity (reduce per-unit ASML dependency) - Lobby governments for subsidies to competing suppliers - Invest in insider competency to reduce equipment dependency

Over a 10-year horizon, maintaining competitor morale and reducing incentive for alternative solutions provides more value than an additional 2-3% margin points today.

Recommendation: Target gross margins of 62-65% through moderate price increases (4-5% annually) coupled with strategic cost reductions. This is still extraordinarily profitable while avoiding the "squeeze too hard and create alternatives" trap that befell OPEC.

We should communicate clearly to customers: "We are partners in building the AI future, not extracting maximum revenue. Our pricing reflects capital investment requirements and fair profit, not exploitation of monopoly position."

This messaging reduces geopolitical pressure and customer desperation to find alternatives.

DECISION 5: RESEARCH INVESTMENT & NEXT-GENERATION TECHNOLOGY

EUV lithography is a scientific achievement. But next-generation approaches are under active development: - High-NA EUV (shorter wavelengths, higher resolution) - Extreme EUV (even shorter wavelengths) - Alternative approaches (quantum lithography, metrology-driven design)

We currently invest 15% of revenue (~€9B) in R&D. This is high by industry standards but appropriate given our technology criticality.

The Challenge: We must simultaneously: 1. Maximize current-generation EUV production 2. Prepare next-generation technology for market transition 3. Maintain manufacturing quality and innovation 4. Absorb geopolitical uncertainty and customer demands

Spreading €9B across all three creates risk that we execute none excellently. We face a resource allocation choice.

Options:

Option A (Defensive R&D): Maintain 12% of revenue to R&D, focused on extending EUV capabilities (High-NA variants, yield improvement). Use freed capital ($5B) for dividends and buybacks.

Option B (Balanced Approach): Maintain 15% R&D investment, allocate 10% to next-generation lithography, 5% to EUV extensions. Modest buybacks but minimal dividends.

Option C (Aggressive R&D): Increase R&D to 18-20% of revenue, with significant investment in post-EUV technologies. Retain all earnings for growth and innovation.

Recommendation: Option B. We cannot afford defensive posture—our monopoly is valuable only if we maintain technology leadership. But we also cannot risk current-generation execution for uncertain next-generation returns.

Commit €3-4B annually to post-EUV technology development (High-NA, next-gen approaches) while maintaining excellence in current EUV production. This requires hiring 1,000+ PhD-level researchers and accepting 5-7 year time horizons for results.

The geopolitical moat gives us breathing room to invest heavily in technology. Use it.

DECISION 6: ORGANIZATIONAL CULTURE & EXECUTION

The most risk to our strategy is internal. We are growing 40%+ YoY while maintaining 35% operating margins. This requires hiring, onboarding, and maintaining culture at scale.

ASML's competitive advantage has rested on concentrated expertise and operational excellence. Rapid scaling threatens both.

Actions: 1. Invest in talent infrastructure: Partner with TU Delft, ETH Zurich, and Caltech to establish ASML-focused engineering programs. Offer 10-year career paths and research funding.

1. Decentralize innovation: Establish R&D hubs in Taiwan, U.S., and potentially Asia to access local talent while maintaining Veldhoven as engineering headquarters.

2. Culture documentation: ASML's operational excellence is tacit knowledge. Systematically document design processes, quality standards, and decision-making frameworks to ensure culture survives scaling.

3. Leadership development: Identify and accelerate 200-300 future leaders over next 5 years. Many of our current senior team will retire by 2035.

THE STRATEGIC NARRATIVE

Our role has changed from vendor to infrastructure provider. This requires us to think differently about: - Success metrics: Not just ASML profitability, but global AI infrastructure capability - Customer relationships: Partnership mentality rather than supplier leverage - Geopolitical engagement: Proactive communication with governments on supply strategy - Long-term value creation: 10-20 year thinking, not quarterly guidance

We are the ASML that built Moore's Law continuation. Now we must build the infrastructure for artificial intelligence economics.

BOARD DECISIONS REQUIRED

1. Approve €10 billion capex commitment for capacity expansion (2030-2032)
2. Authorize supply allocation strategy (hybrid TSMC-focus with diversification)
3. Approve pricing guidance (4-5% annual increases, maintain 62-65% gross margins)
4. Commit to €3-4B annual post-EUV R&D investment
5. Authorize organizational expansion to 15,000+ employees by 2032

Timeline: Board approval required by Q3 2030 to begin 2031 execution.

Our competitive position has never been stronger. The challenge is executing on it without losing the culture and technical excellence that created it.

This memo is confidential and intended for board discussion only.