Public Christophe Fouquet Global market trends Industry & ASML's technology roadmap ESG ASML Investor Day Veldhoven, The Netherlands November 14, 2024 President and Chief Executive Officer Small Talk 2024 Exhibit 99.2
Public • The Semiconductor Industry remains strong and Artificial Intelligence is expected to create further opportunity as major investments in supercomputing are happening and the entire industry is preparing to insert AI in all critical future applications • Our industry will require major innovations to address the anticipated cost and power consumption challenges of AI, and this will further boost the industry roadmap in a product mix shifting towards advanced logic and DRAM • Our customers remain at the core of our strategy, and we believe that lithography will remain at the heart of their innovation. We also anticipate that an increased number of critical lithography exposures for advanced logic and memory processes will continue to support our customers in addressing their challenges • We expect that our ability to 1- scale our EUV technology well into the next decade, 2- extend holistic lithography into supporting 3D front end integration and 3- improve the performance and cost effectiveness of our DUV products will continue to address all our customers’ needs with a flexible and versatile portfolio • We will continue to leverage our large and growing systems installed base (DUV, EUV) to provide high value service and upgrades over a >20 years lifetime • ASML values the strong industry partnerships which are critical to our success and our collective commitment to a leadership position in ESG We expect that our ability to scale EUV technology into the next decade and extend our versatile holistic lithography portfolio could place ASML at the heart of the AI opportunity. This would continue to create significant growth in revenue and profitability in this decade. Page 2November 14, 2024
Public A Customer trust and partnership remains at the core of ASML’s strategy Seon-Yong Cha, SK Hynix CTO We expect that our ability to scale EUV technology into the next decade and extend our versatile holistic lithography portfolio could place ASML at the heart of the AI opportunity. This would continue to create significant growth in revenue and profitability in this decade. Page 3November 14, 2024
Public We see our society going from chips everywhere to AI chips everywhere Gen AI opens endless opportunities, expected to add 6-13T$ value to GDP by 2030 Connected world Climate change and resource scarcity Social and economic shifts Automation Healthcare, medical tech Technological sovereignty Working, learning remotely Cloud infrastructureHyperconnectivity Edge computing Energy transition Electrification, smart mobility Agricultural innovation Smarter use of limited resources Page 4*Source: McKinsey and Company, GenAI: The next S-curve for the semiconductor industry? | March 2024 Internet of Things November 14, 2024
Public Internet of Things Cloud infrastructure Edge computing Energy transition Automation Healthcare, medical tech Technological sovereignty Working, learning remotely Hyperconnectivity Electrification, smart mobility Agricultural innovation Smarter use of limited resources AI has strong potential to drive entire industry forward across many applications Mainstream markets grow in volume while high performance continues to follow Moore's law I-Line Advanced segments Mainstream segments Advanced segments Mainstream segments 300 mm 300 mm 300 mm 300 mm 300 mm 200 mm 300 mm 300 mm 200 mm 150 mm 300 mm ArFi KrF ArF DRAM NAND MPU ANALOG POWER OPTICAL SENSORS NON - OPTICAL SENSORS LOGIC Mainstream LOGICSPECIALTY MEMORY ADVANCED PACKAGING EUV US TW CN EU JP USKR IRL November 14, 2024 Page 5
Public Cloud infrastructure Edge computing Energy transition Automation Healthcare, medical tech Technological sovereignty Working, learning remotely Hyperconnectivity Electrification, smart mobility Agricultural innovation Smarter use of limited resources We reconfirm our view calling for global semi sales >$1T by 2030 Major investments are on-going in AI, the exact pace of its roll-out to consumer products is still unknown >1T$ 2030 Semi Sales 2015 100 200 300 400 500 600 700 800 900 1000 2016 2017 2018 2019 2020 2021 2022 2023 2024 S e m i S a le s [ B $ ] 0 High performance compute 40% Other IoT Automotive Mobile Data AI power compute (+memory) Page 6 *Source: Kevin Zhang, TSMC, Semiconductor Industry: Present and Future, IEEE solid state Circuit Conference, Feb 2024 *Source: WSTS and Gartner Global data center and edge AI demands are expected to grow significantly year on year to represent >40% of semi sales demand in 2030 and generate some upside for our industry Connecting IoT Sensors Mainstream edge compute AI ? *Source: McKinsey and Company, GenAI: The next S-curve for the semiconductor industry? | March 2024November 14, 2024
Public Page 7 Semi Equipment Hardware ASML Peers Semi Non-Semi Loss in 2023 The semiconductor ecosystem has considerable means to drive innovation Despite a market downturn, the ecosystem generated over $865B EBIT in 2023 Logic IDM Foundry Wafers Distributor E D A A u to m o ti v e PC S m a rt p h o n e C o n tr a c t M a n u fa c tu ri n g S e rv e r In d u s tr ia l C o n s u m e r Products Foundry Semi Peers Non-Semi Semi manufacturing Semi design Software and services Source: Corporate Marketing (CMKT) analysis; Company reports; Note: EBIT = Earnings before Interest & Taxes. Box size scales to EBITNovember 14, 2024
Public The semiconductor ecosystem has reinvested around half of its EBIT to drive long-term innovation and growth, and we expect this to continue Page 8 2015 2016 2017 E B IT [ B $ ] 2018 437R & D [ B $ ] 2019 2020 2021 2022 900 800 700 600 500 400 300 200 100 0 2023 449 540 616 564 616 883 817 865 197 220 252 289 318 348 404 468 504 Source: Corporate Marketing (CMKT) analysis; Company reports; Note: EBIT = Earnings before Interest & Taxes. 1046 2024e 541 1000 November 14, 2024
Public Moore’s law for computing power is alive and well Transistors per package continue to double every two years enabling one trillion by 2030 Source: nn Kelleher, ntel, “Moore’s law – now and in the future”, ntel investor day Feb 17, 2022 T ra n s is to rs P e r P a c k a g e ( lo g ) Aspiring to one trillion transistors by 2030 intel internal analysis of Intel products. Future projections based on products still in design. Future transistor counts are projections and are inherently uncertain. 2005 2010 2015 2020 2025 2030200019951990 Page 9 2x/2yrs November 14, 2024
Public High Performance Compute Demand Beyond Moore’s Law We expect that Generative AI will demand an acceleration of Moore’s law Energy use could set a ceiling on training capability moving forward if not addressed Page 10 C o m p u te p e r p a c k a g e [ F L O P s ] 2x/2yrs E n e rg y u s e p e r p a c k a g e [ J ] -0.6x/2yrs 2005 2010 2015 2020 2025 2030200019951990 Compute needs for AI Source: High Performance Computing: https://top500.org/lists/top500/ based on FP64 performance; Consumer GPU: nVidia & AMD data sheets based on FP32 performance Source: EPOCH , “ Notable models” June 2024, updated Nov 2024 Energy Required for AI November 14, 2024
Public High Performance Compute Generative AI, high-performance computing power outpaces Moore’s law In 2030, >70% of data center demand could be driven by AI, representing >90% of Gen AI FLOPs Page 11 C o m p u te p e r p a c k a g e [ F L O P s ] 2x/2yrs E n e rg y u s e p e r p a c k a g e [ J ] -0.6x/2yrs 2005 2010 2015 2020 2025 2030200019951990 November 14, 2024 Source: High Performance Computing: https://top500.org/lists/top500/ based on FP64 performance; Consumer GPU: nVidia & AMD data sheets based on FP32 performance Source: EPOCH , “ Notable models” June 2024, updated Nov 2024
Public Generative AI, high-performance computing power outpaces Moore’s law Supercomputer architecture has the potential to meet the needs of AI Page 12 High Performance Compute C o m p u te [ F lo p s ] 2x/2yrs Moore’s law E n e rg y u s e [ J ] (Energy per flop)-0.4x/2yrs Energy per flop + Compute Eff. -0.6x/2yrs 3x/2yrs Compute efficiency 2005 2010 2015 2020 2025 2030200019951990 Argonne Leadership Computing Facility: https://www.alcf.anl.gov/aurora High-performance computing Aurora supercomputer (HPC and AI applications) delivers 2,000,000,000,000,000,000 FLOPs November 14, 2024
Public Generative AI, high-performance computing power outpaces Moore’s law Architecture cost must be reduced to fully enable AI opportunity Aurora supercomputer (HPC and AI applications) delivers 2,000,000,000,000,000,000 FLOPs using ~85K CPUs/GPUS, >20PB memory for computing and >230PB memory for storage at ~500 M$ cost Cost: >500M$ Page 13Argonne Leadership Computing Facility: https://www.alcf.anl.gov/auroraNovember 14, 2024
Public Generative AI, high-performance computing power outpaces Moore’s law HPC rapidly growing energy consumption must be addressed to prevent a ceiling on training capability 1952 1960 1968 2008 2016 1023 1021 1019 1017 1015 1013 1011 109 107 105 103 101 T ra in in g c o m p u te ( F L O P s ) D e e p l e a rn in g e ra in c lu d in g f e a tu re e x tr a c ti o n L a rg e s c a le e ra c o m b in in g c o m p le x d a ta s e ts Argonne Leadership Computing Facility: https://www.alcf.anl.gov/aurora. | Source : Lisa Su, AMD, ITF May 21 2024 1025 2022 G e n e ra ti v e A I Aurora supercomputer (HPC and AI applications) delivers 2,000,000,000,000,000,000 FLOPs using 60MW and 34,0 0 gallons of water per minute for cooling Energy consumption: >60MW Page 14November 14, 2024
Public Scalability of High-performance computing is endless.. … but new cost and energy challenges must be met to unleash AI opportunity AI computing architecture1 AI architecture cost2 The semiconductor industry needs to deliver the highest computing power / transistor density at the lowest cost, including the lowest possible CO2 emission Page 15 AI energy consumption3 Cost estimate: ~500M$2 exa FLOPS 60MW Source:Lisa Su, AMD, ITF May 21 2024 Argonne Leadership Computing Facility: https://www.alcf.anl.gov/aurora November 14, 2024
Public We expect AI applications to accelerate the need for advanced logic roadmap 2018 2020 2023 2025 2027 2029 2031 2033 2035 2037 2039 Future 2D and 3D innovations are paving the way for the next 15 years Page 16November 14, 2024
Public We expect AI applications to transform DRAM architecture and volume Future DRAM should integrate additional logic functionalities to improve performance & energy efficiency Page 17 “P M can improve performance and energy efficiency of memory-bound workloads by adding additional logic functionality to DR M memory” November 14, 2024
Public 10 100 1000 2015 2020 2025 2030 So, we continue to anticipate strong transistor growth for both logic and DRAM And we expect the semiconductor demand mix to shift towards advanced logic and DRAM Page 18 D R A M G b p ro d u c e d [ 1 0 1 8 / m o n th ] C GR ’15-’23: 18% DRAM Source: Logic history: ASML end market model, DRAM history: WSTS, projections: ASML 10 100 1000 2015 2020 2025 2030 T ra n s is to rs p ro d u c e d [ 1 0 1 8 / m o n th ] C GR ‘15-’23: 26% Logic November 14, 2024
Public Cost and energy reduction should be at the core of future process optimization 1. Increasing the number of good transistor processed at every step 2. Simplifying the overall process flow 3. Minimizing cost and emission of each processing step 40~80 cycles Page 19 Total cost and total emissions of wafer patterning must be reduced to support AI roadmap Total cost and emissions of wafer patterning can be reduced by: CO2 November 14, 2024
Public Page 201Source: ASML, excluding innovations in standard cell design G o o d p ri n te d t ra n s is to r / € [ a .u .] 1 ASML has delivered higher transistor density at lower cost for several decades We believe that EUV scalability & holistic lithography can extend our historical trend into the next decade (Critical layers) 2005 2010 2015 2020 2025 2030 2035 2040 Holistic lithography capability over time ArFi EUV 0.33 EUV 0.55 EUV 0.75 ArFi EUV 0.33 EUV 0.55 EUV 0.75 November 14, 2024
Public Page 211Source: ASML, excluding innovations in standard cell design G o o d p ri n te d t ra n s is to r / € [ a .u .] 1 ASML has delivered higher transistor density at lower cost for several decades We believe that EUV scalability & holistic lithography can extend our historical trend into the next decade (Critical layers) 2005 2010 2015 2020 2025 2030 2035 2040 Holistic lithography capability over time Overall November 14, 2024
Public 2024 2030 2024 2030 2024 2030 Good printed transistor CO2 Cost of Technology Environmental cost +150% /exposure - 30% /exposure - 50% /exposure EUV Extendibility ASML’s innovations can extend the benefits of EUV lithography More good transistors at lower cost and energy for customers, at increased profitability for ASML Page 22 = Good printed transistors Lithography Exposure cost CO2 Note: Cost / exposure reduction not taking into account possible inflation November 14, 2024
Public Page 23 = Good printed transistors Total lithography cost CO2 ASML’s innovations can extend the benefits of EUV lithography More good transistors at lower cost and energy for customers @ increased profitability for ASML November 14, 2024
Public Good printed transistors Resolution ProductivityAccuracyPatterning Yield × × × Operational cost Environmental costSystem cost Lifetime/ + 1 Ton CO2 = 200 EUR + Total lithography cost ASML intends to innovate on all aspects to maximize its product portfolio value nnovation per product (EUV, DUV…) will target specific customer & market needs Page 24 = November 14, 2024
Public ACustomer trust and partnership remains at the core of ASML’s strategy Seon-Yong Cha, SK Hynix CTO We expect that our ability to scale EUV technology into the next decade and extend our versatile holistic lithography portfolio could place ASML at the heart of the AI opportunity. This would continue to create significant growth in revenue and profitability in this decade.
Public We anticipate that semiconductor manufacturers will continue to drive shrink Reducing transistor dimensions likely remains the easiest way to drive density up at reduced cost Page 26 SP: Single Patterning, DP: Double Patterning, QP: Quadruple Patterning, EPE: Edge Placement Error Source: 1Luc van den Hove, IMEC , ITF May 21, 2024 0 5 10 15 30 20 25 21 18 16 14 28 5,2 4,3 3,5 3,5 Public data Customer projection 2020 2023 2025 2027 2029 2031 2033 2035 17 14 3,8 22 23 E P E [ n m ], N o d e , m e ta l p it c h , L in e a r s c a le Year of high-volume production 7 20392037 12 3 5 3 2 1.4 1.0 0.7 0.20.5 0.3 sub-0.2 Logic metal pitch [nm] Edge placement error [nm] Uncertainty 5,5 4,5 5,8 Node name1 [nm] 40 Estimate November 14, 2024
Public Logic and DRAM are expected to drive further critical lithography exposures We expect EUV 0.33 NA layers to continue to increase node on node Page 27 2025 2029 - 2030 2025 2029 - 2030 high low high low low high NXE 0.33 NA low high # o f E U V 0 .3 3 N A E x p o s u re s # o f E U V 0 .3 3 N A E x p o s u re s 10-20% EUV spend CAGR 15-25% EUV spend CAGR DRAMLOGIC EUV lithography 0.33NA exposures (weighted average) EUV lithography 0.33NA exposures (weighted average) November 14, 2024
Public EUV 0.33 NA could enable further multi-patterning to single expose conversion Resulting in cost, yield and cycle time benefits for our customers Page 28 2022 2024 2026 2028 2030 2032 Single Expose (EUV 0.33 NA) A v e ra g e p a tt e rn in g c o s t p e r la y e r 2018 2020 LOGIC (LE-LE-LE – layer) DRAM (SA-LE-LE - layer) LOGIC (LE-LE – layer)2 1 3 Legend: LE = Litho-Etch | SA = Self aligned Cost of multi-patterning Cost of single expose November 14, 2024
Public EUV 0.33 NA could enable further multi-patterning to single expose conversion Resulting in higher number of EUV exposures for ASML between 2025 and 2030 Page 29 2022 2024 2026 2028 2030 2032 SE (EUV) A v e ra g e p a tt e rn in g c o s t p e r la y e r 64% 25% 17% 58% 33% 3% LE-LE-LE SE (EUV) LOGIC – 5 layers 2 DRAM (DUV-EUV) 2025 2027 1 3x LE SE 2019 LOGIC (DUV-EUV) 2x LE SE 2028 LOGIC (DUV-EUV) 3 Litho CAPEX Litho OPEX Non Litho Ref. (SE DUV) SE DUV 4x SPACER 2x SPACER 2x LESE SE SE LOGIC (LE-LE-LE – layer) DRAM (SA-LE-LE - layer) LOGIC (LE-LE – layer) >4X >2X >1.5X>2X >1.5X increase in litho intensity % o f p a tt e rn in g c o s t Legend: LE = Litho-Etch | SA = Self aligned November 14, 2024
Public 0.55 NA EUV opportunity starts now as 0.33 NA EUV transitions to multi-patterning on <26nm pitch critical layers SP: Single Patterning, DP: Double Patterning, QP: Quadruple Patterning Source: 1Luc van den Hove, IMEC , ITF May 21, 2024 0 5 10 15 30 20 Single Expose 0.55 Double Expose 0.55 0.55 NA EUV ADOPTION OPPORTUNITYSingle Expose 0.33 Double Expose 0.3325 21 18 16 14 28 5,2 4,3 3,5 3,5 Public data Customer projection Estimate 2020 2023 2025 2027 2029 2031 2033 2035 17 14 3,8 22 23 E P E [ n m ], N o d e , m e ta l p it c h , L in e a r s c a le Year of high-volume production 7 20392037 12 3 5 3 2 1,4 1,0 0,7 0,20,5 0,3 sub-0.2 Logic metal pitch [nm] Edge placement error [nm] Uncertainty 5,5 4,5 5,8 Node name1 [nm] 40 Page 30November 14, 2024
Public LOGIC: High NA EUV enables more design freedom with 1.5D and 2D designs Single expose simplification reduces process steps, cycle time and improves yield Exposure NXE 1 Exposure NXE 2 Exposure NXE 3 P22 in horizontal direction combined with P28 in vertical direction. Page 31 This multi-exposure method is not performed in high-volume manufacturing due to its complexity. High NA Single Patterning Low NA Multi Patterning High NA Logic metal: ~35% cost benefit and process simplification Exposure High NA1 November 14, 2024
Public Logic and DRAM are expected to drive further critical lithography exposures Insertion of High NA in high volume manufacturing in 2026-27, adoption by end of the decade 2025 2029 - 2030 DRAMLOGIC 4-6 2-3 Expected average High NA Layers Page 32 2025 2029 - 2030 high low high low low high NXE 0.33 NA EXE 0.55 NA low high # o f E U V 0 .3 3 N A E x p o s u re s 4-9 EUV lithography 0.33NA Equivalent exposures (weighted average) >2030 # o f E U V 0 .3 3 N A E x p o s u re s >2030 >3 EUV lithography 0.33NA Equivalent exposures (weighted average) November 14, 2024
Public High NA EUV’s latest optics innovation sets the base for our EUV roadmap Picometer stability (1/200 Si atom) achieved on asymmetric mirrors Page 33 NA >0.5 High NA Mirror Metrology at Zeiss Mask level Wafer level NA 0.33 Wafer level Mask level November 14, 2024
Public EUV source power continues to scale with opportunity to achieve >1000W 740W EUV power demonstrated EUV source power scaling 2010 2015 2020 2025 0 100 200 300 400 500 600 E U V s o u rc e p o w e r [W a tt s ] Products 2030 700 800 900 1000 - measures identified to reach >1000W in the future 1 2 Research 1 EUV Plasma 1μm Pre-Pulse 1μm Rarefaction Pulse 10μm Main Pulse 2 Plasma Process for 1 and 10 µm IR source Target Rarefied Target Tin Droplet 1µm architecture Droplet Repetition Rate Page 34November 14, 2024
Public High NA EUV optics support the vision of a higher productivity EUV platform EUV performance and productivity can be extended far into the next decade (>2030) 0.33 NA 0.55 NA 0.75 NA Today EXE platform NXE platform System commonality~50% ~95% Future: higher productivity platform Page 35November 14, 2024
Public 0.75 NA EUV opportunity is in the next decade when <16nm pitch will be needed Higher productivity platform can be designed to support future Hyper NA needs Page 36 SP: Single Patterning, DP: Double Patterning, QP: Quadruple Patterning Source: 1Luc van den Hove, IMEC , ITF May 21, 2024 Single Expose 0.75 0 5 10 15 30 20 25 21 18 16 14 28 5,2 4,3 3,5 3,5 Public data Customer projection Estimate 2020 2023 2025 2027 2029 2031 2033 2035 17 14 3,8 22 23 E P E [ n m ], N o d e , m e ta l p it c h , L in e a r s c a le Year of high-volume production 7 20392037 12 3 5 3 2 1,4 1,0 0,7 0,20,5 0,3 sub-0.2 Logic metal pitch [nm] Edge placement error [nm] Uncertainty 5,5 4,5 5,8 Node name1 [nm] 40 Multi Expose 0.75 0.75 NA EUV ADOPTION OPPORTUNITY Growing opportunity window for Logic beyond 2032 for replacing multi patterning 0.55 NA by 0.75 NA single patterning Double Expose 0.55 November 14, 2024
Public We expect front end 3D integration to complement 2D shrink in driving density Front end 3D integration challenges will trigger new litho opportunities for all semiconductor products Page 37 Stack LOGIC W-W hybrid W-W Fusion W-W | D-W Fusion Overlay 3D NAND BSPN CFET HVM 2026 >2032 Array CMOS Logic Bare-Si Logic Logic Bonding 5nm → 2nm50nm → 25nm 2.5nm → 1.6nm Litho KrF NXE/EXE NXE/EXE Array W-W / D-W hybrid Array CMOS 50nm → 25nm KrF NAND DRAM W-W hybrid W-W hybrid W-W / D-W hybrid 3D ARRAY4F2 2D ARRAY >2027 >2032 >2032 Array CMOS Array CMOS Array CMOS Array >4.5nm6nm → 3nm 6nm → 3nm ArFiArFi ArFi >2030 November 14, 2024
Public Holistic Lithography should support front end 3D integration Metrology and scanner control at pre and post bonding are critical process points needed for overlay Page 38 Pre-bonding CMOS Scanner correction and control ARRAY Scanner and offline metrology 2 1 Actuators Post-bondingBonding ARRAY CMOS Actuators large wafer deformation 3 Metrology Litho scanner Holistic lithography process control points Actuators 321 >5000 measurements/wafer 50 - 100nm overlay error >2000 measurements/wafer < 5 nm overlay error< 5 nm overlay error Holistic lithography to bring overlay error within spec massive metrology November 14, 2024
Public Lithography is a formidable tool to compensate process fingerprints Correction capability increased ~5 orders of magnitude, now @100 000 parameters per exposure Exposure stage 100% of wafers are measured 100% of wafers are processed field-by-field Metrology stage Even Fingers Odd Fingers Grey Filter X Y Z Optical Centerline Dose manipulator Flexible Illuminator Reticle stage 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025 1000 100 10 1 0.1 O v e rl a y [ n m ] 1,000 100 10 1 10,000 # u s e r s e le c ta b le l it h o c o rr e c ti o n s 100,000 Increasing correction capability per field, wafer, lot Scanner and Process Control Software November 14, 2024 Page 39
Public Wafer bonding can drive transistor density increase for DRAM memory 3 scenarios are currently in play for DRAM roadmap, all include bonding and require major innovations Page 40 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 D1c D1d D0a D0b D0c D0d D0e D0f 1) CBA: CMOS Bonded Array – CMOS logic wafer bonded to the memory array 2) Stacking: Array bonding – 2 or more memory array wafers bonded to each other. Wafer or die based 6F26F2 Source: Corporate Marketing (CMKT) analysis
Public Wafer bonding can drive transistor density increase for DRAM memory 3 scenarios are currently in play for DRAM roadmap, all include bonding and require major innovations Page 41 2025 2026 2027 2028 2029 2030 2031 2033 2034 2035 2036 1c 1d 0a 0b 0c D0d D0e D0f 1) CBA: CMOS Bonded Array – CMOS logic wafer bonded to the memory array 2) Stacking: Array bonding – 2 or more memory array wafers bonded to each other. Wafer or die based 6F26F2 2032 6F2 + CBA 4F2 + CBA4F2 MEMORY CARRIER LOGIC Bonding MEMORY CARRIER LOGIC Bonding Source: Corporate Marketing (CMKT) analysisNovember 14, 2024
Public Wafer bonding can drive transistor density increase for DRAM memory 3 scenarios are currently in play for DRAM roadmap, all include bonding and require major innovations Page 42 2025 2026 2027 2028 2029 2030 2031 2033 2034 2035 2036 1c 1d 0a 0b 0c D0d D0e D0f 1) CBA: CMOS Bonded Array – CMOS logic wafer bonded to the memory array 2) Stacking: Array bonding – 2 or more memory array wafers bonded to each other. Wafer or die based 6F26F2 2032 6F2 + CBA 4F2 + CBA4F2 MEMORY CARRIER LOGIC Bonding MEMORY CARRIER LOGIC Bonding 3D Integration 0d 0e 0f Source: Corporate Marketing (CMKT) analysisNovember 14, 2024
Public Wafer bonding can drive transistor density increase for DRAM memory 3 scenarios are currently in play for DRAM roadmap, all include bonding and require major innovations Page 43 2025 2026 2027 2028 2029 2030 2031 2033 2034 2035 2036 1c 1d 0a 0b 0c 0d 0e 0f 1) CBA: CMOS Bonded Array – CMOS logic wafer bonded to the memory array 2) Stacking: Array bonding – 2 or more memory array wafers bonded to each other. Wafer or die based 6F26F2 4F2 + CBA4F2 2032 6F2 + CBA 3D Array ~>125 layers 2D Array 4F2 + CBA /Array 2D Array 6F2 + CBA /Array MEMORY CARRIER LOGIC MEMORY CARRIER LOGIC Bonding Bonding + Stacking Bonding Bonding + Stacking Bonding Source: Corporate Marketing (CMKT) analysisNovember 14, 2024
Public Bonding CBA Bonding Cost of technology remains a major criteria for our customers’ roadmap choices EUV scalability and holistic lithography can support future front end 3D integration schemes Page 44 2023 2025 2027 2028 2030 2032 2034 2035 2037 2039 2041 B it c o s t s c a lin g ( $ /G B ) – L o g s c a le 3D Array performance today; 5 layers 3D Array cost challenge performance of theoretical 125 layers if produced in HVM with today’s capability 6F2 / 4F2 6F2 3D Array insertion in 2032 requires 125 layers New DRAM factories New equipment 125 L New materials? 3D Array 2D Array Source: Corporate Marketing (CMKT) analysisNovember 14, 2024
Public Page 45 2023 2025 2027 2028 2030 2032 2034 2035 2037 2039 2041 3D Array performance today; 5 layers 3D Array cost challenge performance of theoretical 125 layers if produced in HVM with today’s capability 3D Array 6F2 / 4F2 6F2 3D Array insertion in 2032 requires 125 layers New DRAM factories New equipment 125 L CBA Bonding CBA Bonding + Stacking Bonding + Stacking Bonding Cost of technology remains a major criteria for our customers’ roadmap choices EUV scalability and holistic lithography can support future front end 3D integration schemes 300 L New Materials? B it c o s t s c a lin g ( $ /G B ) – L o g s c a le 2D Array Source: Corporate Marketing (CMKT) analysisNovember 14, 2024
Public ASML remains committed to its greenhouse gas neutrality ambitions Intensifying collaboration in the value chain aimed at accelerating climate action Page 46 Help society limit global temperature rise to 1.5°C Improvement lever 2025 2030 2040 2050 Paris Agreement Category Continue collaboration Collaborate to eliminate emissions from product use ▪ Improve energy efficiency of ASML products ▪ Support industry effort to improve access to affordable renewable energy, facilitating customers to source 100% renewable electricity by 2040 ▪ Collaborate as semiconductor industry to scale credible carbon removal options Collaborate to source GHG neutral products ▪ Set low-carbon design specifications ▪ Support transition to renewable energy ▪ Obtain commitment from suppliers to compensate any residual emissions from products to ASML as of 2030 G re e n h o u s e g a s n e u tra l G re e n h o u s e g a s n e u tra l Continue to drive energy efficiency & renewable energy ▪ Reduce energy use ▪ Use renewable energy ▪ Compensate any residual emissions as of 2025 ▪ Reduce energy use ▪ Use renewable energy ▪ Compensate any residual emissions as of 2025 G re e n h o u s e g a s n e u tra l G re e n h o u s e g a s n e u tra l S c o p e 1 & 2 S c o p e 3 Manufacturing & buildings Business travel & commuting Supply chain Product use Target Target Target This slide must be read in conjunction with, and is qualified by the disclosures set forth in, ASML's most recent Annual Report based on FRS and SML’s most recent nnual Report on Form 20-F filed with the U.S. Securities and Exchange Commission, including the Risk Factors in the Strategic Report with respect to dependencies upon emission reductions by third parties in our value chain and Appendix - Special note regarding emission targets Greenhouse gas neutral: Remaining emissions, after SML’s efforts to reach its GHG emission reduction targets, compensated by the same amount of metric tons of carbon credits that are verified against recognized quality standards
Public 450 400 350 300 250 200 150 100 50 0 Our EUV innovations are also expected to drive EUV energy consumption down Within a 15 year period at customers, we anticipate 80% reduction of energy needed per wafer exposed 2018 2021 2023 2033 Improvement in energy per wafer Total power equivalent 100% Absolute Power Energy per Wafer 2025 2029 20312027 20% 100% 50% A T P t h ro u g h p u t 3 0 m J [ W P H ] Wafers per hour T o ta l p o w e r e q u iv a le n t (k W ) Total wafers per hour November 14, 2024 Page 47
Public Attractive & Inclusive communities STEM ESG innovation Employee giving Contribute to positive improvements and experiences in our communities Increase the talent pool that society needs to solve some of its key challenges Support innovative ideas to solve key ESG challenges Engage with and care for people in our communities ASML is increasing its engagement with communities SML and communities benefit from each other’s presence and support each other’s development November 14, 2024 Page 48
Public Forward Looking Statements This document and related discussions contain statements that are forward-looking within the meaning of the U.S. Private Securities Litigation Reform Act of 1995, including statements with respect to our strategy, plans and expected trends, including trends in end markets and the technology industry and business environment trends, including the emergence of AI and its potential opportunities and expectations for the semiconductor industry, including computing power, advanced logic nodes and DR M memory, statements with respect to Moore’s law and expected transistor growth and aspirations by 2030, global market trends and technology, product and customer roadmaps, long term outlook and expected lithography and semiconductor industry growth and trends and expected growth in semiconductor sales and semiconductor market opportunity through to 2030 and beyond, expected growth in wafer demand and capacity and additional wafer capacity requirements, expected investments by our customers, including investments in our technology and in wafer capacity, plans to increase capacity, expected growth in lithography spend, growth opportunities including opportunities for growth in service and upgrades and opportunities for growth in Installed Base Management sales, expected growth and gross margins in the holistic lithography business and expected addressable market for pplications products, expectations and benefits of a growing installed base, SML’s and its supplier’s capacity, expected production of systems, model scenarios and the updated model for 2030, including annual revenue and gross margin opportunity and development potential for 2030, outlook and expected, modelled or potential financial results, including revenue opportunity, gross margin, R&D costs, SG&A costs, capital expenditure, cash conversion cycle and annualized effective tax rate for 2030 and assumptions and drivers underlying such expected, modelled or potential amounts, and other assumptions underlying our business and financial models, expected trends, outlook and growth in semiconductor end markets and long term growth opportunities, demand and demand drivers, expected opportunities and growth drivers for and technological innovation of our products including DUV EUV, High NA, Hyper NA, Applications, and other products impacting productivity and costs, transistor dimensions, logic and DRAM shrink, foundry competition, statements with respect to dividends and share buybacks and our capital return policy, including expectation to return significant amounts of cash to shareholders through growing dividends and buybacks and statements with respect to energy generation and consumption trends and the drive toward energy efficiency, emissions reduction and greenhouse gas neutrality goals and target dates to achieve greenhouse gas neutrality, zero waste from operations and other ESG targets and ambitions and plans to maintain a leadership position in ESG, increasing technological sovereignty across the world and the expected impact on semiconductor sales, including specific goals of countries across the world, increasing competition in the foundry business, estimates for 2024 and other non-historical statements. You can generally identify these statements by the use of words like "may", "will", "could", "should", "project", "believe", "anticipate", "expect", "plan", "estimate", "forecast", "potential", “opportunity”, “scenario”, “guidance,” "intend", "continue", "target", "future", "progress", "goal" and variations of these words or comparable words. These statements are not historical facts, but rather are based on current expectations, estimates, assumptions, models, opportunities and projections about our business and our future and potential financial results and readers should not place undue reliance on them. Forward- looking statements do not guarantee future performance and involve a number of substantial known and unknown risks and uncertainties. These risks and uncertainties include, without limitation, customer demand, semiconductor equipment industry capacity, worldwide demand for semiconductors and semiconductor manufacturing capacity, lithography tool utilization and semiconductor inventory levels, general trends and consumer confidence in the semiconductor industry and end markets, the impact of general economic conditions, including the impact of the current macroeconomic environment on the semiconductor industry, uncertainty around a market recovery including the timing thereof, the impact of inflation, interest rates, wars and geopolitical developments, the impact of pandemics, the performance of our systems, the success of technology advances and the pace of new product development and customer acceptance of and demand for new products, our production capacity and ability to adjust capacity to meet demand, supply chain capacity, timely availability of parts and components, raw materials, critical manufacturing equipment and qualified employees, our ability to produce systems to meet demand, the number and timing of systems ordered, shipped and recognized in revenue, risks relating to fluctuations in net bookings and our ability to convert bookings into sales, the risk of order cancellation or push outs and restrictions on shipments of ordered systems under export controls, risks relating to technology, product and customer roadmaps and Moore’s law, risks relating to the trade environment, import/export and national security regulations and orders and their impact on us, including the impact of changes in export regulations and the impact of such regulations on our ability to obtain necessary licenses and to sell our systems and provide services to certain customers, exchange rate fluctuations, changes in tax rates, available liquidity and free cash flow and liquidity requirements, our ability to refinance our indebtedness, available cash and distributable reserves for, and other factors impacting, dividend payments and share repurchases, the number of shares that we repurchase under our share repurchase programs, our ability to enforce patents and protect intellectual property rights and the outcome of intellectual property disputes and litigation, our ability to meet ESG goals and execute our ESG strategy, other factors that may impact SML’s business or financial results including the risk that actual results may differ materially from the models, potential and opportunity we present for 2030 and other future periods, and other risks indicated in the risk factors included in SML’s nnual Report on Form 20-F for the year ended December 31, 2023 and other filings with and submissions to the US Securities and Exchange Commission. These forward-looking statements are made only as of the date of this document. We undertake no obligation to update any forward-looking statements after the date of this report or to conform such statements to actual results or revised expectations, except as required by law. This document and related discussions contain statements relating to our approach to and interim progress on achieving certain energy efficiency and greenhouse gas emissions reduction targets, including our ambition to achieve greenhouse gas neutrality. References to “greenhouse gas neutral” means remaining emissions, after SML’s efforts to reach its GHG emission reduction targets, compensated by the same amount of metric tons of carbon credits that are verified against recognised quality standards. Page 49November 14, 2024
Public THANK YOU