Investing in your capabilities Selecting the right development approach AI tools are becoming more commodified. Start- ups are building foundation models that can serve the industrial space; meanwhile, cloud providers are offering increasingly robust AI/ ML ops31 developer toolkits. This has enabled Lighthouses to choose to go in-house for development 37% more often than they did in 2019-2020 – primarily in roles related to AI architecture, design and integration (see Figure 9).32 As a result, while third parties continue to provide services, Lighthouses increasingly opt to own more of their information value chain and outsource only specific elements of tech stack and solution development. 2.2 Lighthouses increasingly opt to own more of their information value chain and outsource only specific elements of tech stack and solution development. Lighthouses are shifting solution development in-house FIGURE 9 Implementation approach by cohort, % of total1 38% 31% In-house Third-party Combination19%48% 19%21%50% 14%27%52% 10%30% Cohorts 1-3 (2018-19) Cohorts 4-6 (2020-21) Cohorts 8-10 (2022-23) Cohorts 11-13 (2023-24)+37% -68% +58%Select in-house capabilities have enabled Lighthouses to deliver more tailored solutions to local sites – Midea in China, for instance, built their own flexible and scalable platforms for product lifecycle management (PLM) R&D, IoT and cloud computing. This platform has enabled agile development and continuous delivery, with 20+ major upgrades per year.33 1. Excludes licensed, acquired and academia approaches (Waves 1-3: 12%, Waves 4-6: 12%, Waves 8-10: 9%, Waves 11-13: 8%) Source: Global Lighthouse Network.EXAMPLE Midea Global Lighthouse Network: The Mindset Shifts Driving Impact and Scale in Digital Transformation 17

Global Lighthouse Network 2025