The Global Risks Report 2024

National security as a driver of industrial policy 28 economies ranked “Ensure national security” as the top objective currently orienting policy. Of these:FIGURE 2.14 Source World Economic Forum Executive Opinion Survey 2023. NotePrivate sector respondents from 113 economies were asked to rank the three objectives that currently orient innovation and industrial policy and three challenges that should orient innovation and industrial policy in their country. The same seven options applied in both questions: (1) Accelerate lifelong learning and talent development; (2) Accelerate the green transition; (3) Address food and water scarcity; (4) Ensure national security; (5) Maximize economic growth and employment; (6) Protect public health and wellbeing; (7) Reduce inequality and promote diversity and inclusion. The final ranking of “Ensure national security” by country is based on a simply tally of the number of times it was selected. 19 economies felt national security should be deprioritized as a top objective in terms of innovation andindustrial policy, including the United Kingdom, Türkiye, and Italy, in favour of, for example,addressing environment and education priorities. 17 economies ranked “Ensure national security” as the top challenge that should orient policy. Of these: 9 economies ranked national security as the top objective and the top challenge that orients innovationand industrial policy in their country, including the United States, Hong Kong, United ArabEmirates and Pakistan. 8 economies felt national security should be prioritized as a top challenge, including Bangladesh, Mongoliaand Malaysia, ahead of current economic objectives, for example. Indicative value chain of generative AI technologies FIGURE 2.13 SourceWorld Economic Forum Global Risks Report 2024 .Note*Subject to restrictions.High-level for illustrative purposes only, does not include all relevant steps (including, for example, inference, web access, etc.).Level of concentration Low HighProprietary data sets can lend a competitive advantage; largercompanies are favoured due toeconomies of scale and networkeffects, particularly Big Tech companiesthat control data generated online.Data* Although comparatively nascent, earlysigns suggest that the market forfoundation models tend towardsconcentration, given both economiesof scale and scope.Foundation model Open-source models will increasecompetitiveness, however the longer-term market structure will depend onthe extent of regulated access tofoundational models.Applications* Most companies access compute powerthrough cloud platforms. Amazon (US)holds around 32% market share,followed by Microsoft (US) at 22%.Computational powerOver 80% of the world’s ATP productionis currently centred in Asia, split betweenIntegrated Device Manufacturers andOutsourced Semiconductor Assemblyand Test (OSAT) companies. Assembly, testing & packaging (ATP) Generative AI is dependent on advanced semiconductor technology. Nvidia (US)has captured 95% of the market forGPU chips that can be used formachine learning.Design*Taiwan Semiconductor ManufacturingCompany (TSMC) holds around a 56%share of the global semiconductorfoundry market, followed by Samsungat nearly 12%. Fabrication* Samsung (South Korea) is a key vendorfor memory storage technologies (DRAMand NAND, at 43% and 33% of globalproduction respectively), alongside SKHynix (South Korea).MemoryThe US has produced more than 70%of the most cited AI research papersover the past three years, followed byChina and the UK.Talent & IP* ASML (Netherlands) reportedly holds90% of the global lithography market,although Canon (Japan) haveannounced that their latest machinecan produce advanced 2-5nm chips. Lithography* As described in last year’s Global RisksReport (Chapter 3: Resource Rivalries),the extraction and processing of criticalmetals and minerals used in advancedAI chips are highly geographicallyconcentrated. Raw resources* Global Risks Report 2024 52