Artificial intelligence (AI) is set to revolutionize produc- tion processes, enable better and new services for cus- tomers, and transform the way that companies work. Asproduction of goods and services become truly digital,the traditional relationship between production facili- ties and human resources will change. AI, besides im- proving productivity, will become a factor of productionlike capital and labor. The new technology will also havea major impact on the competitive positioning of play- ers: Investing in AI not only cuts costs but by the sametoken enhances the competitive advantage of firms that innovate. In AI we are currently witnessing a remarkable shift ininnovation. Rather than coming from multinationalfirms as in the past, innovation now stems largely fromresearch laboratories, digital platforms and startups.These are the players creating algorithms and develop- ing use cases, they are the brains behind innovations inimage recognition, natural language processing and au- tomated driving. With this in mind, Roland Berger and Asgard have car- ried out the first comprehensive global investigation ofstartups working in the field of AI, the results of whichwe present in this paper. Unsurprisingly, the UnitedStates dominates the picture, accounting for nearly 40percent of AI startups. Europe ranks second with 22 per- cent of startups, ahead of China in third place. However,no single European country achieves critical mass on its own. In terms of sectors, some key European industries suchas energy or the automotive industry, and some areas oftechnology such as the Internet of Things (IoT) or robot- ics, are weakly represented among European AI startups.This finding may indicate that the true engines of theEuropean economy have not yet embraced the shift toAI. While the United States and China benefit from theirhighly developed digital industries, research and capital,the European landscape remains fragmented with thecontinent lacking a clear strategy for coordinating andstructuring the AI ecosystem. We see two main challenges for Europe. First, Europeshould promote the emergence of homegrown AI cham- pions that are able to offer the best solutions for enter- prises and consumers. Second, Europe must rapidlyadopt AI technology across all industries in order tobenefit from competitive gains. In the process, Europemust make sure it is not overtaken by early adopters inNorth America and Asia. We believe that Europe can become a third player in thearms race between the United States and China. Togain a competitive position, European policymakersshould promote startups as the main technological andeconomic drivers of AI. This involves profoundly chang- ing the shape of public involvement in the innovationecosystem and adapting support schemes specificallyfor startups. The European Union should quicklystrengthen the competitiveness of startups and supportthe emergence of a European AI ecosystem. Management summary 2 Artifcial Intelligence – A strategy for European startups Geographies, sectors and technologies 6 AI strategies and ecosystems across regions 14 Recommendations for Europe 18 Cov er p hot o: d 1sk/ is toc kph oto Contents 3 CHARTS MAPPING THE GLOBAL & EUROPEANAI STARTUPLANDSCAPE P. 8–13 Artifcial Intelligence – A strategy for European startups 3 Roland Berger and Asgard partnered up to provide the first compre hensive overview of the global AI ecosystem based on data fromstartups in all regions of the world.The data collection carried out by Asgard focused exclusively on startupsworking in AI technology, ignoring firms involved in other digital topicsand technology, such as hardwareonly startups and the R&D depart ments of corporations. Within this, the study further focuses on startupsproducing AI solutions, and expressly excludes startups using existingAI solutions on the market to develop new services or products.This selection procedure is dependent on the availability and qualityof the data gathered at the local level, in an everchanging environment,with different languages and standards in companies' registration.It therefore cannot reflect with scientific certainty the state of the localstartup ecosystems.Roland Berger then analyzed the information on 3,465 startups to define their geographical distribution and reach. The study looks at twokey variables: the main sector that the company focuses on, narroweddown to the 26 commonest industries, and the main technologyused or provided.Methodology 4 Artifcial Intelligence – A strategy for European startups 。。。。。。