Mapping venture capital investments in the (new) space economy with large language models: a comparative analysis of Europe and the United States
摘要
This paper maps Venture Capital (VC) investments received by startups operating in the new space economy located in European countries and the United States of America (US). Relying on a Large Language Model few-shot learning text-based classification methodology implemented via the GPT-5 model, we classify 113,910 VC-backed startups from the PitchBook database and identify 2,336 space startups. Results show that European VC-backed startups are relatively more specialized in the new space economy compared to their US counterparts. However, they raise lower amounts of capital, receive fewer VC rounds, less quickly, and are backed by smaller VC syndicates. They are also less likely to exit via IPO or merger and acquisition. These weaknesses are similar to those European VC-backed startups experience in other industries. Our results highlight that, despite their attractiveness to VC investors, European space startups suffer from the relative underdevelopment of the European VC industry. Overall, this paper provides an original overview of the VC funding landscape in the new space economy, offering relevant insights to improve the European industrial policy in this nascent industry.