Using LLMs to uncover Europe’s green investment blind spot

“If green investment is defined too narrowly, businesses risk misreading where the transition is creating competitive advantage. Combining the use of LLM tools with the EU Taxonomy for Sustainable Activities, Juan Alvarez-Vilanova , Riccardo Crescenzi and Lee Mager show that green FDI is substantially larger and more diverse than conventional statistics show and explain how that changes how projects should be screened, designed and communicated. If asked how much foreign investment into Europe is “green”, most experts would almost certainly reach for figures on renewable energy. Solar , wind and waste management are the go-to proxies: visible, measurable and politically legible. The problem is that this is a bit like measuring a city’s food economy by counting its supermarkets. We would miss the restaurants, the market stalls, catering companies and everything that keeps people fed but does not look like a shop. Just as the food system is much bigger than its supermarkets the green economy is much bigger than its power sector. The green transition is an economy-wide challenge, as well as an investment problem Sectors such as steel, chemicals, automotive and construction are among the largest carbon emitters on the planet. Transitioning these onto greener production processes is just as urgent as building wind farms – and arguably harder. It not only requires the development of new technologies and production processes, but also their diffusion and adoption across long and complex supply chains, spanning multiple countries and involving thousands of different firms, often with highly diverse technological capabilities. At the same time, producing the batteries, cables, heat pumps and advanced materials – as well as the engineering, financial and insurance services – that the green transition depends on is itself a green activity. In short, the transition requires an economy-wide transformation that extends well beyond renewable energy generation. And that transformation needs to be financed. McKinsey estimates the world needs roughly $9 trillion in annual investment in physical assets to achieve net zero by 2050. Private capital, and foreign direct investment (FDI) in particular, is essential to closing that gap. Not only because public budgets alone cannot get close to the scale required, but because foreign investments bring with them the external know-how and capabilities that are essential for new green technologies to diffuse across the economy. If business leaders, investors and policy makers only focus on the renewable energy slice of green investment, they will overlook the full extent to which FDI can contribute to wider sustainability objectives. Yet it is often difficult to distinguish between companies that are truly advancing the green transition and those that are not, especially in hard-to-abate sectors or where emerging green technologies are still developing. Without a clear definition, the door is open to greenwashing and the misallocation of capital. That is precisely the problem the EU Taxonomy for Sustainable Activities was designed to solve. The EU Taxonomy offers a new lens for green investment The EU Taxonomy for Sustainable Activities, adopted in full by the European Commission in 2023, was built on a straightforward but consequential idea: that “green” extends well beyond the energy sector, with advances in technology and innovation enabling environmentally sustainable activities within almost every industry. Source: EU Taxonomy Navigator The Taxonomy defines what counts as a sustainable economic activity across six environmental objectives (shown above), spanning not just power generation but manufacturing, transport, construction, agriculture and services. Crucially, it recognises that economic activities can make an environmental contribution in distinct ways. Some activities are green because of how they are performed (such as an electric arc furnace to make steel that runs on low-carbon electricity); others because they enable other activities to become greener (for example, IT solutions that track air pollution); others because they adopt low-impact inputs in an otherwise conventional process (such as tile-making using recycled ceramics); and others because they represent best-available technology in sectors that cannot yet fully decarbonise (for example, the use of LEAP aircraft engines in aviation). The Taxonomy makes it possible to identify and capture, with much greater precision, the green activities undertaken by foreign investors across all sectors of the host economy — from advanced manufacturing and research and development (R&D) to industrial process upgrading. To qualify, each investment project must meet specific, science-based technical screening criteria for its corresponding sector. Our approach: from the EU Taxonomy to green FDI Applying the EU Taxonomy to tens of thousands of investment projects requires a way of reading and evaluating what those projects actually do and doing so consistently against detailed technical criteria set out in the Taxonomy. We used Large Language Models (LLMs), prompted with the EU Taxonomy’s detailed technical criteria, to read and classify the complex textual descriptions of individual investment projects at scale. Unlike keyword searches, a properly prompted LLM can evaluate the full description, take into account the context and assess whether the specific activity described meets the relevant technical criteria. This approach identifies genuinely green investments that use no “green vocabulary” at all — such as a manufacturer’s new production line focused entirely on more energy-efficient LEAP aircraft engines, which are explicitly recognised within the Taxonomy’s technical screening criteria but described in purely technical terms. It also correctly flags potential greenwashing: investments merely described by a chief executive as “environmentally friendly”, but whose description disclosed no qualifying activity. What our analysis reveals Nearly one in six euros of inward FDI into the EU and Britain (EU27+UK) involves “green” activities. And half of that is missed by standard measures. Across all 109,000 greenfield FDI projects into the EU27+UK between 2013 and 2024, our analysis identifies 15.7 per cent of total inward FDI capital value as green: roughly twice the share captured by conventional approaches. We distinguish two components. The first, which we call “renewable energy and waste” (REW) green FDI, corresponds to the green investment that most sector-based proxies would already capture. The second, which we call “beyond-energy” green FDI, comprises investment involving green activities aligned with the Taxonomy flowing into manufacturing, R&D, industrial processes and other sectors. This beyond-energy component, equivalent to 7.4 per cent of total inward investment, is the “blind spot” our approach reveals. Green FDI into EU27+UK as a share of total inward FDI, 2013–2024 : Beyond-energy green FDI (green) overtook REW green FDI — the renewable energy and waste investment captured by conventional approaches — from 2019 onwards. Source: Alvarez-Vilanova, Crescenzi and Mager 2026 The trend over time reinforces this picture. REW green FDI’s share of total inward investment reached its peak in 2020, hovering around 10 per cent since then. Beyond-energy green FDI, by contrast, was barely visible before 2017 and has since expanded substantially. By 2022–2024, it accounted for the majority of all green inward investment into Europe. The sectoral breakdown is instructive. Manufacturing is driving this beyond-energy green FDI. Knowledge-intensive manufacturing sectors, in particular, account for the largest shares: more than a third of all FDI into computers, electronics and electrical devices qualifies as green, as does roughly a quarter of automotive investment. Basic metals, chemicals and machinery also make a substantial contribution. Moreover, we find that the specific manufacturing sectors attracting the greatest share of green FDI are precisely those that are more likely to be the most innovative. Green FDI intensity by broad sector group, 2013–2024: Within knowledge-intensive manufacturing, the sectors with the highest green FDI shares are computers and electronics (36.8%), automotive (24.5%), and chemicals (17.1%). Source: Alvarez-Vilanova, Crescenzi and Mager 2026 Services, by contrast, still contribute less, reflecting the fact that the Taxonomy’s technical criteria are more readily satisfied through tangible production processes than through the more diffuse activities of services businesses. But this does not mean services are peripheral to the wider green economy our analysis uncovers. On the contrary, the activity-based lens developed here can help identify where new opportunities lie: in the finance, logistics, engineering, software, design, certification and R&D tasks that support greener production beyond renewables. Implications for businesses and policymakers A more complete understanding of the boundaries of green investment matters for both business and policy decisions. Policy decisions about where to direct green industrial support, how to assess the climate credentials of investment attraction strategies and how to monitor progress toward decarbonisation targets are all downstream of how we define and count green investment. So too are questions about where this investment comes from and where it lands with implications that extend well beyond industrial policy into questions of technological sovereignty and economic development. For firms, the implications are equally practical. If green investment is defined too narrowly, businesses risk misreading where the transition is creating competitive advantage. Many of the investment opportunities that matter most now sit in cleaner manufacturing processes, low-carbon materials, energy-efficient equipment, electrified transport, industrial R&D and enabling technical or professional services. For multinational investors, that changes how projects should be screened, designed and communicated. Green foreign investment into Europe is substantially larger and more sectorally diverse than current statistics suggest. It is also, on current trends, growing fast. Policymakers and businesses who are navigating the green transition without this picture are working with a map that covers only half the territory. This blog is based on “ Green Foreign Direct Investment is flowing far beyond renewables: a taxonomy-guided LLM analysis” published as a discussion paper in Economic Geography and Spatial Economics, LSE Geography and Environment Department. This article gives the views of the author, not the position of LSE Business Review or the London School of Economics. You are agreeing with our comment policy when you leave a comment. Image credit: VectorMine provided by Shutterstock. The post Using LLMs to uncover Europe’s green investment blind spot first appeared on LSE Business Review .

Summary generated from the RSS feed of LSE Business Review. All article rights belong to the original publisher. Click through to read the full piece on blogs.lse.ac.uk/businessreview.