“The difficulty lies, not in the new ideas, but in escaping from the old ones,” John Maynard Keynes wrote. For three decades, most economies in the EBRD regions benefited from deepening globalisation, a post-Cold War “peace dividend” and a clear path to technological catch-up. Those favourable conditions are now fading away. Globalisation is giving way to fragmentation, the peace dividend has succumbed to rising defence needs, and the predictable technological catch-up has been engulfed by an artificial intelligence (AI) revolution whose speed and implications remain uncertain. The challenge now is not simply to advance along an established path, but to define a new path entirely.
The post-communist economies of the EBRD regions have reinvented themselves before. In a typical EBRD economy, roughly half of the products in which the economy has a comparative advantage today differ from those of 25 years ago. That transformation has taken time and been supported by open export markets. The old playbook was relatively clear: import technology, invest in capital and upgrade product quality. The next transformation will be harder. Export structures are persistent, emerging sectors are complex, and building new capabilities involves trial and error.
AI illustrates this challenge. The supply chain for advanced chips, data infrastructure and frontier models is concentrated in a handful of countries and firms. The EBRD regions supply just 0.8 per cent of global exports of AI-related products, despite accounting for 8 per cent of total global exports across all goods. Power constraints also restrict their capacity to host large-scale data centres.
For most economies, the greater opportunity lies in developing AI-enabled applications and adopting AI across the wider economy. Realising these gains will require digital and power infrastructure, skills, access to computing capacity, effective data governance and international cooperation. AI adoption is likely to generate significant productivity gains, but differences in adoption capacity mean that these benefits will be uneven. The automation of some tasks will also require workers to move across occupations, firms and sectors, with early signs of this adjustment already visible in divergent hiring patterns for experienced and junior workers.
Defence spending presents a different opportunity. While spending on defence is a burden on the taxpayer, it can lift demand. Yet this additional demand will not automatically create a competitive domestic industry. Where procurement is import-intensive, much of the economic benefit flows abroad. Most economies in the EBRD regions remain highly dependent on imported defence goods, while state ownership, limited scale and weak access to capital constrain domestic production and innovation. In addition, state-owned enterprises in the EBRD regions are less efficient than their counterparts in advanced economies. EBRD economies with a strong industrial base may nevertheless be able to expand production through retooling, joint procurement and the licensing of technology developed elsewhere.
The EBRD regions do not lack entrepreneurs; they lack the financing and investment networks that are needed to turn promising startups into large firms. Many founders are internationally experienced and develop products outside their economies’ established export baskets, but relatively few manage to scale up. Growth-stage finance remains limited, geographically concentrated and increasingly dependent on foreign investors in later funding rounds. This restricts the emergence of large exporters, which play an important role in developing complex products, reaching international markets and generating positive spillovers for other firms.
The policy challenge, therefore, is to make good use of scarce public resources. With defence spending absorbing a growing share of constrained public budgets, support is best directed at the factors that can help existing capabilities to develop into new comparative advantages: skills, research infrastructure, access to finance and integration into cross-border supply chains. Where such capabilities are already present, they point towards AI-enabled applications, regional defence production and specialised roles in global value chains. Which of these are viable is something that firms will establish through experimentation, rather than something that policymakers can determine in advance.
This makes scaleup finance a critical missing link. The EBRD regions have entrepreneurs and ideas; what they lack is the capital and networks that are required to turn startups into globally competitive firms. Consequently, public policy should focus less on subsidising individual businesses and more on ensuring that private capital is willing to back growth through guarantees, co-investment, and deeper domestic and regional capital markets. The prize is not self-sufficiency but reinvention, turning ideas born in the EBRD regions into firms that grow globally and anchor the next generation of high-value activity at home.
EBRD Chief Economist
As technology advances and geopolitics shift, what countries produce and export changes. Rising global demand for artificial intelligence (AI) and increased defence spending could, in principle, create opportunities for economies to develop new export capabilities. Yet, leveraging these demand shocks will be challenging. Economies’ export structures tend to be highly persistent, and developing new export specialisations is more difficult in the case of complex products such as AI and military goods. Building new capabilities involves trial and error, requires active policy support and tends to be more successful where economies already have established specialisations in related upstream or downstream products. Large exporters often play a key role in developing and marketing new complex products, with spillovers to the broader corporate population.
Rising geopolitical tensions have brought the “peace dividend” of the past three decades to an end, with defence spending increasing in many economies where the EBRD operates. While increased defence expenditure can raise debt levels and draw resources away from welfare spending, it can also spread to civilian industries and boost overall output in an economy. At the same time, most economies in the EBRD regions remain highly dependent on imports of defence-related goods, with state ownership, limited scale and weak access to capital constraining domestic production and innovation. For economies with a strong industrial base, retooling, joint procurement schemes and the licensing of technologies developed elsewhere can help to expand domestic defence production.
Advanced AI relies on a capital-intensive supply chain of chips, data centres, electricity and specialised research. Economies in the EBRD regions play a limited role in these upstream segments: their AI-related hardware exports, frontier-model development and AI-firm valuations lag global leaders significantly, while power constraints restrict their capacity to host large-scale data centres. Greater opportunities lie downstream in the development of AI-enabled applications and the use of AI to boost productivity. As adoption capacity varies widely, benefits are likely to be uneven. Policymakers could facilitate successful AI adoption by prioritising infrastructure, skills, access to computing capacity, data governance and international cooperation.
Entrepreneurship is the engine of business dynamism and economic growth, and the EBRD regions are home to many growth-oriented startups. Their founders tend to have international experience, they frequently found more than one firm, often in different countries, and they tend to focus on products that are outside their economies’ established export baskets. Yet relatively few of these firms manage to scale up their operations successfully. Venture-capital funding remains thin, and in later funding rounds it comes largely from abroad, often from economies where founders previously studied or worked. Deepening venture-capital markets, connecting founders to investors and widening the geographical spread of finance are all essential in order to turn more startups into large productive firms.