Ethiopia Can't Wait for the Market to Code Its Future

Ethiopia Can't Wait for the Market to Code Its Future

Jun 13 , 2026. By Anteneh Tadesse ( Anteneh Tadesse is an economist serving as chief strategy officer at Africa Capital Partners, with expertise in Ethiopia’s regulatory landscape and market positioning strategies for government agencies and start-ups. )

The ultimate success of Ethiopia's state-led model will depend on what sociologists call "embedded autonomy", a state independent enough to avoid political capture or funding prestige projects, yet close enough to business to adapt to market feedback. Initiatives like the Sector Skills Body, piloted in agro-processing in 2024, show an effort to bridge the classroom-to-firm gap, but long-term evaluation should focus on actual wage gains, employment metrics, and domestic applications rather than raw enrollment figures, argues Anteneh Tadesse (anteneh@africacapitalp.com), an economist serving as chief strategy officer at Africa Capital Partners, with expertise in Ethiopia’s regulatory landscape and market positioning strategies for government agencies and start-ups.

At a technology conference in Nairobi a few years ago, a product director at one of Africa’s better-known fintech firms talked about why his company had chosen Kenya, not Ethiopia, as its expansion base.

Kenya had better application programming interface (API) infrastructure, a deeper developer community and clearer regulatory pathways. Then he quipped, almost casually, that Ethiopia would probably leapfrog all of it within a decade if the government’s digital agenda delivered. He meant it as praise, not dismissal.

However, this remark captured Ethiopia’s technology dilemma. Its business and citizens’ present frustration is wrapped in long-held confidence. The infrastructure that entrepreneurs admired in Kenya did not appear on its own. M-Pesa grew on regulatory choices and public support. Kenya’s developer ecosystem drew strength from universities, diaspora networks and institutions partly shaped by the state.

Yet Africa’s technology success stories are usually told as market triumphs, with government offstage.

Ethiopia is now making a different bet. Rather than wait for markets to create the conditions for technological development, it is trying to build those conditions directly. Five Million Coders, the Ethiopian Artificial Intelligence Institute (EAII), the law governing startups and the policy for national entrepreneurship development, endorsed in January this year, form a state-led technology architecture. The premise is that private investors alone will not finance such foundations quickly enough for a country of over 100 million people, a median age below 20 and perhaps two million young people entering the labour market each year.

The outcome is uncertain, while the approach is not. It belongs to a tradition of state-led development that has underpinned successful industrial transformations.

Although Mariana Mazzucato’s book gave the “entrepreneurial state” an unambiguous name, the idea is older. Development economists had argued since at least the 1980s that many technologies celebrated as private innovation began with public investment. The global positioning system (GPS) chip in a phone, the internet connection that carries data, and voice-recognition software all trace their roots to United States federal agencies. Silicon Valley commercialised them brilliantly, but it did not invent their foundations.

Alice Amsden’s work on South Korea, Robert Wade’s on Taiwan and Ha-Joon Chang’s survey of industrialised economies reached a similar conclusion. Countries that industrialised intervened in markets, directed credit and protected infant industries. Afterwards, many joined international bodies that advised poorer countries against using the same tools.

The logic is not ideological. Markets work well when returns are knowable and timeframes manageable. They work less well when uncertainty is high, payoffs stretch across decades, and benefits spread beyond a single firm. Basic research, infrastructure and early industries fall into that gap. The state’s role is not to displace markets, but to create conditions under which they function and flourish.

The American case is instructive because the United States is often presented as market-led. DARPA, created in 1958 after Sputnik, funded the ARPANET, which became the internet, along with stealth technology, precision-guided munitions and GPS. ARPA-E, established in 2007, finances energy research that venture capital often avoids because the payoff is distant and uncertain. The National Institutes of Health, with budgets that have at times exceeded 50 billion dollars, funds biological research that drug companies later translate into products.

A 12-year analysis of top-selling approved drugs found that everyone benefited from publicly funded research upstream. NASA’s Apollo programme seeded materials science, computing and satellite communications. Although private enterprise mattered, the state laid much of the groundwork.

The financial crisis in 2008 offered another lesson. The crisis grew from excessive private debt, mortgage-backed securities and derivatives, not public extravagance. Governments became indebted afterwards because they absorbed the shock. Europe responded with austerity, cutting spending to restore confidence. Greece, Spain, Portugal and the United Kingdom (UK) suffered falling output, rising unemployment and weak investment.

Worse, cuts hit education, research, infrastructure maintenance and public health. The International Monetary Fund (IMF) later acknowledged that it had underestimated austerity’s contractionary force in a downturn. The United States (US), through stimulus measures, and Germany, through short-time work, better protected productive capacity.

The lesson for Ethiopia is not that deficits are harmless, but that debt quality matters. Borrowing for a road linking farms to a port or a fibre network enabling digital exports is different from borrowing to subsidise consumption or sustain unviable firms. Fiscal analysis that looks only at the size of the deficit misses what the deficit buys.

Seen separately, Ethiopia’s initiatives can look like announcements. Together, they show a policy architecture. The millions of coders rest on a defensible bet that Ethiopia’s demographic dividend will not appear unless young people gain knowledge-economy skills. South Korea invested heavily in engineering education from the 1960s, not because demand already existed, but because the state intended to create it. Ethiopia’s programme will succeed only if the quality of the curriculum, instructors, and employer links improves. Enrollment numbers alone will mean little.

Digital Ethiopia addresses underprovided infrastructure, including rural broadband, digital identity, payments, and data governance. A telecom company may capture subscription revenue from fibre, but not the wider gains from farmers checking prices, health workers accessing protocols or small businesses reaching new customers. Those spillovers justify public provision.

The Ethiopian Artificial Intelligence Institute follows the same logic. Artificial intelligence is a general-purpose technology, shaping sectors from health to agriculture. If Ethiopia merely imports tools built for other languages and markets, it becomes a consumer dependent on foreign updates and design choices. Building research capacity, local data systems and models adapted to domestic conditions gives the country more agency.

The risks are brain drain, unstable funding and weak commercialisation. But leaving AI entirely to markets would likely leave Ethiopia with access rather than capability.

The proposed AI factory infrastructure, high-performance computing clusters for model training and deployment, resembles a digital industrial zone. A domestic firm cannot justify such computing capacity without a publicly backed facility that lowers entry costs and supports private investment, much as industrial zones have done for manufacturing.

The ecosystem agenda extends beyond infrastructure. The entrepreneurship development policy recognises that entrepreneurship requires more than deregulation. Early-stage capital, mentorship, market links, technical support and legal clarity are usually missing in low-income economies. The Ethiopian Entrepreneurship Development Institute (EEDI), an autonomous quasi-governmental body under the Ministry of Labour & Skills, is meant to provide that connective tissue.

The startup law Parliament passed in July 2025, after nearly five years of inter-ministerial deliberation, gives startups legal recognition. Before it, a technology venture had no distinct legal identity it was treated identically to a microenterprise. Now, a startup is defined as a technology-based enterprise under three years old, built on an innovative idea with the potential to scale. The law offers tax holidays, customs exemptions on equipment, simplified licensing and access to a public Startup Innovation Fund. A Startup Ethiopia Council brings together government, business, academia and civil society.

However, Nigeria serves as a warning, as it struggled because implementation lagged behind the law. Although Ethiopia’s authors of the startup law knew this, awareness alone is not sufficient, as it is not implementation. In the absence of institutions, funds and staff, laws on their own do not build ecosystems.

The energy agenda applies the same platform logic to industry. Apparel and textiles, agro-processing, leather goods and pharmaceuticals are targeted because they combine resource advantages, labour-cost advantages and some existing capacity. Reliable and affordable power, including hydroelectric investment such as the Grand Ethiopian Renaissance Dam (GERD), underpins it. Together, Digital Ethiopia, Made in Ethiopia, and the energy programme suggest the state is building the physical and digital commons on which firms compete.

Skills are the hardest test. The Ministry of Labour & Skills, EDI and the Federal Technical & Vocational Training Institute are trying to shift training toward what employers need. A Sector Skills Body piloted in agro-processing in 2024 aims to connect the curriculum to market demand. But equipment is often outdated, employer partnerships are limited, and the link between classrooms and firms is loose.

Although there may be a temptation for critics to characterise this as central planning, it is platform building. This is not an argument that state intervention is inherently good. Bad intervention creates patronage, protects incumbents and funds prestige projects.

But, intervention works when it is disciplined and accountable. South Korea and Taiwan tied subsidies to performance and adjusted policies to market feedback. Peter Evans called this “embedded autonomy.” It is a state independent enough to resist capture, but close enough to business to understand reality.

Ethiopia’s programmes should be judged that way. The coders should be measured by employment and wage gains, not enrollment. The AI Institute should be judged by domestic applications. The Startup Fund should track survival and revenue, not disbursements. Leapfrogging is real but easily overstated. M-Pesa demonstrated that latecomers can bypass legacy banking models. Yet adopting technology is not the same as building capacity to adapt and extend it.

Ethiopia’s wager is to become more than a user of imported systems. This wager should be urgent, with a young population and a narrow employment window. Waiting for private markets to evolve organically is too slow. The gap between strategy and delivery remains large, and Ethiopia’s record warrants caution. Still, disciplined public investment in technological and industrial capacity is not speculation. It is what much of the past century’s development record shows. For Ethiopia, the case for acting is strong, but the case for acting well is stronger.

PUBLISHED ON Jun 13,2026 [ VOL 27 , NO 1363]