Funding: High Compute Costs Threaten Africa’s $1.3bn AI Investment

African artificial intelligence (AI) businesses have attracted $1.3 billion in disclosed funding over the past decade, but high computing costs, inadequate infrastructure and limited access to growth capital are threatening the sustainability of startups across the continent.

A new study, AI Made in Africa: Voices from AI Businesses Across Africa, published by the German development agency, Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), in partnership with Briter Bridges, found that 311 AI companies across more than 35 sectors secured funding between 2016 and 2025.

However, the report indicates that the funding has not translated into widespread profitability, with many businesses struggling to manage the cost of developing and deploying AI solutions in markets where customers have limited purchasing power.

The study which draws on a survey of 85 AI businesses and interviews with 16 industry stakeholders, highlighted the structural challenges confronting African AI entrepreneurs, particularly the cost of computing infrastructure, power supply and access to specialised talent.

Meanwhile, for Nigerian startups, these challenges are compounded by foreign exchange volatility, as many businesses depend on cloud computing services, graphics processing units (GPUs) and AI application programming interfaces priced in dollars, while generating revenue largely in naira.

The report noted that 65 per cent of surveyed businesses identified cloud computing, GPUs and infrastructure as their largest research and development expense, underscoring the high cost of building AI products on the continent.

The funding landscape also remains uneven. Although the 311 mapped companies attracted a combined $1.3 billion, the median funding ticket was $225,000, while more than 75 per cent of the capital went to ventures established before 2020.

The report further found that nearly one-third of surveyed AI businesses remained entirely bootstrapped, while only 15 per cent were profitable.

This suggests that the growth of AI adoption across Africa is yet to translate into a broad-based investment and profitability cycle for businesses developing the technology.

Similarly, the report identified limited access to computing infrastructure as a major constraint to AI development, noting that Africa accounts for less than one per cent of global data centre capacity despite hosting about 18 per cent of the world’s population.

The infrastructure deficit has implications for Nigerian businesses, where unreliable electricity supply and the cost of powering data centres increase the operational expenses of technology companies.

The report cited Nigeria’s national grid as averaging about 4,000 megawatts of available power in early 2026, against an installed capacity of 13,000 megawatts.

It also noted that currency shocks have significantly increased the cost of cloud services for African startups, with the 2023 naira devaluation reportedly driving local cloud bills up by 224 per cent.

The report identified customer trust and limited AI literacy as major obstacles to commercial expansion, with about 65 per cent of surveyed businesses citing customer trust and low AI literacy as their primary challenge, ahead of high development costs, legacy system integration and lengthy sales cycles.

The findings also showed that AI companies are increasingly targeting institutional customers to improve revenue sustainability, with 88 per cent of surveyed businesses selling to other businesses, while 44 per cent also provide services to governments.

However, government procurement cycles, which can take between 12 and 18 months, remain a challenge for startups that require regular cash flow to sustain operations.

The report further highlighted the shortage of specialised AI talent across the continent, noting that only about one to three per cent of global AI talent is based in Africa.

About 51 per cent of surveyed founders said they struggled to recruit senior AI researchers, while 33 per cent reported difficulties finding machine learning operations engineers.

In response, some businesses have adopted internal training programmes to develop local talent, with 47 per cent of surveyed companies building training pipelines to improve the skills of junior employees.

Meanwhile, the report called for targeted interventions from investors and policymakers to address the sector’s challenges. It found that 62 per cent of surveyed businesses wanted research and development and computing subsidies, while 49 per cent called for harmonised cross-border data regulations to enable businesses to operate across African markets.

A further 26 percent wanted regulatory testbeds enabling them to build solutions with public-sector data in fields like healthcare and agriculture. These results indicate that AI progress in Africa will hinge not just on drawing investment, but also on enhancing access to infrastructure, specialised expertise, and markets that can sustain long-term business growth.