Trust & Safety

Green Compute for Sustainable AI in Africa
Green compute is a strategy to expand access to AI systems in Africa while advancing a more resilient and sustainable digital ecosystem.

AI has the potential to drive economic growth and help tackle critical challenges in Africa, from improving healthcare delivery to expanding access to education. However, two major barriers are holding this progress back: lack of computing resources and the high energy demands of AI systems. Without reliable access to affordable electricity and computing power needed to train AI models, innovation remains out of reach for many. This entrenches the digital divide and affects the future of industries like energy, healthcare, and education.

Green compute is a practical approach to reduce the energy and environmental costs of the hardware and software infrastructure needed for AI. By tackling both compute scarcity and energy inefficiency, it provides a strategy to expand access to AI systems in Africa while building a more resilient and sustainable digital ecosystem.

The Impact of Compute Scarcity in Africa
A lack of computing resources is one of the biggest obstacles to advancing AI innovation in Africa. Compute – essential digital infrastructure like processing power, storage, memory, and networking – is the backbone of AI development. Yet, a study by Zindi, Africa’s largest AI talent network, found that only 5% of Africa’s AI practitioners have access to compute resources they need. Of those, only 1% have reliable on-site resources, like GPUs (graphics processing units) – critical hardware for training AI models. These resources are often expensive to maintain due to high electricity costs. The remaining 4% rely on cloud services, often using cloud credits. They typically spend about US$1,000 per month, which only provides around 2 hours daily usage on outdated hardware.

This scarcity slows progress. While an AI engineer at a start-up in Europe can tweak or retrain their models every 30 minutes, their African counterparts might have to wait up to six days to do the same. These delays don’t just stifle innovation but limit the development of solutions that could improve priority sectors – like energy, infrastructure, healthcare, and education – deepening global inequalities.

AI’s Energy Challenge
Expanding compute access to drive AI innovation in Africa comes at a steep energy and environmental cost. Training a 213 million parameter AI model emits 626,155 pounds of CO2eq emissions, the equivalent lifetime emissions of five cars. Once deployed, AI systems like ChatGPT require about 10x more electricity per query than a standard Google search. And these figures only cover training and deployment; the full AI model lifecycle also involves raw material extraction, manufacturing, and disposal, each stage adding its own costs and impact.

AI already accounts for 10-20% of data center electricity use, a figure projected to double by 2030. Data centers, which house digital services from AI to cloud computing and web hosting, now account for 3% of global energy consumed – double the amount used just a decade ago. Without intervention, these numbers could triple in the next five years, straining energy resources in regions like Africa where electricity is already scarce.

Powering AI: The Current Landscape
Global cloud providers like Google, Microsoft, and Amazon are grappling with AI’s rising energy demands. These tech giants, already some of the largest corporate buyers of renewable energy, have begun investing billions of dollars in nuclear power to maintain their businesses operations without putting their net-zero goals at risk and exacerbating environmental challenges.

Meanwhile, Africa faces a different energy reality entirely. Only half of the population has reliable electricity, with frequent blackouts forcing businesses and households to rely on costly, polluting private generators. Excluded from the electricity industrial revolution (1880-1950), many African countries still lack the infrastructure to meet basic energy demands, despite being among the most resource rich regions in the world.

Given AI’s high energy cost, green compute and sustainable energy infrastructure must be central to Africa’s growth strategies. These investments won’t just advance AI development but help build foundations for a more resilient, inclusive, and innovative digital ecosystem.

***

Green Compute: A Path Forward for AI in Africa
Green computing is an essential part of sustainably expanding AI access in Africa. Green compute maximizes energy efficiency and limits the environmental impact of the infrastructure required to deliver digital products and services. It has the added benefit of reducing operating costs and building resiliency. Limited electricity and water infrastructure on the continent create both an urgent need and opportunity for bold, innovative sustainability solutions.

Key players and their roles:

Public sector: creating the foundations for green compute.
National and local governments are central to creating conditions that attract investment in green compute. To build sustainable digital infrastructure governments must:

• Streamline policies and standards, for example for energy efficient systems, to create open and competitive markets.

• Standardize regulations across the regions to reduce barriers like cross-border data restrictions and high import fees.

• Digitize public services, choosing contractors that prioritize sustainability, to create demand for and encourage private in green compute infrastructure.

• Subsidize research and innovation in green technologies.

  Progress is already under way. Countries like Kenya and Rwanda, have introduced policies that attracted local and foreign investment in data centers. However, closing the gap with regional leaders like South Africa will require an estimated 700 new data centers and 1,000 megawatts of power. While this is a monumental challenge ahead, it also provides an opportunity to innovate in sustainable energy infrastructure and green compute.

Private sector: drive sustainability innovation.
As foundational infrastructure like electricity and internet improves in Africa, more companies are expanding investments in cloud computing and data centers. Currently only 15% of organizations and individuals in sub-Saharan Africa use cloud computing (compared to 71% in Europe), but, the region’s server investment growth is 13% higher than the global average. From foreign tech giants like Amazon, and Microsoft, to local telecom companies like Liquid Telecom and Seacom, the private sector is expanding their investments in the regional data center ecosystem and has a central role in driving sustainability innovation. Companies can:

• Build energy-efficient data centers and incorporate green compute principles at every stage of the AI lifecycle, from raw material extraction to chip design and model development.

• Leverage growing demand for efficient computing to innovate cost-effective solutions. For example, Nvidia’s Blackwell chips are designed with energy efficiency in mind, responding to demand for more cost-effective and sustainable components

• Develop industry standards like AI Energy Star ratings to help users choose the most sustainable model.

  Global policies like the EU AI Act, which will require companies to disclose energy consumption and encourage energy ratings for AI models, will further accelerate sustainability innovation .

NGO sector: orchestrate cross-sector collaboration.
NGOs play a critical role in ensuring that sustainability stays at the forefront of AI expansion in Africa. They can:

• Set standards and metrics for green compute practices.

• Establish green compute hubs to provide affordable, sustainable access to computing resources.

• Promote partnerships between governments, businesses, and civil society.

• Advocate for cleaner, more resilient energy sources to power data centers, advancing sustainable energy infrastructure.

  By focusing on long-term sustainability, NGOs can fill gaps left by businesses and governments.

Green Compute as a Catalyst for AI and Beyond
Investing in green compute is not just about making AI innovation more accessible in Africa, it’s a pathway for more sustainable, resilient, and inclusive digital future. Prioritizing energy efficient digital infrastructure can help African countries overcome some of the barriers of compute scarcity and unreliable electricity – essential components of any thriving digital ecosystem. With governments fostering the right conditions, companies pushing for efficiency, and NGOs ensuring accountability, African countries can become leaders in sustainable AI development.

These investments will not only enable AI innovation but strengthen other critical sectors like energy, agriculture, healthcare, and education, paving the way for long-term economic growth.