AI is moving fast – and a lot of money is following it. But while billions are pouring into innovation in the US, Europe or China, many low- and middle-income countries (LMICs) are being left behind. That’s a missed opportunity, especially as AI has great potential to tackle some of the world’s toughest challenges.

AI innovation in LMICs faces unique challenges but also opportunities. AI solutions can drive economic and societal progress by addressing specific local needs. From AI-powered sign language interpretation in Pakistan to real-time water management in Jordan’s refugee-hosting schools and climate smart farming in Africa, AI is already transforming lives.

Navigating the funding landscape

However, enterprises in these regions, whether AI-focused or not, often struggle to secure the necessary capital to develop and scale their innovations. The funding landscape is tough. For instance, in Africa, startup investment dropped by over 50%, from $650 million in Q1 2023, to $310 million in Q1 2024, while Southeast Asisa saw a 29% decline over the same period.

AI startups in LMICs are building in a challenging and uncertain investment climate. Investors are more cautious, demanding clear pathways to scale and negotiating valuations fiercely. The effects of this ‘funding winter’ are particularly acute in LMICs, where access to capital has always been more limited.

Initiatives like the GSMA Innovation Fund for Impact AI are vital. Backed by the UK FCDO, the Fund is designed to support AI startups driving social impact – offering capital, venture-building assistance – and to de-risk investment in LMICs. By backing local founders solving local problems, it aims to ensure AI tools are relevant, inclusive and sustainable.

Unique challenges in LMICs

AI innovation in LMICs isn’t just a funding issue – it is a matter of overcoming unique and layered structural barriers. While infrastructure and computing capacities are improving, high costs remain a significant obstacle. For instance, in some markets the price of essential AI hardware like GPUs, combined with limited high-speed internet access, hinders the development of AI solutions.

Additionally, AI models trained on global datasets often fail to capture local nuances. Many LMICs lack the vast amounts of region-specific data needed to train models that are truly relevant to their populations, particularly in areas like language processing.

Another challenge is the talent gap. Many universities struggle to keep up with industry demand, making it harder to train – and retain – AI-ready professionals. Brain drain also remains a real issue, with top talent moving to higher-income markets in search of better opportunities, in a global market for AI talent.

AI in action – transformative use cases

Despite these challenges, AI startups in LMICs are demonstrating their potential to deliver profound societal positive impact.

In healthcare, the Microsoft Africa Research Institute (MARI), in partnership with the Central University of Technology in South Africa, has developed AI models that predict malaria outbreaks with 99% accuracy. In Cape Town, startup Aerobotics uses AI powered drone and satellite imagery to help smallholder farmers detect pest infestations, analyse soil health, and optimise irrigation. In the Philippines – one of the world’s most climate-vulnerable countries – startups like Komunidad are using AI-enabled Disaster Risk Reduction systems to distribute early warning alerts to communities via mobile and social media, drawing on data from earth observation satellites and local sources.

These are just a few examples of what’s possible. And frugal innovation has a real potential to scale and replicate. Support and funding can ensure they are isolated success stories.

The path forward – unlocking potential in LMICs

To ensure AI can reach its full potential in LMICs, governments, investors and development organisations must remain steadfast in the following priorities:

• Back local innovators: Prioritise startups embedded in communities and focused on real-world challenges.
• Build AI-ready infrastructure: Improve internet access, address the usage gap, reduce hardware costs and expand access to cloud services.
• Nurture local talent: Align university and training programmes with industry needs to create a pipeline of AI professionals.
• Leverage blended finance: Combine grant funding and commercial capital to de-risk early-stage investment.
• Create an enabling policy environment: Shape national AI strategies with input from startups, NGOs, researchers and underrepresented groups.

Future proofing AI for all

If we want AI to reach the people and places where it can have the greatest impact, we need collaboration – between startups, governments, investors and funders. Support systems matter now more than ever, especially as uncertainty remains around private capital access.

Global governments, development aid organisations and private sector organisations must continue to identify pathways to unlock funding and support where it’s needed most – ensuring resources flow to the innovators building inclusive, sustainable solutions.

With just one in five SDG targets currently on track, the world cannot afford to leave this opportunity on the table. AI must be a catalyst for equitable, long-term progress everywhere.

The question isn’t whether AI can transform lives – but how we ensure it benefits everyone.

Read more on Artificial Intelligence.