How Are Meta’s AI Ventures Tapping Into Previously Untouched Markets?

Meta has joined forces with Data Science Africa to direct its open source AI model, Llama, into areas where the technology has rarely reached. The partnership announced the winners of the Llama Impact Grant at the United Nations General Assembly’s Unstoppable Africa 2025 event. Each of the chosen startups and researchers will receive $20,000 in funding, technical mentoring and access to key networks.

The winners are working across health, education, agriculture as well as digital accessibility. South Africa’s Vambo AI is building infrastructure to support translation and transcription in more than 60 African languages, while PropelMapper is creating podcasts and farm reports to help farmers improve productivity. In Nigeria, Radease is equipping medicine vendors with WhatsApp-based AI to deliver trusted health information.

Uganda’s TeenApp is offering young people accurate and responsible sexual health resources, and Rwanda’s Easy Read Africa is simplifying complex documents for people with cognitive and learning challenges. Meta said its family of Llama models has been downloaded over a billion times, and the latest version, Llama 3.3, is free for organisations to adapt and use.

The grant programme began globally in 2023 and has since drawn more than 800 applications from over 90 countries. Past finalists such as Digital Green and Jacaranda Health showed how AI could support smallholder farmers and maternal health. Meta has described the African projects as examples of AI supporting solutions that are practical and scalable in underserved regions.

“We received an incredible number of applications this year, reflecting the vibrant and growing AI ecosystem across Africa. These projects exemplify the spirit of innovation and impact that the ‘Llama Impact Grant for Startups and Researchers’ stands for. We are excited to support their journeys and look forward to seeing the positive change they will bring to their communities and beyond,” said Sherry Dzinerova, Director of AMET Public Policy, Programmes, Campaigns and Product, Meta, during the announcement at GABI.

 

 

Why Is Meta Extending Llama To Security Partners?

 

In a separate venture, Meta has expanded access to Llama for national security use among its closest US allies. The company said it has opened up its open source models to France, Germany, Italy, Japan, South Korea, NATO and EU institutions. These countries join the Five Eyes partners, which are Australia, Canada, New Zealand, the UK and the US, already use the model.

Meta explained that governments prefer open source systems like Llama because they can host and work on them in secure environments without sharing sensitive data with outside providers. The models can also be adapted for classified purposes and deployed directly in the field. For US agencies, Llama has supported decision making, operational efficiency and mission specific projects.

One project involved the US Army’s Combined Arms Support Command. Meta said the pilot tested how AI together with augmented and virtual reality could help speed up routine repairs and return equipment to service faster. Such uses highlight how AI can serve practical needs in the defence sector while staying within legal and ethical rules.

Meta said that Llama cannot be used for warfare, nuclear industries or espionage. When it launched Llama in the US last year, the company said it was in the interest of the democratic world for American AI models to take the lead over rivals from China. Reports from Reuters had suggested Chinese researchers had reworked an older version of Llama to support their own military.

 

What Does This Say About Meta’s AI Plans?

 

Together, the ventures show how Meta is tapping into new regions and user groups for its AI models. In Africa, the focus is on helping startups and researchers build practical solutions for daily needs in farming, education and healthcare. In Europe and Asia, the extension of access is tied to security and defence partnerships.

Meta is working with companies such as Accenture, Microsoft, IBM, Palantir, Amazon Web Services and Lockheed Martin to develop Llama-based systems. The company described these collaborations as a way to help its allies gain better perception and decision-making tools. Some of the projects involve wearable products developed with Anduril.

Since its launch, Llama has become one of the most downloaded open source AI models in the world. Meta’s grant schemes and its security work show how it is using the model both for community-driven problem solving in regions like Sub-Saharan Africa and for defence use among its closest political partners.