From Silicon Valley to the Global South: Who Owns the Future of AI?

In Silicon Valley, Mark Zuckerberg and Satya Nadella are planning for a world where machines write half of their code. In Africa, we’re still debating whether AI belongs in our schools. The question is, will Africa help shape the design and development of the next generation of AI? or are we just going to remain passive consumers of it ?

This week, Facebook's parent company, Meta hosted its first AI developer conference dubbed “LlamaCon” . The closing highlight was a thought-provoking 40-minute conversation between Meta’s Mark Zuckerberg and Microsoft CEO Satya Nadella. As I attentively listened to their deep dive into the frontiers of AI, discussing things such as distillation factories and a future where about 50% of the code will be written by AI agents, I found myself thinking less about what’s being built in Silicon Valley and more about what’s still missing here at home. It made me reflect deeply on where Africa stands in this rapidly developing AI revolution.

According to the State of AI in Africa by Strathmore University and various other reports, less than 15 countries have an AI policy or strategy in place, and of those, less than 10 have made significant public investments in AI research infrastructure.

Even though much of the Global South was still drafting its first AI policies, there are signs for optimism. AI is only beginning to get some traction across the continent. There many private initiatives on AI in Africa that have taken off recently, led mainly by start ups. However there is still a lot of work to be done.

To be fair, the fear and the caution is very much understandable. Data privacy, digital rights, algorithmic bias, and surveillance are serious concerns. But too often, the debates become an excuse for inaction. We are not training models here. We are not distilling knowledge systems. We are not building the pipelines that will determine who owns intelligence in the future. We are merely consuming.

At LlamaCon, Satya talked about how Microsoft is already seeing AI write 30% of the code in some of its internal projects. Zuckerberg touched on how Meta is projecting 50% of its future model development will be carried out by AI itself.

Microsoft is already seeing AI write 30% of the code in some of its internal projects.

Mr Nadella then mentioned how GitHub Copilot is no longer considered a novelty but a foundational productivity tool embedded in the developer workflow. In other words, all developers are now expected to use AI to help them write better, faster and elegant code.

Meta is projecting that 50% of its future model development will be carried out by AI itself.

For Meta and Microsoft, the conversation has moved beyond whether to use AI. They’re deep into the mechanics of how to build efficient AI models that use less power, fine-tune performance, how to orchestrate agents, and structure workflows. One recurring theme I found in their conversation was how they’re experimenting with multimodel reasoning, distilling large systems into lightweight tools, and embedding intelligence directly into everyday applications. It’s not fear or caution driving the agenda—it’s architecture, infrastructure, and scale.

In contrast, Africa’s most common AI-related headline is still about regulating facial recognition. Somethings will need to change very fast. What we’re facing isn’t merely a gap in tools or talent—it’s a deeper question of direction. As AI becomes the backbone of sectors like healthcare, finance, education, and agriculture, staying on the sidelines means giving up the power to shape the systems that will soon define us. The longer we delay, the more we risk locking ourselves into a future designed elsewhere, on someone else’s terms.