Meta Says Kenya's Digital Economy Could Hit $13 Billion by 2035. Here Is the Honest Assessment.

Meta has just dropped a major economic impact report for Sub-Saharan Africa, researched by public policy consultancy Public First. The Kenya country study estimates that Meta's platforms generated $140 million in economic activity for businesses in Kenya in 2025, and projects that Kenya's digital economy could grow from $4.2 billion today to $13 billion by 2035. That is a more than three-fold increase in under a decade, and Meta is placing its infrastructure, platforms, and open-source AI squarely at the centre of that story.

The numbers are eye-catching. But before we accept them at face value, it is worth asking: what is the report actually saying, who is saying it, and what stands between Kenya and that $13 billion figure?

What the Report Actually Says

The report examines Meta's role in digital infrastructure, digital platforms, and open-source AI across Sub-Saharan Africa. It includes country studies for Kenya, Nigeria, South Africa, and Côte d'Ivoire, and sets out how Facebook, WhatsApp, Instagram, Meta AI, Llama, No Language Left Behind, PyTorch, and other tools are being used by businesses and developers.

At the regional level, Public First estimates that Meta's digital infrastructure investments generated $16 billion in economic value across Sub-Saharan Africa in 2025. The research links this to subsea cables, terrestrial backhaul, and edge infrastructure, as well as the use of Meta platforms by small and medium-sized enterprises.

The specific infrastructure bet is significant. The report says the 2Africa subsea cable will deliver up to 180 Tbps of capacity and could bring 90 million additional Africans online by 2035. Public First says the additional capacity could support real-time telemedicine, cloud-based education tools, AI applications requiring low latency, and remote work.

On the open-source AI side, Kenyan startups are already building real products on Meta's Llama models. Upeo Labs, a Nairobi-based generative AI startup, powers its Somo-GPT education tool on Llama 3 and Llama 3.1. Somo-GPT serves as a multi-subject teaching assistant for high school students and is currently in closed beta across 500 schools in Kenya. Jacaranda Health's PROMPTS system is also cited in the report as a Kenyan example. The AI-powered SMS service gives pregnant women and new mothers care information, triages risk, and connects users to nurses or referrals.

These are not hypothetical use cases. They are live products solving real problems, built on free infrastructure by Kenyan developers who could not have afforded to build equivalent models from scratch.

Read This Report With Clear Eyes

Here is the first thing to understand: this is a commissioned report. Public First conducted the research, but Meta paid for it. That does not automatically make the findings wrong, but it does mean the projections are best understood as scenario modelling rather than independent forecasts. The projections are dependent on favourable policy, infrastructure investment, and adoption trajectories rather than guaranteed outcomes.

The report itself acknowledges that achieving this growth will require the right operating environment, and identifies three major barriers: connectivity constraints, high capital costs for entrepreneurs, and limited access to advanced technologies like AI

In other words, the $13 billion figure is not a prediction. It is a ceiling that Kenya could reach if everything goes right. The distance between that ceiling and where Kenya lands in 2035 depends almost entirely on how well it handles those three barriers, and at least one of them is currently being actively complicated by Kenyan lawmakers.

The AI Bill: Regulation With Good Intentions and Sharp Edges

Kenya's Artificial Intelligence Bill 2026, sponsored by Nominated Senator Karen Nyamu and currently before the Senate, is the country's first comprehensive attempt to regulate the lifecycle of AI systems. The immediate concern is timing: a major report drops projecting AI-driven economic growth of $13 billion, and simultaneously, a bill is working its way through Parliament that could raise the cost and complexity of building AI products in Kenya significantly.

The bill has genuinely important provisions. Any chatbot or automated interaction system must disclose to the user that they are interacting with an AI, not a human. That is a reasonable consumer protection. The deepfake provisions are direct: anyone who generates or distributes AI-created content using another person's image, voice, or likeness without consent faces a fine of up to Ksh 5 million, up to two years in prison, or both. Political deepfakes are explicitly addressed and carry the same liability. With Kenya's 2027 elections approaching, that provision addresses a concrete and documented threat.

The problem is not with those provisions. The problem is with the institutional design and some of the compliance requirements that follow from it.

The bill proposes three new government bodies: the Office of the Artificial Intelligence Commissioner, the Artificial Intelligence Authority, and the Artificial Intelligence Advisory Council. Kenya already has the Office of the Data Protection Commissioner, the Communications Authority, and the Kenya Information and Communications Technology Authority, all of which have existing mandates that overlap with parts of what the AI Commissioner would do.

More critically for developers, the bill requires that developers produce audit trails of how an underlying model was trained, what data was used, and how specific decisions were reached. For a Kenyan developer who downloaded a Hugging Face model and fine-tuned it for a local use case, that information does not exist in a form they can produce. The training data decisions were made by researchers at Meta or Google, not by the Kenyan developer deploying the model.

This is not a theoretical problem for a handful of edge cases. It is the standard operating reality for most Kenyan AI developers. The Upeo Labs team building Somo-GPT on Llama did not train Llama. They adapted it. Imposing criminal liability for compliance failures that are technically impossible for that class of developer would not make AI safer. It would make Kenya a less attractive place to build AI products, at precisely the moment the country is trying to compete for that investment.

The underlying concerns driving the bill are legitimate. The question is whether this particular design is the right instrument for addressing them, or whether it needs refinement before it passes.

The Bigger Problem: Nobody Has Found the Business Case Yet

Even if Kenya resolves the regulatory question perfectly, it still faces the challenge that every market in the world is currently wrestling with: AI's economic promise has not yet been matched by AI's economic performance.

According to MIT's GenAI Divide study, which analyzed 300 public AI deployments alongside executive surveys, 95% of generative AI enterprise pilots failed to deliver measurable financial returns within six months. IBM's research found that only 25% of AI initiatives deliver expected ROI, and only 16% have scaled enterprise-wide.

Average enterprise AI spending is projected to jump 65% from about $7 million per company in 2025 to $11.6 million in 2026, even as most firms cannot demonstrate a return. Companies are spending more because competitors are spending more, not because they have found clear evidence that it works for their specific situation.

For Kenya, this dynamic plays out with an added layer of pressure. A business in Nairobi that spends aggressively on AI tools that deliver no measurable return has far less cushion to absorb that loss than a Fortune 500 company in the same situation. The margin for error is genuinely smaller.

But here is where the Kenya story diverges from the global one in an interesting way. The ROI problem for AI is most acute when businesses try to apply AI to vague mandates: "use AI to improve operations" or "leverage AI for customer experience." The use cases with clear, measurable returns are narrow, well-defined, and matched precisely to AI's actual capabilities.

Kenya does not need to compete with OpenAI or Anthropic at the model layer. What it can do is build at the application layer, deploying these models against problems that are specific to the African context and where no equivalent solution currently exists: agricultural AI tools that help smallholder farmers optimize planting decisions, healthcare diagnostic tools that extend the reach of doctors in areas where the doctor-to-patient ratio is critically low, and financial services tools that deepen credit access in a market where M-Pesa has already demonstrated the extraordinary demand for mobile-first financial infrastructure.

These are not generic AI use cases. They are specific, high-value problems that existing global AI products have not been built to solve. That gap is Kenya's real competitive advantage in this moment.

So Is $13 Billion Actually Possible?

The honest answer is: possibly, but not automatically, and not under current conditions.

Kenya has real foundations to build on. M-Pesa has lifted access to banking from 26% in 2006 to 85% by 2026. In 2025, Kenya secured $1.04 billion in tech investment, a 72% year-on-year surge. Kenyan startups raised $638 million in 2024, the highest total in Africa that year. The developer ecosystem is active, the infrastructure is improving, and open-source AI tools like Llama are genuinely accessible to builders who could not have afforded proprietary alternatives.

The government is committed to growing the digital economy and is in the final stages of developing data governance policies, with AI policies anchored on three key pillars: AI infrastructure, data, and research, skills development and innovation. That alignment matters.

But government commitments and a commissioned report from a company with a direct interest in the outcome are not the same as a roadmap. The $13 billion figure requires Kenya to solve connectivity gaps in rural areas, maintain a regulatory environment that does not price developers out of the market, build the data governance frameworks that AI depends on, and successfully translate AI tools into measurable economic returns in sectors like agriculture and healthcare, at scale.

Meta has launched in-country AI accelerator tracks in Nigeria, Kenya, Senegal, and South Africa to accelerate the development of open-source AI solutions through Llama. That is a real investment in the ecosystem. But accelerators produce early-stage companies. Turning those companies into the kind of scaled, revenue-generating businesses that drive GDP growth takes years and a supportive policy environment.

The Verdict

Meta's report is a marketing document with genuinely useful data inside it. The infrastructure investments are real, the SME numbers are credible, and the application-layer use cases it highlights reflect what Kenyan developers are actually building. The $13 billion projection is a legitimate scenario, not a fantasy.

But scenarios are not forecasts. The gap between where Kenya is now and where the report says it could go is filled with policy choices, regulatory decisions, infrastructure investments, and the hard work of building businesses that actually deliver returns rather than just deploying technology because it exists.

Kenya's AI Bill, in its current form, risks making that gap wider. The ROI problem facing global AI adoption is not Kenya's fault, but it is Kenya's challenge to navigate. And the timeline matters: if the regulatory environment gets sorted in the next year or two and AI inference costs continue falling, Kenya's position as a mobile-first, developer-active market with specific, unsolved problems could make it one of the biggest beneficiaries of the AI wave.

If the bill passes in its current form and stacks compliance costs on the developers who are building exactly the kind of application-layer tools the report is counting on, that opportunity narrows considerably.

The $13 billion is possible. Whether Kenya gets there depends less on what Meta builds and more on what Kenya decides.