Is Kenya Ready for AI?

The Case for Ready

Start with the hard numbers, because numbers don't care about the narrative.

AI adoption in Kenya is outpacing the entire continent. By July 2025, 42.1% of Kenyan internet users aged 16 and older were actively using AI tools like ChatGPT. Compare that to South Africa's 15.3%, Nigeria's 8.2%, Egypt's 9.8%. Kenya isn't just ahead. Kenya is lapping the continent. This isn't policy or infrastructure showing up yet. This is a grassroots adoption rate that's essentially a referendum: Kenyans are ready. We're already using AI. The question is whether our institutions can keep up.

Startup funding is flowing. In 2024, Kenyan startups raised $638 million, nearly 29% of all African venture capital. That's not total equity; that's the single largest concentration of startup cash in Africa, landing in one country. Venture capital is the most unforgiving judge of opportunity. If Kenya wasn't ready, the money wouldn't be here. Instead, the market is shifting. Climate tech and agri tech are now competing with fintech for investment dollars. That shift only happens when the underlying infrastructure can support it.

M-Pesa is the killer advantage nobody talks about anymore because everyone takes it for granted. M-Pesa isn't just payment. It's data infrastructure. Sixty million transaction records. Real-time settlement. No middleman. That's what AI in Kenya will be built on. Every fintech innovation, every agricultural AI system, every healthcare platform that needs instant payment reconciliation is going to run on top of a system that already processes more transactions per day than most countries handle in a month.

Fiber internet is expanding. Kenya's national fiber network reached 13,590 kilometres in 2025, up from 8,900 in 2022. Internet penetration is at 48% of the total population. These numbers sound small if you're from California. They're revolutionary if you've been watching Kenya's digital infrastructure for the past decade. The expansion is uneven (concentrated in urban areas and major routes), but it's moving. And in the sectors where AI matters most for Kenya (agriculture, fintech, healthcare), the fiber is already reaching the key nodes.

Government backing is real and it's happening now. Last year in March 2025, Kenya published its National AI Strategy 2025-2030. This isn't vaporware. The strategy identifies healthcare, agriculture, education, finance, and public service delivery as priority sectors for AI deployment. The fact that the government has a stated strategy means something concrete: policy frameworks are being built to support AI development right now. That's the unglamorous infrastructure that makes everything else possible.

What the Strategy Actually Says

In March 2025, Kenya published the National AI Strategy 2025-2030. Not this year. Not recently. A year ago. It is the first time the country put the full picture in one document, and most of the conversations happening right now about Kenya and AI are happening as if the document does not exist. The numbers inside it are more specific than the general conversation about AI readiness, and they tell a more honest story than most headlines do.

The strategy is built around seven themes. AI Digital Infrastructure: building the compute and data centre capacity, expanding fiber, adding green energy for AI workloads, and setting up the cybersecurity layer. Data: creating governance frameworks, enabling secure cross-sector sharing, and building quality training datasets for local AI. Research and Development: growing a local AI R&D ecosystem and getting universities working directly with industry. Talent: reskilling programs, AI education at every level from primary school to professional development. Governance: regulatory frameworks for ethical and responsible AI deployment. Investment: financing mechanisms, PPP structures, and VC access for local startups. Ethics, Equity and Inclusion: building AI systems that carry Afrocentric values, protect marginalised communities, and incorporate indigenous knowledge.

The Oxford Insights Government AI Readiness Index (2023) is the benchmark the strategy uses for Kenya's starting position. Government readiness: 40.19 percent. Innovation capacity: 48.8 percent, against Sub-Saharan Africa's average of 32.93 percent. Data availability: 44.44 percent. Infrastructure: Kenya ranks 101st worldwide. Only 25 percent of university graduates complete STEM. That last number is the quiet crisis inside the AI talent story.

The big-picture number: AI is projected to contribute 1.2 trillion dollars in economic value to Africa by 2030, a 5.6 percent GDP increase across the continent. Kenya's readiness is the highest in East Africa, but the gap between Kenya and global leaders is large. The strategy is honest about this. It does not pretend Kenya is competing with the US or China on AI right now. It frames the question differently: can Kenya build AI systems that solve African problems, capture African markets, and keep African value inside African institutions?

The honest answer in the document is not yet, but the pieces exist. The 22.71 million internet users, the 65 million plus mobile connections, M-Pesa's 60 million plus transaction records, and the 40,000 Kenyans who have already received AI training. These are real inputs. The 13,590km fiber target and the Microsoft and G42 one billion dollar digital infrastructure deal are real commitments. What has been missing is a coherent framework that connects these inputs into outputs. The strategy is attempting to build that framework. The question is whether Kenya's private sector moves before the framework finishes being built.

The Honest Tensions

But 'ready' breaks down the moment you get specific.

Brain drain is real and it's killing talent density. Kenya produces skilled developers. They leave. Microsoft, Google, Amazon are hiring Kenyans at salaries that no Kenyan company can match. We're losing the people trained to build AI systems to companies that are building AI systems somewhere else. ALX Africa's recent partnership with Claude (200,000 learners across the continent trained by a Socratic mentor powered by AI) is one attempt to build critical thinking at scale locally. But if those learners graduate into a job market where the best opportunities are remote work for foreign companies, we haven't solved brain drain. We've just trained someone for someone else's economy.

Power infrastructure remains unpredictable. This is the thing that never makes the glossy tech headlines. Nanyuki is Africa's Silicon Savannah. It's also a town where the power cuts without warning. AI systems need reliable electricity. You can't train models on intermittent power. You can't run production systems that customers depend on if your electricity is unreliable. Konza Technopolis is 60km from Nairobi and it's supposed to be the hub. But Konza remains a long-term project, not a present reality. By July 2025, Konza was still in active development with partnerships signed with AfriLabs and Korea's electronics industry, but actual deployment remains years away. The infrastructure move happens slower than the policy move.

Internet costs are a hidden barrier. At 48% population internet penetration, Kenya has solid connectivity. But 'connected' and 'able to run AI applications' are different things. Fiber speeds and stability vary wildly. A Nairobi startup runs 100Mbps speeds. A regional town might get 10Mbps intermittently. And the cost of that connectivity is still high relative to incomes. For AI to scale across Kenya's economy, the cost of internet access needs to drop 50% from where it is today. That's infrastructure work, not innovation work. It will take time.

Skills gaps are real even with growing tech education. AWS trained 10,000 students at local universities. African Development Bank opened an ICT Centre of Excellence at US International University-Africa in Nairobi. These are good programs. But they're still producing graduates for a market where local opportunities are limited. A university student graduating from a rigorous CS program sees two options: work in Kenya for a lower salary building systems for a market that's still developing, or work remote for a US company earning US salaries. The brain drain isn't a failure of education. It's a rational economic choice by people who got smart and saw where the money is.

Who's Actually Building

The gap between 'ready' and 'actually building' matters more than either state alone.

M-Pesa's migration to an AI-powered platform shows that it's not hypothetical. Production AI systems are running in Kenya right now. Safaricom chose to upgrade to cloud-native, AI-powered infrastructure because the cost-benefit was there. Not for the press release. For the operations.

Startups in fintech, agri-tech, and health-tech are building AI into their products now. Not 2027. Not after they raise Series C. Now. A climate-tech startup gets VC funding and the first question isn't 'should we add AI?' It's 'how do we build AI into the core product from day one?' The market is already operating with that assumption.

Konza Technopolis is building a Digital Media City for creators and digital innovators. The timeline is slow, but the direction is clear. Kenya's government is not treating AI as a speculative future thing. It's treating it as infrastructure planning.

The Real Question

So is Kenya ready for AI? Yes. The infrastructure is sufficient, though uneven. The talent exists, though it's leaving. The market is adopting, though unevenly. The government is committed, though bureaucracy moves slow.

But 'ready' is a passive word. It suggests something external has to happen for Kenya to be ready. That's not true. What matters now is whether Kenya builds AI for Kenya, or whether Kenya becomes a data mine and a customer base for AI built elsewhere.

The difference between Rwanda and Kenya isn't readiness. Rwanda locked in a three-year partnership with Anthropic for government-wide AI deployment. Rwanda said 'we're building with Claude.' Kenya said 'we're publishing a strategy.' One is readiness. One is readiness plus commitment.

Kenya is technically ready. Kenya's AI future depends on whether we choose to build it ourselves or wait for someone else to build it for us.

What Kenya Should Do Right Now

If the question is 'are we ready?' the follow-up is 'ready for what?' The strategy gives Kenya a framework. These are the concrete things that matter most in execution, not in 2030.

Build the fiber backbone first. The strategy's infrastructure theme targets 13,590km of national fiber and connectivity for all 1,450 sub-counties. That target matters because AI applications in agriculture, healthcare, and education won't scale if the underlying network is unreliable. This is not sexy infrastructure work. It is the foundation everything else depends on. Rural areas and secondary cities are still underserved. Push fiber expansion in agricultural zones before worrying about AI model selection.

Lock in talent through incentives, not patriotism. The strategy identifies brain drain as a structural problem. Telling developers to stay in Kenya because 'it's our country' is performance. The strategy's talent theme calls for reskilling programs and equitable AI access. The private sector translation: pay competitively, offer equity in companies people believe in, and build products they are proud of. That is how you keep talent. Not guilt. Meaning.

Fix the data problem before worrying about the model problem. The strategy's data theme is direct about the bottleneck: government data is siloed, not digitised, and unavailable for local AI model training. Healthcare records, agricultural data, financial records. These are the inputs that would let Kenya build AI systems for Kenyan problems. The strategy is building a national data governance framework. The private sector should be pushing hard for access to anonymised public datasets as soon as those frameworks are live.

Move faster on Konza. The strategy designates Konza Technopolis as Kenya's AI hub. It is a Vision 2030 project, and that timeline is too slow. If Konza is going to be the hub for AI compute infrastructure, it needs to be functional, not aspirational. Government should front-load the investment and get the first data centre fully operational. Geothermal energy gives Kenya a genuine cost advantage for AI infrastructure. That advantage does not last forever if the facility stays on a long-term planning horizon.

Use government as anchor demand. The strategy explicitly calls for government to act as an AI customer, not just a regulator. The Investment theme and the PPP framework are built around this. Government procurement of AI systems from local vendors is the fastest way to build a revenue base for Kenyan AI companies that are not yet large enough to compete for international contracts. Rwanda did this deliberately. Kenya should follow with sector-specific AI procurement targets in healthcare, agriculture, and public service delivery.

Where Firms Like Kaara Works Fit

The strategy's stakeholder map is explicit. Consultancy firms sit in the Developers category, marked high interest, high influence, must be involved. That is not a polite acknowledgement. It is a structural position. The PPP framework is designed to route government AI procurement through implementation partners who can translate policy intent into working systems.

The priority sectors for AI deployment are healthcare, agriculture, education, finance, and public service delivery. These are not abstract categories. They are the sectors where professional services firms in Kenya already operate. An accounting firm automating audit workflows. A healthcare consultancy building patient triage systems. A legal firm deploying document intelligence. The strategy is explicitly building a framework where these engagements become government-supported projects with defined procurement pathways.

The power imbalance the strategy names directly: big tech companies operating in the region wield disproportionate influence and corner opportunities compared to SMEs. The explicit aim of the PPP framework is to create structural entry points for Kenyan consultancies and SMEs that are building locally. Firms that have production AI systems running now, for real clients, generating real revenue, will be positioned to step into those entry points when the frameworks solidify. Firms that are still at the whiteboard stage will not.

The other lever is regulatory timing. Kenya does not yet have AI-specific legislation. The governance theme in the strategy is actively building those frameworks now. Firms that build compliant AI systems today, before the regulations land, will have a first-mover advantage when compliance becomes mandatory. The cost of retrofitting an AI system for regulatory requirements is always higher than building with those requirements in mind from the start.

For Kaara Works, this is not abstract positioning. We build AI systems for professional services firms in Kenya. The strategy has just made the official case for why that work matters at a national infrastructure level. The window to establish a credible position in this market is open. It will not stay open once the procurement frameworks are in place and larger players arrive with proposal teams and government relationships already in hand.

The Honest Verdict

Kenya is ready for AI. Kenya is not ready to lead with AI without making deliberate choices today about whether we're building for Kenya or just allowing AI to be built for everyone else using Kenya as data source and customer base.

The infrastructure works. The talent exists. The market is adopting. The question now is whether Kenya's builders, government, and investors will commit to building AI systems for Kenya's economy, or whether we'll spend the next five years celebrating the AI readiness we have while watching the value go somewhere else.

That choice is still ours. But it won't be for long.