Two FINCA investee companies offer a practical look into what’s necessary to support AI success in clinics and classrooms.
At the Africa AI Village held in conjunction with the 2025 G20 Summit in Johannesburg, a familiar refrain echoed across panels: Africa trails the world in AI computing power and must catch up by building local data centers and AI models. The statistic was clear — Africa accounts for less than 1% of global data center capacity — and so was the implied roadmap: Hardware first; everything else later.
However, that’s not the story we hear from the companies we work with every day. FINCA’s impact investing arm, FINCA Ventures, backs companies across Sub-Saharan Africa that are already running AI‑enabled business models without waiting for a new generation of local supercomputers. Their experience suggests that the question isn’t how Africa can catch up on computing capacity but rather what actually enables AI to work in the contexts where it’s needed most. And that it’s time to reframe the investment narrative around how to support AI applications that are creating real value across the continent today.
Funding patterns are already pointing in this direction. Partech’s 2025 Africa Tech VC report shows investors directing capital to businesses applying AI within essential services like health and education, where improvements in delivery and productivity are already visible. The momentum is behind tools and systems that make AI usable in clinics, classrooms, and daily workflows, not large, up-front infrastructure projects.
To get a clear picture of what this looks like on the ground, we spoke with two of our portfolio companies: Penda Health and Rising Academies, both of which are successfully leveraging AI within the constraints that define African markets.
The View From the Front Lines
“I’m quite bullish on what AI can do,” says Robert Korom, Chief Medical Officer at Penda Health, which partnered with OpenAI to build an AI-powered clinician copilot to support its network of 24‑hour primary care clinics in Nairobi. “We use it across the business — from clinical decision assistance, to patient engagement over WhatsApp, to internal operations — because it helps us deliver better care with the staff and resources we have,” he explains.
Rising Academies, an education company operating in West and East Africa, takes a similarly practical approach with two AI-enabled tools powered by Anthropic’s Claude models. Rori, a virtual math tutor, offers bite‑sized lessons, practice exercises, and personalized feedback for students, while Tari provides teachers with structured guidance and daily instructional resources. Both are designed with lightweight interfaces and optimized for the realities of low-bandwidth connectivity and shared devices.
None of this requires an AI supercomputer sitting next door. “AI is almost unique in how little local infrastructure it requires,” Robert notes. “A community health worker with a low‑cost smartphone can plug into an API. The processing happens in a data center somewhere else, and the value shows up on the phone in her hand.”
So what do these companies need to be able to use AI even more effectively? This is what they told us:
Power First, Always
Reliable electricity is the operating baseline: the thing that keeps healthcare centers open and school systems online long enough for any digital tool to matter. And its absence multiplies the cost and complexity of deploying AI in clinics and classrooms.
“Every facility needs a big generator. That’s upfront capital and a daily operational headache,” Robert explains. “It’s not that the barrier is insurmountable — it just makes everything more expensive.” Generators are also noisy and polluting, adding to their undesirability.
UNDP has framed universal energy access as a prerequisite for any meaningful AI future in Africa, noting that 50% of the population currently doesn’t have reliable electricity. They make the case plainly, noting “You cannot build intelligent systems on unstable grids; but the inverse is equally true: build AI-ready infrastructure, and you accelerate development on every front.”
Connectivity That Doesn’t Blink
Connectivity has improved across much of Sub‑Saharan Africa, but for companies using AI in daily operations, dependability still matters more than raw coverage. Even brief interruptions can derail applications intended to support day-to-day service delivery.
Rising Academies has designed its systems with that reality in mind. Rather than assuming stable network availability, the company builds for intermittent connections. “We design responsibly for limited bandwidth and low‑cost use, pairing a simple front end with a high‑tech back end,” explains CEO Stephanie Dobrowolski.
NextBillion has also argued that connectivity cannot be treated separately from energy. Its concept of connected power recognizes that networks depend on electricity just as much as electricity projects increasingly depend on digital demand. When the two are planned together, infrastructure is more durable and service delivery more reliable.
Digitized Records That Sync
AI tools only add value if frontline records are consistently digital and updateable in daily workflows. In Africa, many of the basic social systems are still largely analog, meaning the raw material that AI needs to be useful often exists only on paper or not at all.
When Penda first launched, patient histories were paper-based, fragmented, and inconsistent across clinics. Rather than immediately deploying AI tools, the company spent years digitizing basic workflows, training staff to record data consistently, and ensuring records synced reliably network-wide. Only then could Penda layer in AI to support decision‑making, automate follow‑ups, and improve clinic operations.
Brookings’s Foresight Africa 2025-2030 report underscores that AI delivers value only when usable, well‑captured data exists; without it, AI remains disconnected from last-mile service delivery. Policies promoting standardized, digital record-keeping across essential sectors like health and education would spare each provider from having to repeat Penda’s years-long digitization effort and free companies to focus on scaling their impact.
What Success Looks Like
The experiences of Penda Health and Rising Academies point to where AI ultimately matters most in Africa: at the last mile, on basic phones, and inside routine workflows. That’s where practical use is growing, and where the impact is clearest.
Picture a teacher in Sierra Leone preparing a lesson. Despite the school’s faulty connection, she opens Tari on WhatsApp, snaps a photo from her teacher’s guide, and gets a set of differentiated exercises tailored to her students’ ability levels. Once the work is completed, she snaps more photos, and Tari gives her a clear sense of where the class needs support, helping her adjust tomorrow’s lesson with confidence.
Or a nurse in Nairobi guiding a patient through triage on her phone. Penda’s decision support suggests a pathway, a follow‑up message is scheduled automatically, and the data goes into an electronic health record instead of a paper folder. The primary internet service falters, the backup link catches, and the clinic keeps moving.
These scenes are not hypothetical. They’re the frontier where AI is already creating value on the continent — despite thin infrastructure — because founders designed for the conditions that actually exist. And they’re the opportunities where investors can help write the next chapter of Africa’s AI success story.
