A fused studio and lab that researches, builds and industrialises frontier ventures — from research bench to factory floor.
The next industrial era — AI, robotics and the infrastructure beneath them — is being built now. Africa has the talent, the markets, the materials and the knowledge to build its share. What decides who benefits is who owns it. Rented technology sends value elsewhere; owned technology keeps it compounding where it's created. So we build, and we own — with the partners and local builders who build alongside us.
We don't move in a straight line. A private, agent-based AI system — running on hardware we own — lets us run many bets in parallel and collapse to the one reality rewards, so a small team explores like a lab many times its size, and backs what validates, fast.
A lab that fits in a room and thinks like a hundred.
And we don't build speculatively — we build with the people who live the problem. See how we co-build →
Africa is where we build and who we build with — and a deep well of knowledge: mathematical traditions, indigenous ecological and agricultural practice. We treat that knowledge as a hypothesis source, validated by our own scientific rigour before anything ships.
AI puts new power in everyone's hands — including those who would misuse it — and brings a new class of threats. So security comes before features: it is the first question we ask of everything we build, and one we never stop asking.
We build to own — with the partners and local builders who build alongside us. Expertise stays close to the work, compounds through practice, and grows where value is created.
We do not move in a straight line. We run many bets in parallel, then collapse to the one reality rewards.
We build with an eco-responsible mindset from day one. We design for loops, not lines: materials, energy, data and value are kept in motion for as long as possible. Waste-to-product is one expression; the principle is broader — build systems that regenerate instead of deplete.
We take ideas through the whole chain: model, train, prototype, prove, and deploy. The work is not finished until it meets the floor.
We embed in real operations, learn what breaks, and ship what works. Client floors become our classroom; shared problems become products; the strongest patterns become ventures.
Live today, Lilenek is the e-commerce infrastructure we built for the African market — designed to connect commerce, transactions and intelligence in one operating layer. Its strength is a set of features made for how this market actually trades — and one of them turns waste into worth: bought through stablecoin rails, looped back into product. It earns its place twice: useful for the businesses using it, and valuable in regions carrying more waste than they can process.
Real infrastructure, already running — the clearest proof that what we promise, we build.
Nothing is lost, everything is transformed.
We own the whole stack
Our AI runs on hardware we own and operate, across multiple regions including the continent. Data and infrastructure sovereignty is invoked far more often than it is built — so we built ours. It decides four things most companies hand to a supplier:
Where our data and our clients' data lives, and who can reach it, is our decision — not a supplier's.
Protecting our models, agents and data is ours to answer for — never left to a supplier's default settings.
Compute we own is bought once, not metered forever — the difference between running experiments and running a business.
No price change, policy shift or supplier decision can switch off what we are building.
Most AI companies rent their compute — every invoice sends value somewhere else. Ours doesn't leave. Built because we needed it; the venture that opens it to others is next.
E-commerce infrastructure for the African market, designed to connect commerce, transactions and intelligence in one operating layer.
View site ↗A next-generation street-football community where teams compete anywhere, anytime.
View site ↗Compute engines and AI inference on infrastructure we own — opening our private cloud to others. Name and launch date to come.
Coming soonAdaptive security for AI-era systems — apps, AI pipelines, agents and self-hosted AI, guarded from build to run. Its aim: no attack works twice. Name and launch date to come.
Coming soonMore launching soon
AI and robotics are the hard frontiers we're building — and we walk them deliberately. We train and adapt our own models, including the brains for our robots. We're finalising the training stage now, and already exploring the first prototype use cases: first floor, warehouse & logistics automation.
Model the task and the environment as a digital twin.
Teaching the robot brain in simulation, at scale — the stage we're finalising now.
Proving it on real hardware. We're already exploring the first use cases.
Take proven systems to the floor: warehouse & logistics first.
Working on a warehouse or logistics problem? We're choosing our first use cases now — we'd like to hear yours.
Let's discussDifferent levels of commitment — and we're explicit about who owns what.
Our engineers embed in your operation and ship. You own the output.
Owned with partners and local builders.
You bring the problem, the floor and the domain expertise. We bring the AI, the engineering and the build, with security designed in from day one — then we prove it in your environment.
A real, costly problem on your floor — ideally one shared across your whole industry.
Your domain expertise, our AI and engineering. Built for your environment, designed to generalise.
A pilot in real operations, with success measures agreed up front.
Agreed terms, time-boxed: a share of the revenue, a head start on your competitors, and/or free access at our highest tier.
You solve the problem you live with. We take the solution to the industry.
A problem to solve, a venture to back or a team to join — whichever way in, the door is the same.