The Frontier Has Moved: Where We See It
Africa is missing at frontier AI development. The talent is here, but very few chose to build at the level where the field is still being shaped. We are making that choice.
We’re building an LLM from scratch in Nigeria. We just don’t talk about it often.
Here's why.
You’ve probably come across Nexenda and seen our focus on “frontier” AI research and product development.
But “frontier” is often misunderstood.
For some, it means the biggest models. For others, the most capable systems. For a few, it’s benchmark performance. But to us at Nexenda, it means something even more precise.
Pretraining a large model is capital intensive, infrastructure heavy, and technically demanding. Incredible teams have spent years solving it. And they have. The field now knows how to do it.
We are doing it too.
But that is not where the hardest problems are anymore.
The frontier has moved.
The frontier is safety.
The frontier is alignment.
The frontier is: how do we build AI systems that remain safe, steerable, and aligned with human values as they scale toward AGI?
As Mustafa Suleyman, CEO at Microsoft AI recently framed it, the shift is toward capable, agentic systems that can act, not just respond.
But there's another real question:
What about the alignment challenges of these systems?
How safe are the underlying architectures?
How interpretable are the systems we are orchestrating and stacking increasingly powerful capabilities on top of?
That is where we are building.
We believe this is how Africa meaningfully participates in shaping the future of AI. Not as a consumer. Not as a late adopter. But as a co-creator, contributing at the layer where the field is still being defined.
There is important work happening across the continent. Indigenous language systems, culturally grounded AI, infrastructure, ecosystem building.
Teams like Lelapa AI, Cassava Technologies, Awarri, Crane AI Labs, Masakhane African Languages Hub are pushing critical boundaries.
That work matters. The ecosystem needs it.
But we are deliberate about what we mean when we say frontier.
It is not about model size.
It is not about parameter counts.
It is not even primarily about language representation and inclusion.
It is about solving the hardest unsolved problems in AI. And today, that problem is alignment.
We're building our own pretrained LLM from scratch. We also leverage open-source models where it makes sense.
But we didn't start at the transformer.
We started at the problem.
Our work spans original research in model architecture, datasets, mathematically grounded alignment and safety theorems, philosophical foundations, and novel training methodologies, all developed in-house by our global team.
Africa is missing at the frontier. Not because the talent isn't here, but because very few chose to build at the level where the field is still being shaped.
We are making that choice.
That is where our differentiation sits.
That is what frontier means to us.
We conduct research in AI safety, interpretability, alignment, model efficiency, and language processing — and translate that research into real‑world systems, innovative products, and ecosystem initiatives anchored in safety and long‑term human benefit.
We operate at the frontier where the field is still being shaped, tackling the most consequential challenges at the intersection of Africa realities and global needs.
To ensure the path to beneficial AGI is safe, steerable, and shaped by Africa.
Follow our social handles to stay updated as we build safe, reliable, and human-centered AI born in Africa for global benefit.
Linkedin | X (Twitter) | Instagram
Navigation: Home | About this Publication | Nexenda founders
Visit our website to learn more: https://nexendahq.com
Discover our approach: https://nexendahq.com/about
Learn more about our research: https://nexendahq.com/research
Join our team: https://nexendahq.com/careers
Stay updated: https://nexendahq.com/news
Follow our journey: https://nexendahq.com/journey-to-superintelligence