Access to frontier AI has mostly been a privilege of companies rich enough to buy it. A new programme wants to hand it to African founders working on harder problems. Novastar Ventures has partnered with Google to launch an Applied AI Lab for startups and researchers building AI solutions in healthcare, agriculture and education, with applications open since July 1 and the first cohort of five to ten teams to be announced in September.
What selected teams get
The package goes well beyond a typical accelerator. Participants receive early access to Google’s latest AI models, technical mentorship from Google engineers, Google Cloud credits and go-to-market support, with the lab backed by Google’s AI Futures Fund and supported by Google DeepMind and Google Research. Participants may also be eligible for equity investment and non-dilutive funding. Three more venture firms, Ventures Platform, 4DX Ventures and Norrsken22, join Novastar in providing mentorship and operational guidance, stacking four African-focused investors behind a single cohort.
AI for fundamentals, not productivity
The lab’s framing is its most interesting feature. Novastar co-founder and managing partner Steve Beck argues that where developed markets deploy AI mostly for workplace productivity and entertainment, African startups are pointing it at fundamentals: healthcare delivery, food systems, access to education. Novastar’s own portfolio sketches the pattern, with Penda Health using AI-powered clinical support to improve diagnosis and treatment, NewGlobe applying it to teaching quality across schools, and Agrails using it to connect smallholder farmers to climate risk insurance.
Why Google is here
For Google, the lab extends a deliberate African AI strategy through its AI Futures Fund, and the logic is not charity. The continent’s young population and expanding digital economy make it a proving ground for AI applications that global products were never designed for, and the startups solving those problems generate exactly the kind of use cases, data and talent relationships a model provider wants close. Direct access to DeepMind researchers is the part money cannot usually buy, and it is the clearest sign this programme is meant to produce working companies, not workshop certificates.
What founders should do with it
For African AI founders, the calculation is simple: this is currently among the most comprehensive support packages available on the continent, and the cost of applying is an application. The deeper takeaway is about positioning. The teams most likely to win selection, and to matter afterwards, are those applying AI to problems with life-sized consequences, because that is where Africa’s advantage is real: proximity to problems worth solving. The lab opens later this year; the September cohort announcement will show what Google and four VC firms believe that looks like.
