What do buyers wish to see in African AI startups?

Each Thursday, Delve Into AI will present nuanced insights on how the continent’s AI trajectory is shaping up. On this column, we study how AI influences tradition, coverage, companies, and vice versa. Learn to get smarter concerning the individuals, tasks, and questions shaping Africa’s AI future. Tell us your ideas on the column via this kind.

Globally, AI startups have attracted over $100 billion in funding within the first half of 2025, practically triple the $38.5 billion raised over the identical interval final 12 months. 

Africa’s share of that pie stays small, however is steadily increasing. Final week, I highlighted a string of notable offers from eight startups from the continent which have raised over $40 million this 12 months. 

On this week’s column, I spoke to a number of buyers who lately invested in AI-focused startups to search out out what sorts of AI-enabled options entice enterprise capital and their funding theses for a fast-evolving sector in a continent taking part in catch-up. 

Infrastructure gaps as a possibility

Throughout the board, buyers and stakeholders cite an absence of infrastructure and datasets as roadblocks to AI innovation on the continent. “We’re nonetheless lacking essential infrastructure like electrical energy and information,” mentioned Adjei Boateng, Chief Funding Officer at Accra-based funding agency, Black Star Group. 

These constraints will not be going to go away within the brief time period, but they affect how buyers select to deploy their capital.  

For Nairobi-based Enza Capital, nevertheless, infrastructure challenges will not be deterrents.  

“We’re most enthusiastic about AI use circumstances that handle Africa’s infrastructure gaps and create defensible worth via native context,” mentioned Mike Mompi, Managing Associate at Enza Capital. In February, the VC agency participated in a $3 million pre-series A spherical for Widebot AI, an Egyptian-founded AI startup centered on constructing enterprise-grade language options for firms and authorities establishments within the MENA area. 

Widebot plans to make use of the funding to construct an Arabic-focused massive language mannequin referred to as AQL Thoughts. Enza Capital believes firms that may create defensible datasets in contexts that world incumbents like OpenAI and Anthropic can not simply replicate will carry extra worth. 

“Going ahead, we’ll proceed to again firms that deal with the numerous constraints in capital, infrastructure, expertise, and information as design ideas, not obstacles,” Mompi mentioned.  

Capria Ventures, a International South-focused VC agency, takes a unique method. The agency chooses explicitly to not spend money on firms centered on constructing AI fashions and instruments. As an alternative, Capria prefers to spend money on African startups with utilized AI use circumstances, principally current firms at present embedding AI into their pre-existing fintech, jobtech, healthtech, or logistics options. The agency at present has six startups in Africa underneath its portfolio, together with Moniepoint, SeamlessHR, and Helium Well being. They’ve a devoted AI innovation crew that works intently with their startups to determine and develop worthwhile AI use circumstances.

“Investing in conventional AI firms is extra within the ballpark of huge buyers within the US, UK, and different developed markets,” Mobola Da-Silva, associate at Capria Ventures, defined. She worries about corporations constructing what she describes as “AI only for AI’s sake”. 

Atlantica Ventures, an African early-stage VC fund, takes a broader method to AI startup investments, believing that alternatives for AI startups can come at completely different phases. 

“We have a look at the total AI stack, from foundational fashions to infrastructure to customer-facing functions. We again founders constructing the place aggressive moats will be established,” mentioned Aniko Szigetvari, founding associate at Atlantica Ventures.

In February 2023, the agency participated in a $2.5 million enterprise spherical to again Lelapa, a startup constructing language fashions for low-resource African languages like isiZulu, Hausa, Yoruba, Afrikaans, and extra. In March 2025, in addition they backed Nosible’s $1 million pre-seed spherical. Nosible gives two principal merchandise: an API that produces real-time and historic information insights for companies, and specialised AI brokers designed to simplify duties in world asset administration. AI merchandise primarily designed for developed markets and imported to the continent may fail to handle these native circumstances, making localisation a aggressive benefit. 

“We can not merely transplant Silicon Valley AI fashions onto the continent,” Uwem Uwemakpan, head of funding at Launch Africa Ventures, mentioned. “I’m not saying we should always construct our massive language fashions; that’s going to be too costly. However we must be asking: how are we adapting and coaching fashions on information distinctive to Africa, and turning these into options with world functions?”  

In February, Launch Africa led a $1 million pre-seed spherical for ToumAI, a Moroccan AI startup centered on constructing speech recognition and conversational AI instruments tailor-made for African companies. 

Is there a proper enterprise mannequin?

For Mompi, there’s a must deal with balancing hype and noise within the AI house. Enza Capital is looking out for AI options constructed for particular industries with a excessive willingness to pay similar to agriculture, well being, and fintech, the place native context and information are every little thing. 

Szigetvari of Atlantica Ventures holds an analogous view. “AI is extra than simply LLMs. We see monumental alternative in utilized AI for chilly chain logistics, agriculture, cybersecurity, DevOps, and compliance. Fashions with out defensible IP, or these over-reliant on generic AI with out a market moat, are much less compelling.” 

Launch Africa’s Uwemakpan desires to see extra AI startups centered on B2B (business-to-business) and B2B2C (business-to-business-to-consumer) options. Basically, startups that promote to different firms, and people who promote to companies, which then serve the tip client. 

“The explanation for that is that there’s a stable construction round revenues and you’ll normally predict this, particularly for B2B startups, by signing long-term contracts to lock organisations and establishments right into a plan,” he talked about. “For B2B2C, there are decrease buyer acquisition prices.” 

Various methods for AI founders

Past the B2B and B2B2C mannequin, VCs are looking out for 2 different kinds of approaches to AI innovation. 

These centered on combining AI with human experience to ship automation and reliability, and people centered on frugal innovation. Extra particularly, edge computing, small language fashions (SLMs), and infrastructure optimised for gradual or unreliable web on the continent. 

Edge computing permits for information to be processed nearer to the place it’s generated relatively than sending it to distant cloud information centres within the US and Europe. This may help AI-powered instruments on the continent in sectors like healthcare diagnostics run sooner and extra reliably, even in rural areas. SLMs are compact, extra environment friendly variations of huge language fashions. They don’t want as a lot computational energy and are higher fitted to use circumstances the place information safety is a precedence. These instruments are more and more gaining reputation amongst firms on the continent. As an example, UBA Financial institution, a tier-one business financial institution in Nigeria, has inspired its in-house ‘Superior Analytics and AI’ crew because the begin of 2025 to develop these SLM instruments for software throughout completely different departments. 

Chris Maranga, regional director for East and West Africa at impression funding agency Acumen, believes that AI startups prioritising working in low-resource, high-constraint environments maintain promise. 

“We’re waiting for AI instruments that combine seamlessly with characteristic telephones and low-bandwidth environments as a result of that’s the place the biggest untapped person base is,” Maranga instructed TechCabal. 

Massive Tech can be watching

Massive Tech can be looking out for AI-focused startups on the continent. Google lately dedicated $37 million to help Africa’s AI innovation, together with its newest Google for Startups Accelerator Africa program. The cohort, introduced in June, consists of 15 AI-driven startups from seven international locations, together with Ghana, Rwanda, Nigeria, and South Africa. 

“The choice to incorporate AI-focused startups on this 12 months’s cohort displays a long-standing perception that Africa’s innovators are uniquely positioned to create options that may profit each the continent and the broader world,” mentioned Taiwo Kola-Ogunlade, Communications & Public Affairs Supervisor for Google West Africa.

The chosen startups will obtain the mentorship wanted to develop strong enterprise methods and refine their merchandise. They can even obtain coaching on AI improvement, $350,000 in credit on Google Cloud, and the chance to fulfill buyers and companions who will assist them scale their impression. Google plans to offer these sources with out fairness necessities to the startups. 

Uwemakpan believes buyers ought to take note of rising curiosity from bigger,  world gamers when investing in AI startups on the continent. “How are we investing in firms which might be strategically positioning themselves for acquisitions? When the worldwide manufacturers start to have a look at Africa, who will they probably purchase?” 

In June, Meta concluded a name for early-stage African startups growing domestically related AI functions that handle key sectors, together with training, healthcare, and agriculture. Profitable groups will obtain as much as $25,000 in equity-free funding to help the deployment of their AI options. This system features a six-week-long accelerator and 6 further months of post-program help. 

These investments, which can seem like philanthropic, can function an essential technique for securing Africa’s AI market, which is anticipated to be value over $16 billion by 2030. 

Sceptics nonetheless stay

Some buyers query whether or not these short-term-accelerator applications can have an effect on the continent’s evolving AI house. 

“What number of startups are you able to level to that went via accelerators that survived and succeeded?” requested David Lanre Messan, angel investor and founding father of enterprise studio, FirstFounders. 

The studio gives three-year-long applications that they imagine can unlock the potential of early-stage entrepreneurs to construct high-value AI-focused startups. Some startups at present underneath the FirstFounders portfolio embody PocketLawyers, a Nigeria-based startup offering AI-powered productiveness instruments for legal professionals and different purchasers with legal-related wants, and KorinAI, a generative AI music lab for Africans. 

Wanting forward

For now, the AI panorama continues to be in its formative phases. Betting small on the continent’s evolving house could shield in opposition to threat. However, it may additionally shut the door on long-term good points to be constituted of investing in startups innovating regardless of current infrastructure and useful resource gaps. Some buyers view these gaps as limiting, whereas others nonetheless see a possibility to raised help impactful AI startups. 

“VCs also needs to be fascinated by the buildings that may help these startups. They need to start to spend money on partnership with enterprise studios,” Messan added. 

Black Star Group’s Boateng desires to see extra personal, affected person, long-term capital going into supporting context-specific AI options like language fashions for customers on the continent. He welcomes efforts lobbying for Ghanaian pension funds to allocate 5%  to Enterprise Capital and Personal Fairness within the nation.  “In a world that’s altering so quickly, not taking dangers is the largest threat itself.”

The regulatory panorama for AI startups on the continent can be nonetheless in its early phases, with solely 16 international locations up to now releasing nationwide methods or insurance policies to help AI innovation. This regulatory uncertainty may discourage investments in AI startups on the continent. 

“We want regulatory sandboxes. Individuals have concepts. Individuals wish to connect with particular databases,”  Launch Africa’s Uwemakpan mentioned. “It solves two issues: governments can see what’s taking place, whereas startups and founders can experiment.”

We’d like to know what you concentrate on this column and another subjects associated to AI in Africa that you really want us to discover! Fill out the kind right here. 

Mark your calendars! Moonshot by TechCabal is again in Lagos on October 15–16! Be part of Africa’s prime founders, creatives & tech leaders for two days of keynotes, mixers & future-forward concepts. Early chicken tickets now 20% off—don’t snooze! moonshot.techcabal.com