In the space of a year and a half, Atlantica Ventures has led or co-led funding rounds for four African or Africa-linked start-ups building the underlying layers of artificial-intelligence infrastructure — semiconductor design talent, vector search, DevOps automation and compliance orchestration — a concentration of activity that marks a departure from the fintech- and logistics-heavy portfolio the Lagos firm built over its first five years.
The latest, announced September 1, is a $1.9 million round for ChipMango, an AI-native semiconductor technology company that trains engineers in chip design and verification and deploys them on commercial hardware programmes. Atlantica led the round, with DFS, Kaleo Ventures, Madica, Trilinear Technologies and Malta Ventures participating.
The ChipMango bet
ChipMango’s pitch rests on a labour shortage rather than a technology gap. Deloitte has estimated that the global semiconductor industry will need one million additional skilled workers by 2030, roughly 100,000 a year, even as the industry’s revenue is expected to grow more than 80 percent from its 2021 base of just over $550 billion — pushing it toward the trillion-dollar mark. ChipMango’s model is to train engineers on browser-based design environments built around industry-standard electronic design automation tools, place them on paid commercial engagements — including verification work for Trilinear Technologies, a supplier of DisplayPort and multimedia intellectual property — and route the resulting expertise into its own edge-AI and hardware products. It has also been named an Arm Approved Training Partner and works with Ohlone College in California and the University of Pretoria’s Carl and Emily Fuchs Institute for Microelectronics.
The company, headquartered in Delaware with operations in the United States, South Africa, Nigeria, Rwanda and Malta, said the new capital would fund a European design centre in Malta and workforce-development activity in Kigali. “The semiconductor industry’s binding constraint is no longer capital, it is people,” said Aniko Szigetvari, a founding partner at Atlantica Ventures, arguing that ChipMango converts African engineering talent into chip-design capacity already used by commercial customers.
A pattern, not an outlier
ChipMango is the fourth AI-related lead or co-lead Atlantica has taken since March 2025. Atlantica led NOSIBLE’s Pre-Seed round in March 2025, then led Salus Cloud’s Seed round in June 2025, and co-led Cybervergent’s Seed round in March 2026 — a cluster of deals that, analyzed together, spans most of the AI stack rather than repeating a single thesis.
NOSIBLE, a South African-founded start-up led by Stuart Reid, raised $1 million in pre-seed funding to commercialise a disk-optimised approximate-nearest-neighbour search index that it says cuts the cost of large-scale search relative to existing methods; the company has not published independent benchmarks for that comparison. Salus Cloud, a Cape Town-based AI-native DevOps platform founded by Deimos chief executive Andrew Mori, raised $3.7 million co-led by Atlantica and pan-African investor P1 Ventures, with participation from Lofty Inc. Capital’s Idris Bello, Everywhere Ventures and angel investor Timothy Chen; it prices its managed enterprise tier from $5,000 a month against what it says is the cost of hiring a single DevOps engineer in Nigeria, Kenya or South Africa. Cybervergent, a governance and security-posture platform, raised $3 million co-led by Ventures Platform and Atlantica; the company says more than 150 organisations across West, East and Southern Africa use the platform and that it has cut compliance and audit work by more than 70 percent, a figure it has not had independently verified.
Atlantica’s involvement with AI-adjacent companies predates this run: it co-invested, alongside Mozilla Ventures, in the roughly $2.5 million seed round that backed Lelapa AI, a Johannesburg-based natural-language-processing company founded by Pelonomi Moiloa and Jade Abbott in 2022. Lelapa builds Vulavula, a speech and text engine for African languages, and InkubaLM, a compact multilingual model trained on African-language data intended to run on low-resource hardware.
An infrastructure thesis
Individually, none of the five deals is large; combined, the disclosed amounts across ChipMango, NOSIBLE, Salus Cloud and Cybervergent come to under $10 million. What distinguishes the run is its shape rather than its size: chip-design talent sits at the hardware layer, NOSIBLE at the retrieval-infrastructure layer, Salus Cloud in developer tooling, Cybervergent in governance, and Lelapa in language and data. Szigetvari has described the firm’s approach in similar terms in earlier public remarks, saying Atlantica looks at the full AI stack, from foundational models to infrastructure to customer-facing applications, and that it sees opportunity in applied AI for areas such as cold-chain logistics, agriculture, cybersecurity, DevOps and compliance rather than in large language models alone.
That stack-wide approach is a shift in emphasis for a firm whose earlier, better-known bets — Sabi, OnePipe, Sendy and Curacel — sat mostly in fintech, e-commerce infrastructure and logistics. Atlantica, founded in 2019 by Szigetvari and Ik Kanu, closed a $50 million first fund backed by the European Investment Bank and the EU’s Boost Africa programme, and has generally written seed to Series B cheques of $10 million to $50 million across the continent, according to third-party investor data — larger, in most cases, than the sub-$4 million tickets it has taken in this recent AI cluster, suggesting the firm is treating these as early, thesis-testing positions rather than scaled bets.
The wider funding backdrop
The activity sits against a continental picture in which AI investment remains small relative to overall deal flow. African AI start-ups raised about $1.25 billion cumulatively between January 2019 and March 2025, and roughly 207 AI-labelled start-ups were tracked across the continent between 2022 and 2025 — figures that researchers caution are incomplete given inconsistent sector-tagging. More broadly, African start-ups have raised more than $1.6 billion so far in 2026, roughly matching the amount raised over the same period in 2025, even as the number of disclosed deals has fallen. The shift points to a market characterised by fewer, larger rounds, increasingly concentrated in later-stage and debt financing, with fintech and mobility still accounting for the bulk of capital deployed across the continent. Atlantica’s AI cluster, by contrast, is composed of small, early rounds in a segment that has not yet attracted the mega-deals seen elsewhere in African tech.
Whether that positioning proves prescient will depend on factors outside any single investor’s control — global chip demand, the pace at which African enterprises adopt AI-native tooling, and whether larger funds follow into these companies’ next rounds. For now, Atlantica’s recent cheque-writing suggests a firm betting that the more defensible openings in African AI lie in the infrastructure underneath the applications, not in the applications themselves.