The Nairobi Securities Exchange wants to let Kenyans invest in global AI companies in shillings, while its chief warns of a bubble. In Accra, MEST is training the next cohort of African AI founders. One bet is on owning the boom; the other is on making it local.
ABUJA – Two announcements this year, one from a stock exchange and one from a startup school, point to different answers to the same question: how should Africa take part in the AI economy?
The Nairobi Securities Exchange (NSE) is building what it says would be East Africa’s first exchange-traded fund (ETF) focused on AI stocks. And the Meltwater Entrepreneurial School of Technology (MEST) is preparing its second cohort of AI founders, with training due to begin in Ghana in January 2027.
The first lets African savers buy into AI companies that are mostly built elsewhere. The second tries to produce African AI companies. Read together, they show how much of the continent’s AI story is still being written.
The NSE’s plan, and its caveat
NSE Chief Executive Frank Mwiti told Reuters on August 5 that the exchange is creating an AI-focused ETF and plans to offer it to investors before the end of the year. The fund would track a basket of companies with direct exposure to artificial intelligence, he said, and would most likely be denominated in Kenyan shillings to limit foreign-exchange risk for local investors.
The pitch is about access. Kenyan investors can already buy AI-related products abroad, Mwiti said, but Kenyans are investing in foreign markets because of a lack of product diversity at home. He said demand is strongest among younger investors. The Nairobi market, he noted, draws many foreign equity investors thanks to its banks and the telecoms group Safaricom, but offers no AI exposure today.
Then came the caveat. “There is… a vibe in the market that there might be a bubble around AI,” Mwiti said, adding that the exchange would study the global rally and delay the launch if investors risked being exposed to an overdone move. It is unusual for an exchange chief to flag valuation risk in his own product announcement. It also fits the way AI has dominated markets this year: the boom has pushed stock markets to record highs while raising fears of a sharp correction.
What the fund can and cannot hold
Details are thin. Mwiti said companies such as Microsoft, Anthropic and OpenAI could be used as references for the fund. Anthropic and OpenAI are privately held and not listed on any exchange, so an ETF cannot hold their shares directly. The final basket will therefore depend on listed companies with exposure to AI, such as developers, chipmakers, cloud providers or investors in private labs. The fund manager, benchmark, constituents, weighting rules and fees have not been disclosed.
The proposal is being discussed with Kenya’s Capital Markets Authority (CMA), which must approve it. As of this writing, I found no public report that approval had been granted or a launch date set.
A small product shelf with big ambitions
Context matters here. The NSE has a short ETF history. Reporting at the time of the announcement said it had listed only two: a gold-tracking fund in 2017 and the Satrix MSCI World Feeder ETF in 2025. On August 11 the CMA approved a third, a shilling-denominated banking index ETF. The exchange’s 2025–2029 strategy targets 50 new funds and names AI and blockchain among the technologies it wants to bring into its offering.
Nairobi would also not be first on the continent. Johannesburg Stock Exchange listed a comparable AI-focused fund, IVYA, in March 2026, so the NSE’s would be a regional first for East Africa rather than a continental one. Kenya’s equity market has been strong: reports citing Reuters put total market capitalisation at a record of about 4 trillion shillings in 2026 after a rise of more than 30 percent. That is the backdrop for a product launch, and for the bubble worry.
MEST: building the companies
Away from the trading floor, MEST is tackling the other end of the problem. Founded in 2008 as a software-entrepreneurship school, it redesigned its programme around AI and ran its first AI cohort in 2026, which drew founders from seven African countries, according to the tech outlet tech-ish.
The second edition, the MEST AI Startup Program for the Class of 2027, is open to software developers aged 21 to 35 based in Ghana, Nigeria, Senegal or Kenya. It is fully sponsored and lasts up to 12 months. Participants spend seven months in residential training in Ghana, from January to July 2027, forming teams, validating ideas and building working products. A select group of ventures then enters a four-month incubation phase, from September to December 2027, and may pitch for pre-seed investment of up to $100,000.
MEST says training will draw on experts from organisations including OpenAI, Perplexity, Google and Meltwater. Emily Fiagbedzi, the programme’s director, has argued that meaningful participation in the global AI economy requires being able to build rather than only access tools.
One correction for readers: some listings, including a headline aggregator, still describe applications as open. MEST’s own application page says applications opened on May 19 and closed on July 20, 2026. The programme’s published timeline has contracts going out this month and virtual pre-learning in November and December, so the 2027 cohort is already being assembled.
The gap between exposure and ownership
Set side by side, the two stories raise the same uncomfortable point. An AI ETF in shillings is a convenient way for a Kenyan saver to own a slice of companies that are almost all based abroad. A $100,000 cheque to an African founder is meant to create something that will, one day, be worth owning at home.
The funding picture shows why the second is hard. Tech In Africa reported in September that African startups raised about $1.36 billion in the first half of 2026, but only 190 companies secured rounds of at least $100,000. It also found that genuinely AI-native companies receive a relatively small share, with much “AI” funding going to firms that apply AI in sectors such as fintech or agriculture, and that the financing gap opens at the earliest stages. Programmes like MEST’s target that early stage, but a single cohort is small against the need.
What to watch
- CMA approval. The NSE has not announced a regulator decision, a fund manager or a launch date. Watch for those, and for whether the exchange delays the launch if global AI valuations fall or swing.
- The basket. Whether the fund holds large US technology companies, chipmakers or a broader set will determine both its risk and its link to Africa, which is likely to be limited.
- MEST’s results. The first cohort’s outcomes, including how many ventures raise follow-on funding, will say more about the model than the intake figures do.
- The bubble question. The NSE chief’s caution may prove prudent or premature. Either way, it is a reminder that exposure to AI is also exposure to its market cycle.
Sources and verification notes
- Reuters (August 5, 2026), as syndicated by BusinessDay, KFGO and others, on the NSE’s planned AI ETF.
- iAfrica, News Ghana and Serrari Group on the ETF’s structure, the NSE’s strategy and open questions.
- Capital Markets Authority of Kenya press release (August 11, 2026) on the banking index ETF.
- MEST AI Startup Program application page and timeline; Disrupt Africa (May 25, 2026); tech-ish (May 19, 2026); TechAwkNG.
- Tech In Africa (September 14, 2026) on African startup funding.
Editor’s checks before publication:
- The NSE announcement is two months old. I found no later news on CMA approval or a launch date. Ask the NSE for a current status.
- Reports differ on how many ETFs the NSE had before August. The CMA says its approval brings the total to three, naming an Absa-branded gold fund (earlier reports call it Barclays NewGold). Confirm the names with the NSE.
- Mwiti’s bubble quote is a short extract from Reuters. Re-check wording against the Reuters original before running it as a direct quote.
- The first MEST cohort’s “seven African countries” figure comes from tech-ish. Confirm with MEST, and request outcome data from the first cohort.
- The iAfrica listing that prompted this story presents the MEST applications as open. The programme’s own page shows they closed on July 20.
