Kenya’s startup funding headline needs three denominators

How to tell a record total from a broad financing market

A founder sees the headline: Kenya led African startup funding in 2025, then reported funding fell sharply in the first half of 2026. An investor asks whether that means capital has disappeared, shifted into debt, or become more selective. Those are different questions. Before deciding which story is true, put three denominators on the table: dollars raised, ventures that raised, and the deals and companies included in the source.

The short answer is that the 2025 total was large and unusually exposed to a handful of energy and asset-financing transactions, while the number of Kenyan ventures crossing a stated funding threshold fell. In H1 2026, the tracker reported lower Kenya funding again. That is a useful signal about disclosed transaction activity. It does not, by itself, show that the typical startup raised less, that all Kenyan ventures found capital harder to get, or that investors became more confident in Nigeria and Egypt. The reader has to separate what the data show from what they might mean.

Start with what “funding raised” counts

Startup databases do not observe every Kenyan company borrowing, receiving a grant, raising a convertible note or selling equity. The Africa: The Big Deal database describes a monthly updated record of disclosed or privately shared funding deals at or above a threshold, currently $100,000, with more than 30 data fields per deal [2]. That is useful coverage. It is not a census of all business finance. Small grants, undisclosed angel cheques, local bank overdrafts, founder loans, supplier credit and some confidential facilities may be absent or treated differently.

A source’s “amount raised” is therefore a measured total under that source’s rules. It depends on which entities qualify as startups; whether debt, grants and asset finance are included; whether exits are excluded; how multi-country companies are tagged; whether a deal is counted when announced, signed, closed or drawn; what happens when the amount is confidential; and how revisions are handled. Those rules are not footnotes. They define the number.

The 2025 tracker post says Kenya “almost” reached $1 billion. It reports $582 million of debt and $383 million of equity, with debt representing 60% of the amount, and identifies large energy players as major drivers [2]. The screenshot supplied for this research says $984 million [1]. Contemporary reporting also attributes an exact $984 million 2025 total to Africa: The Big Deal [10, 12], although the tracker’s own public post rounds the country figure to “almost” $1 billion. The reported categories still do not reconcile to that exact value: $582m + $383m = $965m. Subtracting the subtotal from $984m leaves $19m. Because the public country summary does not provide a same-version deal ledger with every instrument, we cannot say whether that difference reflects grants, other instruments, rounding, a later update or another coverage difference. The practical treatment is to keep the reported total and categories visible, identify where the exact total came from, and leave the residual unassigned.

That is a small arithmetic discrepancy with a useful lesson. A precise-looking total can be less informative than a clearly defined approximate one. Here, the exact $984m number is corroborated by contemporary coverage of the tracker, but it is not printed as an exact figure in the public ATBD country summary. The reader should keep that source lineage visible and should not combine the number, rounded category amounts and a different report’s deal list into one exact table without confirming that they share a publication date and dataset version.

The 2025 record says something about dollars and something else about breadth

The tracker reports that 75 Kenyan ventures raised at least $100,000 in 2025, a 23% year-on-year decline [2]. The national dollar amount increased 52% year on year, while the number of ventures over that threshold fell. Those measures can move in opposite directions because a small number of very large rounds can lift the total while fewer companies receive qualifying tickets.

That is not a contradiction. It is the difference between scale and reach. The total asks how many dollars appear in the covered transactions. The venture count asks how many distinct ventures cleared a minimum disclosed amount. Neither tells us the median cheque, the number of first-time raises, the stage mix, the proportion of companies that sought funding but failed, or what share of all Kenyan startups had enough financing to reach their next milestone.

The phrase “equity did not scale” needs a defined comparison. Within ATBD’s 2025 series, the source says the dollar amount of equity almost doubled. If the claim means early-stage equity access remained narrow, the decline in the number of qualifying ventures is relevant but not enough: the count combines financing stages and is limited to deals crossing the source threshold. A more precise question is: did the number and value of early-stage equity rounds rise or fall after controlling for stage, company age and source coverage? The public summary does not answer that. As the provider comparison below shows, the size and mix of the same country-year also change when a second tracker applies its own coverage rules.

A concentration claim lives or dies on its denominator

The screenshot says five firms—d.light, Sun King, M-KOPA, BURN and PowerGen—accounted for 82% of Kenya’s 2025 funding [1]. The tracker identifies these companies as major drivers, but the public country summary does not publish a verified Kenya-only share for those five companies [2]. A separate tracker analysis reports that the top five energy companies across Africa represented $701m, or 82% of the continent-wide energy-sector total [5]. That is a different numerator and denominator. It cannot prove the screenshot’s Kenya-wide claim.

There is a second calculation hiding in the source material. The screenshot listing 13 businesses gives claimed amounts that add to KSh 93.31 billion. The five largest entries in that selected list total KSh 75.7bn. Dividing the two yields about 81.1%: 75.7 ÷ 93.31 = 81.1%. That is a concentration ratio within the screenshot’s selected 13-company list, using unverified KSh claims. It is not the share of Kenya’s 2025 venture funding. It does not establish that all 13 firms raised those amounts, that the money was received in Kenya, or that the outcomes were all failures.

To calculate a true country share, we would need a transaction ledger with one row per deal: company and recipient entity; amount and currency; date; instrument; disclosed, committed, closed or drawn status; sector; source; and the database’s country-allocation rule. Then the analyst would total the five qualifying companies and divide by the same-version Kenya total, including or excluding grants and debt consistently. If the numerator is based on deal announcements while the denominator includes drawn facilities, the result is not comparable.

A useful concentration measure is the top-five share: C5 = funding to the five largest firms ÷ total funding. It is intuitive but cannot show whether the remainder is shared by many small firms or a few medium ones. A Herfindahl-Hirschman Index can add that information: HHI = sum of each firm’s funding share squared. If one firm has 50% of the market and five others have 10% each, HHI is 0.50² + 5 × 0.10² = 0.30. If ten firms share the other half equally, the index is different. But HHI is only useful if the unit is consistent—company or deal, not a mixture—and missing transactions are not ignored as zeros.

Even that does not answer whether the market is healthy. Funding can be concentrated because one sector has bankable assets and long-term contracts. It can be broad but inefficient if many firms receive small tickets that do not match a credible use. Concentration is a diagnostic about distribution, not a verdict on company quality.

H1 2026: the fall is real in the source series, but not a universal confidence score

The Big Deal reported $227m of Kenyan startup funding in H1 2025 [3] and $126m in H1 2026 [4]. On those rounded values, the change is (126 ÷ 227) − 1 = −44.5%. This is a valid matched half-year comparison within that tracker’s method. It is not appropriate to compare H1 2026 with all of calendar 2025 and call that the year-on-year fall.

The same H1 2026 summary reports Kenya’s equity funding at $46m, compared with $214m in Nigeria and $183m in Egypt [4]. That can matter to a founder looking for equity capital, but it is not necessarily a countrywide ranking of investor confidence. The same post explains that Spiro, a Benin-born company with a Nairobi headquarters and Dubai holding company, was tagged as “Africa” going forward because of its pan-African operations, even though much of its activity is in Kenya [4]. In H1 2026 Spiro alone announced $270m of equity and $57m of debt, and the tracker says its Kenya operations put the national decline in perspective. Whether and where that transaction belongs depends on the country-tagging rule.

This is not a criticism of the tracker. It is a reminder that cross-border businesses do not fit neatly into one country box. A transaction may be raised by a holding company in one jurisdiction, contracted through an operating subsidiary in another, deployed across several markets and reported by headquarters in a third. For a country capital-market question, the researcher should show at least two views where possible: headquarters or legal recipient; and operations or use of proceeds. If the database publishes only one tag, state that rule and do not present it as domestic cash invested in Kenyan companies.

The H1 2025 source provides another useful comparison. It reports $227m for Kenya and a $100k+ venture count, with the Big Four’s share of African funding at 78% and 67% of qualifying deal counts outside the Big Four [3]. The H1 2026 summary says the Big Four accounted for 58% of total funding and 58% of $100k+ deals, but also notes the Spiro attribution issue [4]. Those continental shares should not be used as Kenya-only measures; they show why the geographical scope must accompany every percentage.

What would tell us whether access got harder?

A market dashboard should have more than one number. For each period, split funding by instrument, stage, sector and ticket size, then count unique recipients and disclosed rounds separately. Track first-time recipients and follow-on recipients. Compare the median and distribution of deal size, not only the total. Show the top-one and top-five shares, and where the underlying data permit, a concentration index. Keep grant, equity, debt, asset finance, trade credit and exits in separate columns.

The denominator should include companies that sought capital, not only those that received it. That information is much harder to obtain. An investor survey, accelerator application cohort, bank pipeline or deal process dataset can show the number of applications, term sheets, closings and rejections. Without it, a lower count of recorded transactions cannot distinguish fewer companies seeking money from more companies being denied.

A practical research table might look like this:

The peer-reviewed research warns against a simple boom/bust story

Research on venture investment cycles helps explain why “more funding” and “healthier companies” are not interchangeable. Nanda and Rhodes-Kropf study venture-backed firms receiving early-stage financing in the United States from 1985 to 2004. They find startups first funded in more active investment periods were more likely to go bankrupt, but among successful firms they were also associated with higher IPO values and more patenting [6]. Their interpretation is that hot markets can finance riskier and more novel experimentation. The study is not evidence about Kenya; it demonstrates that a funding boom can change which ventures investors choose, not simply how much money existing firms receive.

Ewens and Farre-Mensa review changes in entrepreneurial finance and the way private capital has altered company financing paths, governance and time to public markets [7]. Their synthesis is largely about the US ecosystem. It reinforces the need to look at who provides capital, what rights they receive, and what alternative financing channels exist. A funding total cannot answer those institutional questions.

For Kenya, both processes can happen together. Investors may demand stronger unit economics after losses in some sectors while also funding new asset-backed businesses through debt. Early-stage firms may face less equity even when overall dollars rise. An energy asset transaction can create a large national total while a software startup with a working product cannot raise its next round. These observations are compatible; the data need to test each rather than force a single national explanation.

What founders, investors and policy teams can do with the dashboard

A founder should compare funding offers to the next operational milestone, not to the national total. Ask: how much cash is needed before the product or route has evidence of repeat demand; what gross contribution remains after delivery, support and collection costs; what working capital is required; and what is the smallest instrument that can finance that path without creating a mismatch? A lower national equity count is a reason to improve the financing plan, not proof that no investor will fund the company.

An investor should separate access from quality. A smaller number of deals may reflect a narrower pipeline, longer diligence, reduced risk appetite, a move toward larger transactions, or strategic selectivity. A high deal count may include rounds too small to finance the next milestone. The useful questions are deal stage, time-to-close, valuation, terms, follow-on rate, realized outcomes and the counterfactual: which firms were not funded and why?

A policy or ecosystem team should avoid treating a large debt-financed asset transaction as proof that early-stage venture finance is plentiful. It should also avoid labeling all debt as low-quality finance. Measure the instrument against the financed asset and cash flow: whether local companies can obtain working capital, whether productive assets are financed at a tenable currency and tenor, whether equity is available for experimentation, and whether firms can graduate from grant-supported pilots to repeatable customer revenue.

A useful reading of the screenshot

The screenshot's strongest point is that aggregate funding can be concentrated and a single total can hide uneven access. Its arithmetic claim about five companies and 82% is not established by the public material reviewed here. Its concern about fewer deals deserves testing, while its statement that equity “did not scale” conflicts with the tracker’s reported near-doubling of Kenya equity in calendar 2025. Its suggestion that fewer deals mean stronger companies is a hypothesis, not a conclusion. The screenshot's question is worth keeping; its causal explanations need better evidence.

The reader can now make a practical judgment: H1 2026 funding in the cited tracker was about 44.5% below the comparable H1 2025 total. The 2025 Kenya headline was concentrated in large transactions and debt formed a majority of the published mix. The count of qualifying ventures fell year on year. Those facts point to unevenness and concentration. They do not yet tell us whether investor selection improved, whether viable firms were shut out, or what funding did to operating survival. Part 2 turns to the next practical question: what exactly does the money require in return?

A five-step check before repeating the headline

If you are preparing an investment memo or public presentation, use this sequence before copying a funding figure.

First, freeze the release. Save the exact report or dataset version, publication date and capture date. A website can update a total after later deal disclosures. The 2025 report was published in January 2026; a later tracker analysis may update sector data or the way an issuer is categorized. Record which version produced the number on your slide.

Second, write the unit in a sentence. For example: “US-dollar equivalent of disclosed funding transactions of at least $100,000 secured by source-defined African startups during H1 2026, excluding exits.” If the source does not disclose enough detail to complete that sentence, list the missing parts rather than implying the measure is universal. State whether the table counts dollars, rounds, companies or financing recipients. Do not name a dollar total as “number of investments.”

Third, reconstruct every subtotal you can. Keep equity, debt and grants on separate lines. Recalculate percentages from the stated base. Preserve a residual category when the total and categories do not match. Round only after the calculation. In this case, 582 ÷ 965 = 60.3%, while 582 ÷ 984 = 59.1%; both might be described as “about 60%,” but they use different denominators. That is precisely why the source's rounded prose and the screenshot's exact total should not be combined as if all inputs came from one reconciled table.

Fourth, test whether a large deal explains the story. Recalculate total funding excluding the largest transaction and the largest five, and compare their shares with the company count. That sensitivity shows how much the aggregate depends on megadeals. It does not mean those deals are inappropriate or should be discarded; it helps readers see what the total describes. If a large securitization finances thousands of productive assets over several years, the amount may be economically meaningful even though it is not equivalent to a seed-equity round.

Fifth, keep a “known/unknown/next record” row beside every public claim. For the screenshot's “82% from five firms,” the known item is that the tracker names the five energy players as major drivers. The unknown item is the Kenya-only numerator under the same source version and country definition. The next record is a deal-level tracker export or a public breakdown. A disciplined note can still be useful while the export is unavailable: “The five firms were described as major drivers; the claimed 82% share could not be reproduced from the public country summary.”

This is not paperwork for its own sake. The sentence used in the investment committee often becomes the premise for a funding strategy. When a claim has no denominator, a founder may mistakenly believe a whole financing channel has closed. When an attribution rule is opaque, a policy team may subsidize a market total that mostly reflects cross-border deployment. The five-step routine makes the next capital decision testable instead of rhetorical.

Country comparisons need common definitions, not flag colors

The H1 2026 comparison places Kenya at $126m total, below Nigeria and Egypt in total funding, and at $46m in equity [4]. The same tracker puts Spiro's $327m aggregate in an “Africa” category going forward rather than assigning it entirely to Kenya or Benin [4]. Other reports, including AVCA’s venture-capital report [8], are also useful context, but a headline comparison needs each provider’s scope and instrument rules beside the value. Partech’s primary 2025 report makes that comparison concrete: it gives Kenya a larger total than the ATBD figure, while also publishing a substantially different equity/debt split [9]. That disagreement is an invitation to reconcile scope, not choose whichever number confirms a preferred story.

Before comparing two providers, place their rules side by side: transaction threshold; equity-only or equity-plus-debt; grant treatment; startup eligibility; headquarters, legal recipient or operating geography; disclosed versus confidential deals; and reporting cutoff. If one source includes large asset-finance facilities and another focuses on venture-equity rounds, the difference is partly a definition. It still matters for founders, but it does not indicate that one source is necessarily wrong.

A useful country view can show two separate columns: source-defined total and disclosed equity total. A third column should show company-count coverage. If the sources cannot be made like-for-like, publish them as separate measures and explain the reason. Do not average them or add them together. A national “rank” that changes when a pan-African company moves from a country label to an Africa-wide label is a classification outcome as well as a capital-market observation.

Cross-checking Kenya 2025 against a second provider

The primary Partech Africa Tech Venture Capital report gives a useful independent check because it publishes both its headline and the rules behind it. For Kenya in 2025, Partech reports about $1.04bn in total equity and debt, including $539m equity and $498m debt. It says the total rose 72% year over year, with equity up 144% and debt up 30%. The equity and debt figures sum to $1.037bn, which is close to the rounded $1.04bn headline. Partech also reports 91 Kenya deals: 72 equity rounds and 19 debt transactions [9].

ATBD and Partech are not counting the same thing. ATBD’s database tracks funding deals of at least $100,000 across African startups and covers equity, debt and grants, excluding exits; Partech reports equity and debt rounds of at least $200,000 for African tech and digital startups. Partech defines “African” through the company’s primary market, using operations and/or revenue rather than requiring an African headquarters [9]. ATBD’s public database description confirms the $100,000 threshold, while its 2024 and 2025 summaries identify the included instruments and exclusions [2, 11]. These are both useful curated datasets, but neither is a census of every financing transaction in Kenya.

The total spread is 1,040 − 984 = $56m, which is 56 ÷ 984 = 5.69% of the ATBD exact value reported by the two news outlets. The instrument differences are larger and run in opposite directions: 539 − 383 = +$156m in equity for Partech, while 498 − 582 = −$84m in debt. A reader who only repeats the near-$1bn total would miss that the providers tell different stories about what financed that total. The figures do not show which individual deals explain the spread; that requires matching their deal-level records.

There is also an internal growth-rate check. ATBD’s official 2024 Round-Up reports $638m for Kenya [11], and the 2025 roundup says Kenya grew 52% [2]. Taking the exact $984m figure quoted in contemporary reports and the official 2024 $638m base gives (984 ÷ 638) − 1 = 54.23%, or about 54.2%. That is 2.23 percentage points above the tracker’s stated 52%. It could reflect rounding or different data vintages, but no inspected public source resolves which. The primary ATBD post says “almost” $1bn rather than publishing $984m; the exact amount is corroborated in reporting published after that post [10, 12, 14]. We should not silently relabel ATBD’s +52% as wrong; quote its reported rate and show our arithmetic beside it.

A reporting date is a vector, not a timestamp

The nine-day interval between ATBD’s 13 January post and Partech’s 22 January release is relevant, but it does not tell us that Partech observed nine more days of deals. A report publication date is when a document became public. It is not necessarily the last transaction date in the dataset, the date the provider first recorded a deal, or the date the provider froze its export. Similarly, “raised in 2025” is a period label, not a documented rule for whether a deal belongs to 2025 because it was announced, signed, closed, drawn or first disclosed during that year.

The inspected timeline is:

The comparison therefore needs two layers of dates. At the report level, record the funding period, release date, source, publication precision and any disclosed cutoff or version. At the transaction level, record announcement, agreement/signing, close, facility commitment, drawdown, cash receipt, provider first-seen date, the provider's date-assignment rule, and the snapshot/version containing that transaction. Keep these fields separate: a solar lender may announce a facility in December, sign it in January, disburse tranches over later months and disclose it publicly still later. The amount might mean commitment, facility size or cash drawn; each observation answers a different question.

So, are the totals consistent with different reporting dates? A timing effect is plausible, but the inspected dates cannot establish it. The reports cover the same nominal calendar year, but their underlying cutoff and assignment rules are not disclosed. The nine-day publication gap tells us when readers received each headline, not which transactions each provider had captured. ATBD's later statement demonstrates that disclosure delay exists in its process; it does not prove a post-publication revision to this particular annual total. Partech's use of confidential and partial disclosures also means its report date cannot be treated as a complete public-news cutoff.

What we can establish is that the scalar-only comparison was insufficient. The total spread is $56m, but the equity and debt differences are +$156m and −$84m. The instrument scope, eligibility threshold, geographic attribution and count units also differ. Those rule differences are visible in the published material; the share of the $56m attributable to date/vintage versus classification or transaction coverage is not. Do not write that the later report “includes deals ATBD missed,” and do not assign the $19m between ATBD's $984m secondary-reported total and $965m category subtotal to grants. Both claims require same-version, deal-level transaction records.

For a real reconciliation, obtain timestamped exports from each provider and create one row per provider-deal record. A minimum comparison key would include company and legal recipient, investor/lender, amount and original currency, instrument, announcement date, signing/closing date, commitment and draw amounts, cash-received date/status, provider first-seen date, activity year assigned, country rule, source citation, disclosure status and snapshot ID. Match records first; then explain duplicates, amounts, FX treatment, split tranches and eligibility exclusions. Only that bridge can quantify a vintage effect. Until then, show the two provider totals side by side, each with its definitions and publication date.

There is a practical route to better evidence, but the public summaries are not the underlying records. ATBD's year-end post directs readers to subscribe for its full data; its database page describes a monthly updated Excel file with 30-plus fields for African startup deals of at least $100,000 [15]. That makes a record-level comparison attainable if the export and the right historical version are obtained. Ask for the snapshot used to prepare the January 2026 report, the current snapshot, and the change history. A current monthly file alone may not reproduce what the January report counted. Partech's public report describes fully disclosed, partially disclosed and confidential entries; for Africa overall, it says confidential deals represented 24% of equity funding and 7% of debt funding [9]. Those are not Kenya-specific shares, but they show why public announcements alone cannot recreate Partech's full denominator. Until same-vintage records are available, the $56m difference remains unallocated across timing, eligibility, instrument, country and disclosure effects.

This is why the workbook has a provider-comparison sheet instead of a single blended “best” estimate. In an investor memo, put the source name in the column heading and preserve the publication date and units. If an IC member asks for one market total, the honest answer is to specify the intended question first. A venture-backed technology pipeline, a broad disclosed funding database that counts grants, and cash actually received by Kenyan operating entities are three different measures. A transaction export might allow a deal-by-deal reconciliation, but only if it records amounts, instruments, dates, legal recipients, country tags and the revision date for each provider.

The resulting recommendation is practical: use the tracker that best matches the decision, not the one with the preferred result. A venture investor comparing equity opportunities may prefer Partech’s narrower technology and round definitions, then inspect the company-level equity records. A government program measuring all disclosed startup support might prefer ATBD’s broader threshold and grant coverage. A lender judging debt capacity should move past both totals and inspect the obligor, drawn amount, security, currency, tenor and repayment source. Keep the alternate provider visible as a sensitivity. If a ranking or growth conclusion changes with the provider, that is an important result to disclose.

The same caution applies to local currency. A social post may convert a company’s multiyear dollars raised into Kenyan shillings at a present-day exchange rate, while the original rounds occurred across several years and legal entities. That number may be useful as a current-scale illustration, but it is not a historical cash-flow total. Record each original amount and currency, transaction date, recipient and source before translating. The first source should remain visible beside every converted value.

A short memo for a founder deciding whether to raise

A national funding chart is background, not a financing plan. A founder deciding whether to raise should write a one-page capital request containing: the amount needed before the next verifiable milestone; the timing of each cash use; the operating result that will be measurable; the customer or asset generating repayment or future value; the currency of the cost and revenue; the smallest financing that gets to the milestone; and what will happen if the next round is delayed. Then compare the offer's instrument and control terms against that plan.

If the business has contracted customer payments and long-lived equipment, asset finance may be more appropriate than giving up permanent equity. If the product still needs experiments before any predictable cash arrives, fixed repayment may make the downside worse. If customers pay slowly, the cash forecast should show whether working capital—not product-market fit—is the immediate funding gap. If customer acquisition requires a subsidy, record how much cash each cohort consumes before repeat purchases or margin improvement is visible. Neither a debt-market headline nor an equity-market headline can resolve those company questions.

For investors and ecosystem teams, the corresponding brief is to compare how much capital reaches companies at each stage with what those companies must prove or repay before their next financing event. Track the cost of capital, dilution, repayment burden, investor protections, operating performance and follow-on access together. Deal count alone does not tell whether financing terms are sustainable, and a lower deal count does not automatically mean selectivity improved.

Workbook map

The Startup Funding Market and Outcomes workbook is built. Market Metrics compares H1 2025 US$227m with H1 2026 US$126m and calculates a 44.5% decline. Provider Comparison lays out ATBD’s secondary-reported US$984m and Partech’s approximate US$1.04bn beside their separate equity, debt, threshold and count measures; it calculates the US$56m provider spread, the 5.7% relative difference, ATBD’s 54.2% recomputed annual change versus its stated 52%, and both providers’ category shares. Reporting Vectors records the activity period, report dates, observed-publication timeline, precision, source scope, the subscriber route to ATBD's underlying database and the missing historical snapshot, transaction cutoff and assignment-date rule. Its nine-day calculation is a publication lag only. Deal Ledger appends separate fields for provider, report date, first-seen date, export as-of date, assigned period, assignment basis and release vintage; those fields remain blank until source records and a dated vintage are obtained. Funding Mix keeps the ATBD US$582m debt and US$383m equity subtotal (US$965m) beside the reported total, leaving the US$19m unassigned. Screenshot List reproduces the 13 KES claims, their KES93.31bn sum and the first five's 81.1% share of that selected list only. Appendix 01 and Appendix 04 reproduce the calculations and source-definition audit.

Technical appendix

• Appendix 01 — Headline reconciliation and concentration
• Appendix 04 — Provider comparison and denominator audit

References

[1] Nasra Mohamed and Alex Owiti, user-supplied LinkedIn screenshots captured September 22, 2026; transcription, claims and SHA-256 fingerprints in ../01 - Screenshot Source Capture.md. The visible posts have relative timestamps and no stable URL.

[2] Africa: The Big Deal, “2025 in review: Mapping the money,” January 13, 2026, https://thebigdeal.substack.com/p/2025ir2; and “2025 in review: Fintech in gold; energy boosted by debt,” February 16, 2026, https://thebigdeal.substack.com/p/2025ct.

[3] Africa: The Big Deal, Start-up funding in Africa: H1 2025, October 2025, https://africathebigdeal.com/wp-content/uploads/2025/10/ATBD-H1-2025-Round-Up-1.pdf.

[4] Africa: The Big Deal, “H1 2026: Mapping the money,” July 20, 2026, https://thebigdeal.substack.com/p/h12026c.

[5] Africa: The Big Deal, “2025 in review: Fintech in gold; energy boosted by debt,” February 16, 2026, https://thebigdeal.substack.com/p/2025ct. The 82% claim is Africa-wide energy-sector concentration, not a Kenya-only share.

[6] R. Nanda and M. Rhodes-Kropf, “Investment Cycles and Startup Innovation,” Journal of Financial Economics, vol. 110, no. 2, pp. 403–418, 2013, https://doi.org/10.1016/j.jfineco.2013.07.001.

[7] M. Ewens and J. Farre-Mensa, “Private or Public Equity? The Evolving Entrepreneurial Finance Landscape,” Annual Review of Financial Economics, vol. 14, pp. 271–293, 2022, https://doi.org/10.1146/annurev-financial-101821-121115.

[8] African Private Capital Association, Venture Capital in Africa Report, February 2026, https://www.avca.africa/media/umlh132j/avca25-15-vc-annual-report_4.pdf.

[9] Partech, 2025 Africa Tech Venture Capital Report, January 2026, https://partechpartners.com/africa-reports/2025-africa-tech-venture-capital-report. Kenya’s country total, equity/debt split, deal counts and methodology.

[10] Business Daily Africa, “Kenya remains on top with Sh127bn startup funding in 2025,” January 2026, https://www.businessdailyafrica.com/bd/markets/market-news/kenya-remains-on-top-with-sh127bn-startup-funding-in-2025-5331672. Contemporary secondary report quoting the ATBD-attributed US$984m value.

[11] Africa: The Big Deal, 2024 Round-Up, January 2025, https://africathebigdeal.com/wp-content/uploads/2025/01/ATBD-2024-Round-Up.pdf. Kenya’s published 2024 total is US$638m; the chart describes equity, debt and grant deals, excluding exits.

[12] The EastAfrican, “Kenya takes largest share of Africa’s venture capital funding,” January 2026, https://www.theeastafrican.co.ke/tea/business-tech/kenya-takes-largest-share-of-african-venture-capital-funding-5329968. Independent contemporary corroboration of the exact US$984m figure attributed to ATBD.

[13] Africa: The Big Deal, “Where are the investors?”, May 12, 2026, https://thebigdeal.substack.com/p/invest426. Later methodology context: deals may be reported with delay through public sources and confidential investor disclosures; it does not establish a revision to the 2025 Kenya total.

[14] The Star, “Kenya tightens grip on Africa startup capital, attracts Sh126bn in 2025,” January 15, 2026, https://www.the-star.co.ke/business/2026-01-15-kenya-tightens-grip-on-africa-startup-capital-attracts-sh126bn-in-2025. Earliest exact-total report inspected for this comparison; not claimed as first publication.

[15] Africa: The Big Deal, startup deals database product page, https://africathebigdeal.com/, and “2025 in review: Mapping the money,” January 13, 2026, https://thebigdeal.substack.com/p/2025ir2. The site describes a monthly updated Excel database of African startup deals of at least US$100,000 with 30-plus fields; the year-end post directs readers to subscribe for the full underlying data. No transaction export or historical January 2026 snapshot was acquired for this analysis.

Companion resources

The workbook and technical appendices are available with research membership. The video production guide maps this article to a planned episode; a published video is not yet available.