What happened after the raise?

Follow the company, the cash and the legal claim separately

A social post says 13 Kenyan startups “collapsed after raising” a combined Sh93 billion. The sentence feels conclusive: large funding, failed companies, money gone. But it puts three questions into one label. Which legal entity received the financing? What happened to each business and its operating companies? What money, if any, did investors or creditors recover? Without those answers, the total is not a failure rate or an investor-loss figure.

The screenshot’s 13 amounts do add to KSh 93.31 billion, which rounds to Sh93 billion [1]. The five largest items add to KSh 75.7 billion, or about 81.1% of that selected list. That calculation is useful for understanding the list’s internal concentration. It does not verify the amounts, establish a national funding share or prove a loss. The post supplies no underlying deal ledger, transaction status, currency-conversion dates, legal recipients or source methods beyond “MoneyAcademyKE.”

This article gives the list a more practical treatment. It separates financing events from operating outcomes, gives each legal entity its own dated status, and shows what records would be needed to estimate recovery. It also explains why a company can raise a large amount and still run out of cash—and why an adverse outcome does not prove that every earlier investment decision was irrational.

“Raised” has several possible meanings

A funding story can refer to an amount announced in a press release, a signed term sheet, a committed facility, a closed transaction, cash drawn, or funds that cleared into the company’s bank account. Those stages can be months apart. A credit facility may have a maximum commitment but permit drawings only against eligible assets. A multi-tranche equity round may close in stages. A grant may be awarded subject to milestones. A public amount may include capital invested in subsidiaries, assets acquired in a transaction or a refinancing that pays an older creditor.

For each company, record the financial event in a transaction ledger rather than one cumulative headline:

When totals are converted to Kenyan shillings, keep the original currency and the exchange rate/date beside the converted value. Translating a US-dollar round at today’s exchange rate can be a valid illustration of current scale. It is not the amount of cash received in Kenya at the time. If a US-dollar figure was announced but paid in several tranches, use the cash receipt dates and rates for the cash-flow record.

“Collapsed” is not one company status

A company may continue selling after a funding round ends. It may close one country operation while trading elsewhere, restructure debt, sell assets, merge with a competitor, enter administration, be liquidated, or cease operations without a formal insolvency process. These outcomes matter differently to employees, customers, lenders, investors and suppliers. They should not be collapsed into a single “failed” category.

Use a status table with exact company names and dates. Keep at least two fields: operating status and legal process. “Trading” is an operating status. “Under administration” is a legal process. A company can trade during administration; an entity can cease activity without liquidation. The legal status of a parent does not automatically determine the status of every subsidiary, and a brand can outlive the original corporate entity.

A source can also be accurate but too broad for a company-level claim. Financial Sector Deepening Kenya (FSD Kenya) reported in 2024 that several prominent firms, including Sendy, Swvl, WeFarm, Copia Global, iProcure and Gro Intelligence, had either closed, entered administration or exited Kenya; it separately described MarketForce, Twiga Foods and Wasoko as facing cash shortages, downsizing, restructuring or pursuing mergers [2]. That groups several outcomes together. It does not tell us which exact status applies to each company at every date. The report is valuable context, but company-level assertions still need dated primary records.

What can be said about the 13-company list today?

The table below records the screenshot’s claimed amounts as leads. It deliberately does not treat them as verified funding received. The outcomes are bounded by the public materials identified in the source capture and verification ledger.

The amount column is in KSh billions and comes from a screenshot, not audited filings. The table does not claim that the listed firm itself received every shilling: an investor may have funded a parent, foreign holding company, project vehicle or separate operating subsidiary. It also does not imply that the company still owes all financing or that an investor recovered nothing.

The distinction is visible in the Twiga example. The September reports cited in the research materials refer to GT Flow Limited (formerly Twiga Foods One Limited) and Templar Field Limited (formerly Twiga Foods Limited) as separate companies. One report describes an administration effective in August 2026; another describes a September board appointment. The underlying Gazette notice and administrator communication have not both been directly verified in this package [8]. The responsible wording is attribution to the report, with an open primary-source check. Even if the appointments are confirmed, they do not by themselves establish that every entity using the Twiga name entered administration or that all investor claims are unrecoverable.

Copia provides a stronger primary-record example. The Kenya Gazette of 31 May 2024 states that administrators were appointed to Copia Kenya Limited [3]. A ruling delivered on 23 April 2026 is available here only as a secondary reproduction; it reports that the court granted a two-month extension to the administrators, although they had sought six months [4]. I did not locate the ruling on Kenya Law during this check. Treat the term extension as a dated, secondary-reported court outcome, not proof of Copia's status after that order or of any creditor's final recovery. The Gazette appointment identifies an entity and process; it does not tell us the company's pre-administration cash economics or final creditor waterfall.

Recalculate what the screenshot actually proves

The 13 listed amounts sum to Sh93.31bn. The top five entries—Twiga, Copia, Gro Intelligence, Koko Networks and Mobius—sum to Sh75.70bn. The ratio is:

Sh75.70bn ÷ Sh93.31bn = 81.13%.

That proves only the arithmetic relationship among the 13 displayed claims. It says nothing about national venture funding because the list is selected around firms described as collapsed, mixes outcomes, covers more than one financing period and uses an unverified currency conversion. It is a sample chosen after adverse events, not a complete list of all funded startups. A high concentration among the largest failed firms may be interesting, but it cannot estimate the probability that a funded Kenyan startup fails.

FSD Kenya listed dollar-denominated funding claims for nine companies in a 2024 discussion: Twiga $195m, Wasoko $144m, Gro Intelligence $125m, Copia Global $120m, Swvl $100m, MarketForce $42.5m, WeFarm $32m, Sendy $26.5m and iProcure $17.2m [2]. These amounts sum to $802.2m. FSD describes the firms collectively as having raised over $800m and cites several commercial databases and media sources. That arithmetic is a check of FSD’s displayed entries, not a second audit. It differs in companies, currencies, source date and classification from the screenshot’s Sh93.31bn list; the two should not be converted and compared without original round schedules and common FX dates.

The task is not to find a single “correct” cumulative number by averaging lists. It is to build an event-level record and show the range of transactions confirmed, attributed or unresolved. If the list includes a loan that was refinanced, count the original loan and refinancing as separate financing events but do not treat both as new cash available for operations. If it includes a secondary share purchase, distinguish the seller’s proceeds from capital injected into the company. If a facility was never drawn, do not add its full committed amount to cash received.

Raised capital is not the same as investor loss

A company can raise KSh 1bn in equity, spend part on product and payroll, return some assets, sell a subsidiary, raise a later round and eventually enter administration. An equity investor’s realized result depends on the amount actually paid in, any distributions or secondary-sale proceeds, final share value and transaction costs. A creditor’s result depends on allowed claims, priority, security, costs of the process and distributions from the borrower’s estate. The national funding amount is neither result.

For an investor, a simple realized equity measure is:

Realized cash loss = cash invested − cash distributions − net proceeds received on share sales or liquidation.

This is not a complete investment return. It ignores timing, follow-on contributions, taxes, fees, FX and unrealized value. For an annualized return, use dated cash flows and a clearly defined valuation date. If the company remains under administration, the shareholder value may be unresolved rather than zero.

For a creditor, a simple recovery ratio could be:

Recovery rate = cash distributions received ÷ allowed claim amount.

The numerator should identify whether it is principal, interest, collateral proceeds or a compromise. The denominator should distinguish secured and unsecured claims and state the date. A petition amount is not automatically an allowed claim; a court judgment may record what a party asserted without finally adjudicating every amount. A company’s “raised” sum cannot be substituted for the creditor denominator.

The Nakumatt court record gives a useful example of source status. A 2021 ruling summarizes an administrator report stating that KSh 5,218,737,409 had been raised and more than KSh 3.5bn paid to creditors, including KSh 766,641,309 to landlords [5]. The “raised” amount is what the administrator reportedly stated about the process, not a full audited annual cash-flow series. The landlord amount is included within the reported creditor payments; adding it again would double count. These are valuable case facts, but not a final recovery rate for every claimant.

The Tuskys judgment offers another arithmetic test. Three petitioners stated claims of KSh 248,220,797.38, KSh 11,284,096.50 and KSh 30,892,518.92 [6]. Their sum is KSh 290,397,412.80. It remains the sum of three petitioners’ claims, not total company debt or final creditor loss. The later ruling identifies Tusker Mattresses Limited as in liquidation and discusses a liquidator’s request for time to conduct a forensic audit and file a statement of affairs [7]. That request is not the completed audit result.

A raise can extend runway without fixing unit economics

Funding can pay payroll, marketing, inventory or debt service for a period. Whether that extends a viable business depends on contribution, collection timing and the cost of serving customers. Take a fictional delivery route: it invoices KSh 100,000 of sales, spends KSh 66,000 on goods, KSh 14,000 on delivery and handling, and KSh 8,000 on customer support and payment costs. Contribution is KSh 12,000 before hub rent, central staff, financing, vehicle replacement, tax and growth spending.

If customer cash arrives 45 days after delivery but suppliers require payment in 15 days, the route can show positive contribution while consuming working capital. More orders may improve route density and reduce cost per delivery; more orders can also deepen the cash gap if customers pay late or new inventory must be purchased in advance. Equity can finance that experiment without a fixed monthly repayment, but only if management sets a test period and stop criteria. A receivables facility could finance accepted invoices, but only if the terms and customer quality support the borrowing base. Neither instrument creates contribution by itself.

An investor reviewing a company after a raise should trace the “sources and uses” schedule: opening cash; cash received by date; restricted proceeds; operating receipts; inventory and supplier payments; payroll and customer acquisition; capex; debt service; and closing unrestricted cash. Then ask which metrics changed after the financing: gross contribution per order, returns, retention, collections, service quality, staff productivity and cash required per unit of growth. If all that is available is a fundraising headline and an administration notice, the right conclusion is that the operating bridge remains unknown.

What peer-reviewed research adds to the retrospective

Venture-capital funding cycles can affect both the quantity and the type of ventures funded. Janeway, Nanda and Rhodes-Kropf review research on VC booms and busts, including evidence that capital supply, sector concentration and investor incentives can affect the kinds of firms selected and the terms offered [9]. Nanda and Rhodes-Kropf’s study of investment cycles finds that firms initially financed in more active periods were more likely to go bankrupt but successful firms also had stronger innovation and exit outcomes [10]. The sample is US venture-backed companies. It does not prove that a Kenyan funding boom caused any company in this list to fail. It does caution against using either funding abundance or funding scarcity as a direct measure of company quality.

Shumway’s hazard-model research demonstrates why financial distress is a time-to-event problem rather than a single ratio observed once [11]. Fischhoff’s work on hindsight bias is relevant when analysts revisit a company after a public administration or closure [12]. Once an outcome is known, earlier expansion, hiring or fundraising decisions can look more obviously reckless than they did to people acting with partial information at the time. A fair analysis saves the information set by date, then separately adds later records.

The work of the case study is not to excuse management or investor choices. It is to identify the mechanism supported by evidence: a missed customer collection, an asset financed longer than the loan tenor, an unprofitable customer cohort, a failed refinancing, supplier arrears, a loss of distribution access or a governance decision. If the record cannot discriminate among causes, write what it establishes and name the missing document.

Build a company outcome file in five passes

Pass 1: Map the legal entities. List parent, holding company, operating subsidiaries, asset-owning entities and trade names. Record company numbers and former names. Draw the ownership relationships only where a filing, registry record, contract or administrator notice supports them. Do not infer that shared branding means shared liabilities.

Pass 2: Reconcile funding. Add one row per transaction with announcement, close and cash dates, original currency, legal recipient, instrument, security, amount drawn and source. Separate a company's own shares sold from secondary sales; separate cash equity from loan commitments; flag later refinancing.

Pass 3: Build the operating timeline. Record what was public or known to management at each decision date: product launches, customer terms, route openings, hiring, shutdowns, leadership changes, missed payments, fundraising, acquisition, court filing and administrator appointment. Use exact dates and identify who could access the information.

Pass 4: Trace cash and unit economics. If available, link customer orders to delivered sales, cash collections, direct cost, returns, delivery cost, working-capital investment, capex and debt service. Calculate contribution by cohort or unit. Keep company-level cash history blank if records are absent; do not fill it with a fictional example and label it as history.

Pass 5: Classify the outcome and recovery. Date the trading status, legal status, asset sale or merger separately. Estimate creditor recoveries only from administrator distributions or confirmed payments. Estimate equity outcomes only from distributions, secondary proceeds or a supportable remaining valuation. If unresolved, retain “unknown.”

A decision record for the next founder and investor

For founders, the lesson is to connect every raise to a milestone and a cash plan. Before signing, test what happens if the next round is delayed six months, if a top customer pays late, if inventory turns more slowly, or if the currency moves against the company. Name the action for each trigger and the person accountable. Record any personal guarantees, liens or security granted by affiliates.

For investors, record the financing thesis before the outcome is known. What evidence supported the customer opportunity? What was the expected capital need to reach a durable margin? Which assumptions were uncertain? What would count as a successful adjustment, a failed hypothesis or a new funding need? A later failure does not alone show that the original investment was irrational. A contemporaneous investment memo makes that test possible.

For journalists, ecosystem teams and policymakers, do not add all “raised” values for a selected group and call the result losses. Separate financing instrument, entity, outcome, time period, national market and source. If the data are incomplete, state which records could change the conclusion. Readers can handle an unresolved finding when the next research step is concrete.

What the 13-company screenshot currently supports

The screenshot supports two limited facts: it displayed 13 named brands with claimed amounts, and those amounts arithmetically total KSh 93.31bn. The five largest displayed claims form about 81.1% of that selected list. FSD Kenya’s 2024 report supplies a later public snapshot of several companies’ financing and broad operating/legal outcomes, while court and Gazette materials give more precise entity and legal-process records for Copia, Nakumatt and Tuskys. September 2026 reporting adds a new, still-to-be-verified administration status for two Twiga-related legal entities.

The research does not yet support the claim that every listed startup “collapsed,” that each amount reached the named Kenyan operating company, or that investors lost KSh 93bn. It also cannot reconstruct each company’s owner earnings or declare one cause of distress. The useful next output is a verified ledger and matched workbook, not a stronger adjective.

Do not turn a selected list into a startup failure rate

The screenshot begins with companies already described as collapsed. It is therefore a case list selected on an outcome. To estimate a failure rate, start with a defined population: for example, every Kenya-tagged venture that received a source-defined first equity or debt transaction in 2020, then observe each for five years. The researcher would need a denominator of all qualifying recipients, including firms that continue trading, were acquired, changed business model, stopped operating or remain unresolved. If companies are included only after a reported closure, the sample cannot tell us the share of all funded ventures that closed.

The outcome also needs a fixed definition and horizon. “Closed operations in Kenya” is different from “liquidated legal entity”; “administration” is a legal procedure, not necessarily a terminal operating state. An acquisition can preserve employees and products while ending the original shareholders' investment. A voluntary pivot can close one business line without ending the parent company. Code each dimension separately and allow more than one event date. State whether the event is first legal filing, cessation of trading, asset sale, acquisition or final dissolution.

Follow-up periods differ. A company funded in 2025 has had less time to fail than one funded in 2018. If the analysis counts only companies whose outcomes are already known, it will overrepresent older cohorts. Survival or hazard methods can handle unequal observation time when dates and event definitions are reliable, but they cannot fix missing companies or inconsistent source coverage. Shumway's public-company bankruptcy study treats failure as a time-dependent outcome; it does not supply a Kenyan startup base rate [11].

An evidence-quality column should accompany the outcome code. A court judgment can verify that an order was issued, and may summarize a creditor's statement. It does not independently establish every allegation in that statement. A company release can verify what management announced, not the ultimate cash recovery. A media report can date an event and point to a Gazette notice. A screenshot can preserve the wording of a claim. Those source types can all be useful, but their uses differ.

A practical case register can record one row per dated event with: legal entity; former name; country operation; event date; public date; source; source passage; event category; operational status; legal status; evidence level; and follow-up record needed. Keep the earlier status in history rather than overwriting it when a later event arrives. This lets a reviewer answer both “What is the status now?” and “What could a lender or supplier have known then?”

Workbook map

The Startup Funding Market and Outcomes workbook has a Company Outcomes sheet that places each screenshot amount beside its current evidence note, but does not label every company a failure. Deal Ledger separates announcement, closing, drawdown and cash receipt and stays blank until evidence supports a transaction. Sources and Checks records the Gazette, court and tracker sources and flags where the company or funding entity still needs reconciliation. There is no recovery total because investor-level recoveries have not been established, and the selected 13-company list is not a cohort from which a failure rate can be estimated. Appendix 03 provides the full claim-by-claim register and the next evidence request for each case.

Technical appendix

• Appendix 03 — Thirteen-company claim and outcome register

References

[1] Alex Owiti, user-supplied LinkedIn screenshot captured September 22, 2026; transcription and image hash in ../01 - Screenshot Source Capture.md. The post attributes its source to “#moneyacademyke” but does not include a stable URL or underlying deal ledger.

[2] Financial Sector Deepening Kenya, “Building resilient innovation ecosystems in Kenya,” October 15, 2024, https://www.fsdkenya.org/blogs-publications/building-resilient-innovation-ecosystems-in-kenya/.

[3] Republic of Kenya, Kenya Gazette, Vol. CXXVI No. 76, May 31, 2024, Insolvency Notice E106 of 2024, appointment of administrators to Copia Kenya Limited, https://new.kenyalaw.org/akn/ke/officialGazette/gazette/2024-05-31/76/eng@2024-05-31/source.

[4] In re Copia Kenya Limited (Under Administration), Insolvency Notice E106 of 2024, [2026] KEHC 5926 (KLR), ruling delivered April 23, 2026, https://sheriahub.com/cases/ke/caselaw/in-re-copia-kenya-limited-under-administrstion-2026-kehc-5926-klr. Secondary reproduction; the official Kenya Law copy was not located in this check.

[5] In re Nakumatt Holdings Limited (Under Administration), Insolvency Cause No. 10 of 2017, [2021] KEHC 8603 (KLR), ruling delivered March 11, 2021, https://new.kenyalaw.org/akn/ke/judgment/kehc/2021/8603/eng@2021-03-11.

[6] In re Tusker Mattresses Limited, Insolvency Petitions E018 of 2020 and E026 of 2021 and Insolvency Cause E002 of 2020 (consolidated), [2023] KEHC 18107 (KLR), judgment delivered May 31, 2023, https://new.kenyalaw.org/akn/ke/judgment/kehc/2023/18107/eng@2023-05-31.

[7] Hotpoint Appliances Limited & another v Tusker Mattresses Limited & another; Njenga (Suing as Interim Liquidator), [2025] KEHC 14728 (KLR), ruling delivered October 16, 2025, https://new.kenyalaw.org/akn/ke/judgment/kehc/2025/14728/eng@2025-10-16.

[8] “Twiga Foods is in administration, 16 months after the leaked plan we published,” Tech-ish, September 14, 2026, https://tech-ish.com/2026/09/14/twiga-foods-is-in-administration-16-months-after-the-leaked-plan-we-published/; and “The second Twiga company has gone into administration,” Tech-ish, September 15, 2026, https://tech-ish.com/2026/09/15/twiga-foods-templar-field-administration/. Both are secondary reports; the cited primary notices remain a publication check.

[9] W. H. Janeway, R. Nanda, and M. Rhodes-Kropf, “Venture Capital Booms and Start-Up Financing,” Annual Review of Financial Economics, vol. 13, pp. 111–127, 2021, https://doi.org/10.1146/annurev-financial-010621-115801.

[10] 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.

[11] T. Shumway, “Forecasting Bankruptcy More Accurately: A Simple Hazard Model,” The Journal of Business, vol. 74, no. 1, pp. 101–124, 2001, https://doi.org/10.1086/209665.

[12] B. Fischhoff, “Hindsight Is Not Equal to Foresight: The Effect of Outcome Knowledge on Judgment Under Uncertainty,” Journal of Experimental Psychology: Human Perception and Performance, vol. 1, no. 3, pp. 288–299, 1975, https://doi.org/10.1037/0096-1523.1.3.288.

[13] P. Gompers and J. Lerner, “The Venture Capital Revolution,” Journal of Economic Perspectives, vol. 15, no. 2, pp. 145–168, 2001, https://doi.org/10.1257/jep.15.2.145.

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.