Investor brief / Nairobi

Health intelligence.
Shared prosperity.

Build the infrastructure that turns extraordinary intelligence into everyday care.

Clinician-supervised AI. Protected health capital. Devices within reach. African discovery that earns its next experiment.

Explore the $50m programmeRequest the manifesto
Proposed programme · all funding uncommitted
$50mstaged programme capital
36 monthsevidence before expansion
4 instrumentsseparate financial responsibilities

01 / The thesis

Intelligence is abundant.
Completed care is the opportunity.

The investment case starts with a narrow, paid clinical workflow. Every adjacent layer must make that care journey more reliable, more accessible or more measurable.

01

Deliver useful care

Reduce clinical administration and coordinate care, with accountable clinicians and measured utility.

THE FIRST COMMERCIAL WEDGE
02

Make access possible

Connect consented health contributions and validated devices, each with its own permissions and risk bearer.

EARN THE RIGHT TO EXPAND
03

Compound the evidence

Build permitted local data, skilled operators and reproducible African discovery. Fund research through gates.

LONG-TERM OPTIONS

Why this institution? Distribution, clinical accountability, permitted data and verifiable delivery can retain value as frontier intelligence becomes cheaper. Faster AI progress also raises competition; success must be earned.

02 / Capital architecture

Ambitious by design.
Disciplined by structure.

$24m builds the company. Separate facilities protect care, finance devices and fund research. Programme mobilization and corporate proceeds are different things.

Programme capital / USD$50million

Four instruments. Distinct mandates.

01 / CAPITAL MANDATE
$24m

Corporate equity

Build the clinical platform, validate delivery and earn distribution.

Shareholders bear corporate operating risk. Successive closes; valuation negotiated against evidence.

Source: investment manifesto, p. 5

Successive closes / USD millions

The next dollar follows the next proof.

First evidence$12m

A narrow paid clinical workflow, a lawful fund structure, explicit consent and one approved device SKU.

Corporate equity
$6m
Device warehouse
$2m
Protected health capital
$2m
Research grants
$2m

Stage totals: $12m + $16m + $22m. Warehouse amounts add conditional capacity; they are not automatic cash receipts. The $6m first equity close can progress with separately negotiated pilot funds.

Source: investment manifesto, p. 6

Corporate equity / $24m

Spend on the platform.
And the proof.

$20m planned fixed deployment over three years, including the SPV equity transfer. $4m reserve before revenue and variable costs.

The $1.5m device line includes a $1m SPV equity transfer and $0.5m integration. This is one allocation, counted once.

Source: investment manifesto, p. 7
Clinical platform & validation$7m
Distribution partnerships$4m
Cash reserve$4m
Data, models & infrastructure$3m
Management, operations & legal$3m
Device integration & prototypes$1.5m
Academy & consultancy$0.8m
Noninvasive BCI exploration$0.5m
Venture experiments$0.2m

03 / Matawis / the physical intelligence layer

Care does not happen
outside the real world.

Minerals, energy and fertilizers connect the device, the clinic and the household. The elegant integration is better operating decisions across that shared system.

The materials behind access.

Concentrated processing and component supply can affect device lead times, backup power and replacement costs. Matawis turns those exposures into procurement questions.

IEA · structural evidence, 2025
  1. 01Minerals & components
  2. 02Batteries + validated devices
  3. 03Landed cost + repairability

Gate: approved suppliers, service quotes, spare parts and loss-adjusted device economics.

Matawis is a distinct intelligence and solution partner. These are operating-risk channels; the $50m programme includes no separate mining, energy-project or fertilizer investment allocation. Intercompany services, IP and data rights require contracts.

04 / World state / African choices

A forecast you can inspect.
And eventually score.

Six horizons from January 2027 to October 2029. A public-evidence mixture and a conditional AGI branch. Both remain visible, challengeable and uncalibrated.

613retained source documents
1,361source-linked evidence records
138 + 12binary questions + continuous parameters
19 Oct 2026frozen evidence cutoff

Continuous planning envelope

Brent futures proxy

86.1USD / barrel
1 year · P10–P90: 63.2–140.4
Brent futures proxy: public mixture, six horizons55.9105.9155.93m6m9m12m24m36mMonths from 19 October 2026
Public-mixture medianP10–P90 model envelope

Exact supplied model outputs, displayed to one decimal; not live quotes. Bands have no demonstrated historical coverage. The AGI-assumed branch is conditional planning, not proof of achieved AGI.

World forecast, p. 12 · all values are analyst judgments
View all six values
Brent futures proxy · USD / barrel · public mixture
HorizonMedianP10P90
3 months105.988.8137.7
6 months96.171.3155.9
9 months89.564.3151.8
1 year86.163.2140.4
2 years86.659.3141.2
3 years87.655.9146

Two planning mixtures

Faster AI changes
both sides of the distribution.

Assuming achieved AGI increases the weight on rapid productivity and rapid destabilization. It does not remove power, evidence, funding or institutional constraints.

Weights are analyst judgments, not learned frequencies or a probability that AGI exists.

World forecast, p. 9
ScenarioPublicAGI assumed
Managed stabilization
35%
15%
Prolonged supply shock
20%
10%
Financial & AI deleveraging
15%
10%
Fast AI productivity diffusion
15%
30%
Fast AI destabilization
10%
30%
Persistent systemic cascade
5%
5%

The question register

Conviction, with conditions.

Download JSON

Evidence-informed public planning mixture. Probabilities are uncalibrated analyst judgments.

10 questions · deadline 2027-10-19 · event / state definitions matter

F038 / Cumulative event

Will the monthly Europe Brent spot price average exceed US$130 per barrel in any month after the cutoff and by the horizon?

Resolver: www.eia.govPDF p. 23
Resolution rule & assumptions

YES if EIA Europe Brent Spot Price FOB monthly series has at least one post-2026-10-19 calendar-month average >130; allow first revision and retain version.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: IEA Oct 14 physical shortage and inventory draws raise tail risk; EIA Oct 6 base projects easing. This is cumulative monthly price breach, not horizon-end spot.

Public-branch input sensitivity: 22–51%. Input sensitivity is not a confidence interval.

Evidence context: S0382. Source links support audit; they do not estimate probabilities.

38%
J · uncalibrated
F040 / Cumulative event

Will an EPRA monthly Nairobi diesel retail ceiling exceed KSh 250/litre by the horizon?

Resolver: www.epra.go.kePDF p. 23
Resolution rule & assumptions

Any petroleum pricing schedule issued after 2026-10-19 and effective by horizon lists Nairobi automotive gas oil ceiling >250 KES/litre; exclude local noncompliant pump price.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: October schedule rises to 227.76; imported distillate, FX and administered pass-through create upside tail.

Public-branch input sensitivity: 23–52%. Input sensitivity is not a confidence interval.

Evidence context: S0488. Source links support audit; they do not estimate probabilities.

39%
J · uncalibrated
F041 / Cumulative event

Will the FAO Food Price Index exceed 150 in any post-cutoff month by the horizon?

Resolver: www.fao.orgPDF p. 23
Resolution rule & assumptions

FAO nominal monthly all-food index 2014-2016=100 >150 for a month November 2026 onward and completed by horizon.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: September 136; fuel, fertilizer, weather and export restrictions can co-move, while large global production buffers resist extreme outcomes.

Public-branch input sensitivity: 24–54%. Input sensitivity is not a confidence interval.

Evidence context: S0378, S0407. Source links support audit; they do not estimate probabilities.

41%
J · uncalibrated
F062 / Cumulative event

Will global utility battery system turnkey average cost fall below US$100/kWh in a post-cutoff annual survey by horizon?

Resolver: www.irena.orgPDF p. 26
Resolution rule & assumptions

BNEF or IRENA source explicitly citing global volume-weighted turnkey battery energy storage-system price, battery plus balance of system scope comparable to 2025 survey, <100 USD/kWh; exclude cell-only and China-only lowest bid.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: 2025 average below 120 and 31 percent fall; commodity/logistics shock and tariffs could slow decline.

Public-branch input sensitivity: 32–67%. Input sensitivity is not a confidence interval.

Evidence context: S0390, S0389. Source links support audit; they do not estimate probabilities.

53%
J · uncalibrated
F089 / Cumulative event

Will METR publish a frontier-model 50%-task horizon of at least 48 hours?

Resolver: metr.orgPDF p. 31
Resolution rule & assumptions

YES on a dated METR technical release reporting a point estimate >=48 human-work hours on a named software-task suite, with at least 30 tasks having estimated human duration >=8h. Preserve method version; saturating estimate without finite point estimate does not qualify.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: Long-task capability is advancing, but measurement ceiling, estimated human baselines and reward hacks create large uncertainty.

Public-branch input sensitivity: 52–85%. Input sensitivity is not a confidence interval.

Evidence context: S0587, S0588. Source links support audit; they do not estimate probabilities.

74%
J · uncalibrated
F091 / Cumulative event

Will a peer-reviewed trial in at least two African countries show improved patient outcomes from clinician-facing generative AI?

Resolver: pubmed.ncbi.nlm.nih.govPDF p. 31
Resolution rule & assumptions

YES for published prospective randomized trial, >=1000 total patients, >=2 African countries, prespecified patient clinical outcome with corrected two-sided P<.05 and effect favouring AI. Documentation, vignette accuracy and uncorrected secondary outcomes do not qualify.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: The Kenyan Penda trial improved process measures but did not meet its primary treatment-failure endpoint.

Public-branch input sensitivity: 15–39%. Input sensitivity is not a confidence interval.

Evidence context: S0512, S0542. Source links support audit; they do not estimate probabilities.

28%
J · uncalibrated
F094 / Cumulative event

Will the cost of a frozen Pawanax-like evaluation workload fall by at least 50%?

Resolver: www.anthropic.comPDF p. 31
Resolution rule & assumptions

Freeze 100 anonymized clinician-approved synthetic workflow tasks and an expert pass threshold at baseline. YES if a generally available commercial API meets that same threshold at <=50% baseline measured invoice cost including retries, tools and review time valued at fixed wage. Until baseline is measured, keep UNRESOLVED rather than using token-price comparisons.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: Constant-quality benchmark prices are falling rapidly; production retries and human review can offset token-price reductions.

Public-branch input sensitivity: 50–84%. Input sensitivity is not a confidence interval.

Evidence context: S0518, S0519. Source links support audit; they do not estimate probabilities.

73%
J · uncalibrated
F114 / State at deadline

Kenya headline CPI inflation is at least 8.0% year-on-year in the latest monthly KNBS release available at the horizon.

Resolver: www.knbs.or.kePDF p. 35
Resolution rule & assumptions

Use latest KNBS national CPI annual percentage rate whose reference month ends on or before horizon; resolve true iff rate >=8.0. Later revisions ignored after 60-day resolution window.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: Food and transport pressures can transmit to household affordability; 6.8% Sep2026 baseline. Analyst prior.

Public-branch input sensitivity: 24–48%. Input sensitivity is not a confidence interval.

Evidence context: S0491. Source links support audit; they do not estimate probabilities.

38%
J · uncalibrated
F116 / State at deadline

KES/USD official daily indicative exchange rate is at least 145.00 at the horizon.

Resolver: www.centralbank.go.kePDF p. 35
Resolution rule & assumptions

CBK indicative mean KES per USD rate on horizon or last preceding business day; true >=145.00. Avoid reciprocal conversion.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: Imported equipment, cloud and USD debt increase FX mismatch. Analyst prior.

Public-branch input sensitivity: 22–44%. Input sensitivity is not a confidence interval.

Evidence context: S0490, S0262. Source links support audit; they do not estimate probabilities.

35%
J · uncalibrated
F120 / State at deadline

Kenya trailing 12-month formal remittance inflows are below their preceding 12-month total.

Resolver: www.centralbank.go.kePDF p. 35
Resolution rule & assumptions

Sum CBK monthly USD formal remittance inflows for last12 complete reported months with reference months on/before horizon and preceding12; true if recent < preceding. Max90-day publication lag.

Observe qualifying post-cutoff event by deadline for event questions; latest eligible reference period at deadline for states. Allow 90 days for scheduled statistical release, record period and vintage. No qualifying new release -> unavailable, not NO. Resolve events already confirmed YES without waiting. Two independent primary records for contested geopolitical claims; otherwise disputed/unresolved.

Rationale: Diaspora employment/policy shocks matter to health contributions; no guaranteed national pool capture.

Public-branch input sensitivity: 20–42%. Input sensitivity is not a confidence interval.

Evidence context: S0262, S0307. Source links support audit; they do not estimate probabilities.

33%
J · uncalibrated

Source: world forecast, pp. 17–38; frozen 19 Oct 2026. Questions overlap; probabilities must not be added. Missing annual observations remain unresolved. Exact measurement, release-window and revision rules appear with each question and in the JSON register.

Method, scoring and the previous forecast freeze

Event timing uses a constant-hazard approximation; state probabilities use a separate odds model. Seasonality, policy responses and structural uncertainty remain limitations. Score resolved binary outcomes with Brier and log loss; continuous forecasts with quantile loss or CRPS and coverage. Compare same-vintage benchmarks. No predictive skill or calibration is claimed.

The 12 questions published on 16 October are preserved below as a separate, unscored historical freeze. The 19 October framework changes definitions; it does not retrospectively overwrite them.

16 Oct / F1 · 65% — Brent settles above $90 on 15 Jan 2027. Resolution: 2027-01-15. Unscored.

16 Oct / F2 · 25% — Brent settles above $120 on any day before 15 Jan 2027. Resolution: 2027-01-15. Unscored.

16 Oct / F3 · 40% — Fed raises rates again by its January 2027 meeting. Resolution: 2027-01-31. Unscored.

16 Oct / F4 · 30% — Central Bank of Kenya raises the CBR before 28 Feb 2027. Resolution: 2027-02-28. Unscored.

16 Oct / F5 · 55% — Kenya December 2026 CPI ≥ 7.0%. Resolution: 2027-01-05. Unscored.

16 Oct / F6 · 80% — WHO Bundibugyo Ebola PHEIC still in force on 31 Jan 2027. Resolution: 2027-01-31. Unscored.

16 Oct / F7 · 20% — Kenya confirms a locally acquired Bundibugyo case by 31 Jan 2027. Resolution: 2027-01-31. Unscored.

16 Oct / F8 · 60% — FAO Food Price Index for December 2026 above 136.0. Resolution: 2027-01-08. Unscored.

16 Oct / F9 · 30% — US 10-year Treasury closes above 5.50% before 31 Jan 2027. Resolution: 2027-01-31. Unscored.

16 Oct / F10 · 35% — Nasdaq closes ≥ 15% below its 2026 high before 16 Apr 2027. Resolution: 2027-04-16. Unscored.

16 Oct / F11 · 85% — Two of OpenAI, Anthropic, Google ship a new flagship generation before 31 Jan 2027. Resolution: 2027-01-31. Unscored.

16 Oct / F12 · 45% — Kenya reports ≥ 50 flood deaths in the Oct–Dec 2026 short rains. Resolution: 2027-01-15. Unscored.

05 / Economic engine

Show the assumptions.
Then stress them.

Targets are built from payers, units and prices. Corporate revenue, protected contributions, credit principal and research grants remain separate.

Management revenue targets / USD

A bottom-up commercial ambition.

TARGETS · NOT ACTUALS
$0.45m
Year 1
$1.8m
Year 2
$5.5m
Year 3
Clinic subscriptionsInstitutional workflowsConsultancy projectsEnterprise academy cohortsImplementation / contracted platform fees
Inspect the revenue math
Annual revenue = paid units × USD price × billing periods. Clinics / institutions use average paid units.
StreamYear 1Year 2Year 3
Clinic subscriptions50 × $250 × 12
$150,000		200 × $300 × 12
$720,000		600 × $350 × 12
$2,520,000
Institutional workflows5 × $1,500 × 12
$90,000		20 × $2,000 × 12
$480,000		60 × $2,500 × 12
$1,800,000
Consultancy projects6 × $20,000 × 1
$120,000		12 × $25,000 × 1
$300,000		20 × $30,000 × 1
$600,000
Enterprise academy cohorts6 × $10,000 × 1
$60,000		15 × $15,000 × 1
$225,000		20 × $20,000 × 1
$400,000
Implementation / contracted platform fees1 × $30,000 × 1
$30,000		1 × $75,000 × 1
$75,000		1 × $180,000 × 1
$180,000

Assumed variable cash cost: 35% of revenue. Collections lag: two months. Uncollected revenue: 3%. Target prices are planning assumptions, not a replacement for public product pricing. No additional SPV profit enters parent cash.

Source: investment manifesto, p. 17

Harambee / annual steady-state illustration

Follow every dollar.

01 / Sponsored transfers$600m

$500 / household / month

02 / Protected contributions$12m

2% explicitly authorized

03 / Eligible care cost$7.2m

2 lives / donor · $36 / life / year

04 / Restricted balance$3.84m

After care + 8% administration

Proposed platform service fee: $480,000 / year. 4% of contributions, already inside the 8% administration allowance ($960,000). Transfers and protected contributions are excluded from corporate revenue. Fees require a separate contract.

All inputs are design assumptions, not demand evidence or actuarial pricing. Enrollment of 100,000 by month 36 is a target. The lawful choice between discretionary assistance and contractual insurance precedes receiving funds.

Source: investment manifesto, p. 9

Devices / separate credit economics

Access needs
sound underwriting.

Start with approved third-party devices. Scale after at least two aged, paid cohorts demonstrate service quality and positive contribution after losses.

Source: investment manifesto, p. 11
Illustrative base / kit+$11.59
Joint adverse / kit−$50.33

$100 landed cost · $120 cash price · $20 down · 12 × $9.25 installments. Performing total: $131. Nominal APR 19.72%; effective annual 21.60%.

Base: 8% annual PD, 85% LGD, 12% KES funding, 80% advance and $10 service. Adverse: 25% PD, 95% LGD, 18% funding, $15 service plus 20% supplier-cost and 20% FX shocks. All terms are illustrative, not quotes.

Protected health funds stay outside device recourse. Essential device functions cannot be disabled as a collection penalty.

Corporate liquidity / supplied monthly model

The downside belongs in the room.

Month 36 cash$7.92m
Minimum modeled cash$1,210,000

All $24m equity closes on schedule; revenue targets and base cost assumptions hold.

Monthly corporate cash · USD millions · Base plan
Base plan: corporate cash by month0.06.212.41122436Month from first evidence phase

Actual supplied model path, locally reproduced. The dashed line marks zero cash. Committed funding and actual costs determine the operating decision.

Inspect monthly cash values
Corporate cash · USD · Base plan
MonthCash
1$4,653,542
2$4,307,083
3$3,997,000
4$3,686,917
5$3,376,833
6$3,066,750
7$2,756,667
8$2,446,583
9$2,136,500
10$1,826,417
11$1,516,333
12$1,206,250
13$8,606,792
14$8,007,333
15$7,517,000
16$7,026,667
17$6,536,333
18$6,046,000
19$5,555,667
20$5,065,333
21$4,575,000
22$4,084,667
23$3,594,333
24$3,104,000
25$12,422,417
26$11,740,833
27$11,358,333
28$10,975,833
29$10,593,333
30$10,210,833
31$9,828,333
32$9,445,833
33$9,063,333
34$8,680,833
35$8,298,333
36$7,915,833

Figures reconcile to the manifesto and a local numerical rerun of the supplied finance model. Before forward six-month committed cash falls below fixed obligations: freeze optional hiring and capex, secure a bridge or reduce burn. Restricted funds cannot bridge parent cash.

Source: investment manifesto, p. 18

06 / Evidence & accountability

Big ambition.
Precisely stated proof.

A useful investment page makes it easy to see what exists, what is proposed and what would stop the next release of capital.

Recorded platform activity

Internal aggregate snapshot · 2026-10-16. Operational activity, not paid customers, audited traction or clinical outcomes.

1,776internship & academy applications
12,158WhatsApp agent messages handled
5,329epidemic-surveillance alerts generated
259registered platform accounts

Source: existing internal production-count snapshot. Independent product and financial verification remains due.

Commercial proof

Invoices, collections, renewals, customer value and actual acquisition / support cost.

Clinical proof

Qualified clinical owner, prospective utility and safety evaluation, incident review and patient appeal.

Capital proof

Entity and IP chain, negotiated instruments, segregated accounts, lender terms and insurer pathway.

Scientific proof

Native-species provenance, permissions and benefit sharing; blinded external replication before lab expansion.

African discovery earns its infrastructure.

The $6m restricted research envelope: $2.5m staged robotics, $2m native compound screens, $1m replication and staff, $0.5m noninvasive BCI. Corporate BCI exploration is a separate $0.5m line. External assays come first; no discovery windfall funds the base plan.

Source: investment manifesto, p. 15
01 / RESEARCH & PLANNING

The investment manifesto

Capital, economics, milestones & ownership

23 pages · PDF · approval required02 / RESEARCH & PLANNING

The world forecast

Six horizons, two branches & full source catalogue

82 pages · PDF · approval required03 / RESEARCH & PLANNING

The forecast audit

What holds up. What still needs to be proved.

12 pages · PDF · approval required

Readable by people.
Inspectable by agents.

Versioned data, explicit units, stable identifiers and source pages. The same assumptions sit behind this page and its machine-readable investment brief.

Programme JSONForecast JSONSource catalogueAgent guide

Evidence quality matters more than source count: the report records 322 sources with opened text or excerpts and 291 indexed / access-limited sources across 154 domains. Those are not 613 independent datasets. The audit’s 5.07/10 is a desk-review score of the earlier report, not forecast accuracy. The supplied companion forecast and finance models were reproduced locally; matching outputs demonstrate reproducibility, not predictive skill. The full source package is restricted to approved investors.

07 / Questions worth asking

Clarity is part
of the proposition.

What is Pawanax raising?

A proposed $50 million programme over 36 months: $24 million corporate equity, $10 million conditional device warehouse capacity, $10 million protected health capital and $6 million restricted research grants. All funding is uncommitted. The first stage targets $12 million, including $6 million equity.

How do critical minerals, energy and fertilizers connect to healthcare?

Matawis supplies physical-risk and commodity intelligence. Mineral and battery supply affect device procurement and backup power; energy affects clinic uptime and transport; fertilizer affordability affects farm economics and food access. These are operating-risk transmission channels, not additional mining or fertilizer allocations in the $50 million programme.

Is the $50 million programme the company valuation?

No. Programme size is not valuation. Only the proposed $24 million equity purchases shareholder rights on negotiated terms. Device debt, protected health assets and grants have separate mandates and loss bearers.

Are the forecasts calibrated or proof of AGI?

No. The 19 October 2026 report contains 138 binary questions and 12 continuous parameters at six horizons. Probabilities and model envelopes are uncalibrated analyst judgments. The AGI-assumed branch is conditional planning, not evidence that AGI has been achieved.

What is the Year 3 revenue target?

$5.50 million of management target revenue: $2.52 million clinic subscriptions, $1.80 million institutional workflows, $600,000 consultancy, $400,000 academy and $180,000 other implementation or contracted platform fees. This is not achieved revenue. Device principal, health contributions, grants and discovery windfalls are excluded.

What must be proved before expansion?

Paid clinical renewals, independent safety and utility evaluation, lawful fund structure, informed consent, actual care costs, seasoned repayment cohorts, scientific replication and committed funding. Serious safety incidents or reserve breaches pause the affected activity.

The invitation

Back the institution.
Release capital against proof.

Join the first evidence phase of an African care infrastructure programme. Start with the $6m equity anchor, a bounded pilot or a separately governed facility.

Start an investment conversationMeet Pawanax

Nairobi, Kenya · Founder contact: Derrick Macharia
Proposed programme; all funding uncommitted. Programme size is not valuation. Research and planning materials; not an offer of securities. Terms, permissions and independent diligence remain open.