The complete Songbird model, opened up.
25 variables · 14 cost line items · 20 monetisation ideas · 4 preset scenarios. Every input is defined, benchmarked and stress-tested. Every cost is sourced. Every idea is quantified. Where the v0.4 spreadsheet has assumptions, they are surfaced — not hidden.
Executive summary
Songbird is a wholesale connectivity initiative bridging Europe and Africa across up to 20 Points of Presence in 11 countries. The v0.4 financial model targets dual outcomes — mission impact (schools connectivity routed via Giga at deeply discounted rates) and commercial credibility (a path to wholesale-grade EBITDA margins).
Preset scenarios, side by side
Four pre-configured runs of the model. Each varies the funding stack and unit economics to express a different strategic posture — from minimum-viable proof to full EU public-stack deployment.
| Scenario | Peak fund | Y3 rev | Y10 margin | Payback | IRR | Schools Y10 |
|---|---|---|---|---|---|---|
| Base v0.4 | $34.8M | $0.4M | -5.7% | — | N/M | 23,408 |
| Min Viable | $0.0M | $0.3M | -4.6% | — | N/M | 4,290 |
| Patient Capital | $23.5M | $2.4M | -2.1% | — | N/M | 28,600 |
| EU Public Stack | $25.4M | $1.0M | -4.4% | — | N/M | 23,408 |
Every variable, defined
The model has 25 input variables organised in three sections. Each card lists the definition, why it matters, suggested low/high scenarios, certainty, sources and the principal risk if the input is wrong.
Number of Points of Presence (POPs) where Songbird operates network equipment and terminates capacity. Each POP is a physical facility with router, switch, optical equipment, IXP connection, and (usually) IRU termination on a subsea/terrestrial cable.
Discount applied to traffic routed through Giga-mediated procurement (vs commercial wholesale rates). In v0.4 this is the cell labelled "Giga IPT discount to market price" — set to 0.2 (20% of market) which effectively means 80% discount.
Percentage of total capacity sold at full commercial wholesale rates (to ISPs, hyperscalers, banks, etc.) rather than via Giga-discounted school routing. v0.4 base case assumes 0% — all traffic flows through Giga firewall.
Percentage of circuit capacity purchased as IRU (Indefeasible Right of Use — 15-year prepaid) vs leased monthly. IRU = high capex, low ongoing opex. Lease = no capex, high opex.
Months of prepaid revenue from anchor customers (UNICEF, large telco, hyperscaler) committed at project start. Treated as cash-in at month 0, reduces peak funding need.
Capex per POP for active equipment: router (Cisco/Juniper/Nokia 100G platform), switches, optical transceivers, power systems. Excludes IRU costs and facility leases.
Average prepaid IRU cost per POP for 100G subsea/terrestrial circuit (15-year term). v0.4 total IRU spend is $17.5M across 20 POPs = $875k avg, but varies widely by route.
Annual cost per POP for leased (non-IRU) capacity. Used as alternative to IRU when capital is constrained or route uncertain.
List price per school per year for connectivity service (before Giga discount). The actual revenue per school = price × (1 − Giga discount). At default 80% discount and $600 list, schools pay $120/yr.
Average number of schools mapped in the catchment area of each POP. Highly variable: Lagos catchment ~10,000+; rural Niger ~1,000.
Percentage of mapped schools in catchment that are paying customers by end of Year 3. Drives ramp-up curve. Used by penetration curve which assumes growth after Y3 to ceiling.
Annual percentage decline in IP transit (IPT) unit pricing. Reflects industry-wide bandwidth deflation. v0.4 default ~11%/yr matches historical African wholesale IPT trends.
Annual commercial wholesale revenue per POP in Year 1 (ramps with maturity). Only applies to the commercial mix percentage.
Allocated staff cost per POP per year. Includes NOC, field engineers, country presence, partner management. v0.4 has $528k Y1 (4 staff) rising to $782k Y10 (~6 staff) — i.e., total not per-POP.
Local overhead per POP: legal, regulatory, accounting, country office, travel. Separate from central HQ overhead.
HQ overhead: executive team, finance, legal, compliance, board governance, audit. Foundation governance adds to this.
Annual inflation applied to staff and overhead costs. v0.4 uses ~4% (staff goes from $528k Y1 to $782k Y10 = ~4.5%/yr).
Working capital as percentage of annual revenue. v0.4 zeros this with comment "Not yet calculated — assume small." Telecom norm 10–15% (DSO + inventory + prepayments).
Annual depreciation of local currencies vs USD. v0.4 model has FX column but flagged "Not currently used." Real exposure exists — costs USD/EUR-denominated, revenue partly local currency.
Effective tax rate on positive operating income. v0.4 uses a simplified flat 1% of revenue (cell formula = revenue × 0.01) — which is wrong; real corporate tax in SSA is 25–35%.
EV/EBITDA multiple at Year 10 for terminal value calculation. Used in IRR computation.
Total grant financing across the project. Non-dilutive, non-repayable. v0.4 base case has $20M grant in Y1.
Senior concessional debt from DFIs (EIB, EAIF, IFC, Proparco). Below-market rate, often 12–20 year tenor.
All-in annual interest rate on concessional debt.
Equity from private investors (PE, DFIs as equity, family offices). Dilutive but permanent capital.
Sensitivity — top 5 levers
Across the full parameter range, here is how peak funding need responds to each of the five most consequential inputs. Steeper curves = more leverage.
Cost line items, grounded
Every cost item from the v0.4 spreadsheet, with per-unit benchmarks, industry references, optimisation opportunities and an estimate of potential saving.
- Negotiate with Medusa consortium for capacity at construction-time pricing (vs ready-for-service pricing). Could save 20–30%.
- IRU only the routes with confirmed anchor demand (Nairobi-Marseille, Lagos-Lisbon, Algiers-Barcelona). Lease secondary routes for first 3 years.
- Join 2Africa consortium for upgrades — get capacity at member pricing rather than wholesale.
- Bundle IRU purchases across the 20 routes for volume discount — single procurement vs route-by-route.
- Co-build with regional terrestrial operators (Liquid, MainOne, Wingu) — sharing fiber cost.
- Use mobile operator backhaul networks where they already exist (towers + fiber-to-tower).
- Negotiate with national fiber operators (Kenya's NOFBI, Egypt's WE) for government-grade IRU pricing tied to public-good positioning.
- White-box switching (Edgecore, Mellanox) for non-anchor POPs — 50-70% cheaper than Cisco/Juniper.
- Refurbished/end-of-life gear for tail POPs (with vendor maintenance contracts) — 40-60% capex reduction.
- Vendor financing from Nokia/Cisco/Juniper — 5-year amortisation moves capex to opex.
- OEM revenue-share or strategic vendor partnership: Cisco/Nokia loss-leader pricing in exchange for case study + reference customer status.
- Partner with existing licensed operators in early markets — Songbird operates as wholesale supplier to their retail licence.
- Negotiate "infrastructure-as-a-service" exemptions where regulators offer them (Rwanda, Senegal have).
- Use Cures Gateway / Foundation status to argue for development-mission licence discounts in donor-priority countries.
- Bundle O&M across the entire footprint — single SLA covering 7 circuits cheaper than 7 separate contracts.
- Direct relationship with cable consortium rather than intermediate operator markup.
- Founding-member status at smaller IXPs (Niamey, Bizerte) — typically free for foundational members.
- Trade transit-for-port arrangements: provide international transit to small African ISPs in exchange for free IXP port.
- Bundle ports across operator-managed IXPs (AfricaINX consortium).
- Peering instead of paid transit at Marseille, Lisbon, Barcelona, Marseille — large European IXPs (DE-CIX, AMS-IX, FranceIX) offer free peering with willing partners.
- CDN/hyperscaler private peering: Cloudflare, Akamai, Meta, Google all offer free private network interconnects in exchange for hosting their content.
- Tier-1 dual-source strategy at scale (Cogent + Lumen) drives 30-40% discount vs single-source.
- Co-maintenance arrangements with partner operators sharing fiber.
- Outsource NOC to existing 24/7 NOC operator (Vodacom, Liquid, Seacom) — pay per ticket rather than build internal team.
- Partner-operated POPs in tail markets (Niger, Malawi) — local partner provides field ops in exchange for partnership rev share.
- Engineering centralised in Spain (Pablo's base, cheap vs UK/FR) with regional liaisons.
- Use Africa50 talent network for in-country leadership hires.
- Open-source stack (Netbox, LibreNMS, Suzieq) — saves $50–80k/yr on monitoring/inventory.
- AI-native ops tools (potential category — Pablo's Agentic AI angle could replace staff).
- Equity stake in regional data centre operators (Raxio expanding) — locks in below-market rates in exchange for capital commitment.
- Co-location swaps with peer operators — Songbird hosts in their DC in exchange for their hosting in Songbird POPs.
- Carrier-neutral host vs telco-owned DC: ~30% pricing gap.
- Foundation governance overhead spread across multiple Songbird-style initiatives.
- Outsourced legal/finance/audit in operating countries — partner with Big4 or regional firm at fixed monthly retainer.
- Prepayment-favoured pricing — discount for annual prepay.
- Government channel: structured PO + escrow with donor backing.
- Factor receivables via DFI-backed facility.
- Foundation-owned commercial entity may qualify for development-stage tax exemptions in some operating countries.
- Swiss Foundation HQ enables transfer pricing optimization (but with substance requirements).
- Negotiate tax holidays in priority markets with development case.
20 monetisation & cost-reduction ideas
Each idea is a separately-toggleable lever in the simulator. The report lists them by tier (1 = highest potential) with stakeholder, mechanism, monetary parameters, ramp time and effort.
Sell premium 100/400G transit capacity on Songbird routes, particularly Medusa-aligned Mediterranean axis. Hyperscalers need AI training data and cloud regions backhauled to EU. Songbird becomes the trusted-route alternative to Chinese-built capacity.
Host CDN edge nodes at Songbird POPs in exchange for free transit (eliminates IP transit fees) and rack rental income. Each CDN pays $50–150k/yr per POP they want to be in.
Sell SLA-grade premium connectivity for cross-border payments, reconciliation, real-time fraud systems. 2-3x baseline IPT pricing for guaranteed latency and 99.99% uptime.
DIB structure: investors pay upfront for school connectivity outcomes; donors pay back with premium when verified outcomes achieved. Songbird is the delivery operator.
Multi-year prepayment for committed capacity. Effectively non-dilutive bridge financing. IFC has structured similar arrangements before.
Vendor provides equipment on 5-year amortisation; Songbird pays via opex rather than upfront capex. Standard for emerging-market telecom.
Host their content caches at Songbird POPs for free; in exchange Songbird saves on IPT (their traffic doesn't use paid transit). Indirect monetisation via cost reduction.
EU-compliant African gateway for sovereign data flows. Post-Russia and Houthi cable attacks, premium for route-diverse trusted infrastructure. Sell capacity reservation contracts to EU defence/intelligence and African governments.
Lease unused rack space at Songbird POPs to other operators. Passive infrastructure sharing model that's standard in African telecom.
Lease excess fibre pairs on terrestrial spans Songbird already operates. Asset-light revenue (no incremental capex).
Songbird operates as regulated utility — government pays per-school subscription via national education budget. Like rural electrification PPPs.
Bundled connectivity + health data routing for cross-border teleconsultation. Maghreb-Europe corridor particularly relevant (European-trained doctors do remote consultations).
Sell dedicated capacity to research networks at non-profit pricing but meaningful scale. Connects African universities to European research infrastructure.
Replace satellite backhaul at remote sites with Songbird fibre + last-mile wireless. Mining operators currently pay $5,000–20,000/month for VSAT.
Dedicated capacity + edge compute for smart port operations. EU funding for African port modernisation (Global Gateway) creates buyer alignment.
Host or operate national IXPs where they don't yet exist (Niger, Malawi, parts of Maghreb). Earn per-port fees + strategic position.
Lower-rate debt linked to schools-connected metrics. Equivalent to ~50bps interest rate reduction vs straight commercial debt.
Enterprise WiFi + IPT for hotel chains. High-margin per-property revenue.
Songbird's network replaces satellite/GSM backhaul → lower emissions. Verified carbon credits + sustainability-linked loan discount.
Premium routing for diaspora-driven services. Africa-Europe remittance corridors particularly relevant given Maghreb-EU axis.
Sanity checks at base case
Each check compares a model output against an industry benchmark or internal consistency rule. Run with the v0.4 default parameters.