West Africa just appointed a clearing bank for its regional electricity market. Boring? It may be the most important infrastructure news of the year. Lines are not a market — settlement is. And the intelligence economy is next. New essay — The Corridor Market: blog.josephnganga.com/the-co…

1/ Nobody who kept their lights on last night ever signed a power purchase agreement. That one fact answers the sharpest objection to AI-infrastructure finance. New essay: The Compute Utility 🧵 2/ A data centre is financed over 15 years. Its customers buy compute by the hour. Power solved that mismatch a century ago — it invented the utility. The intelligence economy hasn't. 3/ The Compute Purchase Agreement is the contract. The compute utility is the counterparty. Who signs it, who never should, and why the institution gets built at the frontier — where demand is real but fragmented. 4/ It also answers what a guarantee must never do, why deflation helps the utility, and how Africa's climate knowledge — held in cultures, not databases — gets compensated, not extracted. Climate is the unlock. AI is the enabler. blog.josephnganga.com/the-co… #SovereignAI

1/ Kenya had the clean power. Microsoft wanted to build. A $1B AI data centre stalled anyway. Everyone blamed the grid. The real story is better — and it exposes the missing instrument in AI infrastructure finance. 🧵 2/ What stalled it: Microsoft/G42 asked Kenya's Treasury to guarantee BOTH the compute payments and the power supply. The Treasury said no. Unpopular opinion from someone who wants this facility built: the Treasury was right. 3/ The rule the global market just proved: lenders don't finance buildings, they finance contracts. CoreWeave's $8.5B GPU loan hit investment grade because of committed offtake — not the chips. In AI infrastructure, the offtake agreement IS the asset. 4/ Africa's regulated inference demand — banking, government, ID, health — is real, creditworthy, and legally can't leave the continent. But it's fragmented across hundreds of institutions. No single anchor tenant → nothing finances. 5/ Power solved this with the PPA. Vaccines solved it with the Advance Market Commitment. Nobody has done it for compute offtake. So: the Compute Purchase Agreement — pooled, cross-border, regulation-anchored. No sovereign signs a take-or-pay alone. 6/ And it inverts the Kenya problem: aggregated AI demand becomes the anchor client that finances POWER. Two liabilities a treasury couldn't hold become two solutions. AI demand, aggregated at scale, creates capital for power at scale. 7/ Instruments are invented where constraints bind hardest, then travel — M-Pesa was born in Nairobi and now moves value on every continent. Designed in Africa; deployed everywhere. Full essay: #AIInfrastructure

Technology companies are investing billions into lunar data centres while the continent with 30% of the world's critical minerals struggles to attract a fraction of that capital. Not because the risk is higher. Because the intelligence to assess it doesn't exist yet. The mispricing of emerging market infrastructure may be the largest capital allocation failure in the global economy. I wrote a founder's reflection on Year One. blog.josephnganga.com/buildi…

A year ago I founded AfCEN to fix it — an infrastructure intelligence platform connecting data across 55 countries. We started in energy. Then discovered that every tool we built demanded the next: minerals data, trade corridors, agricultural demand, regulatory compliance. Convergence isn't a strategy. It's an engineering requirement.

Capital is not the constraint. Intelligence is. The connective tissue between projects, data, and capital — across emerging markets — simply does not exist. Investors can't see the pipeline clearly. Governments can't benchmark their own assets. Institutions make consequential decisions with incomplete information. That's a market failure, not a funding gap.

I helped launch @EnergyAlliance — a $1.5B initiative with @RockefellerFdn, @IKEAfoundation, and @BezosEarthFund. Led a summit that generated $26 billion in commitments. Served as Special Envoy for a $50 billion+ electrification initiative with the @AfDB_Group and @WorldBankGroup. Every time: the capital showed up. The pipeline didn't move.

Eighteen years ago I was selling solar lanterns from the boot of my car in Kenya. This year, we are co-designing the pipeline architecture for a $10 billion AI fund with the @AfDB_Group and @AIHub4SD . Between those two points, one lesson kept repeating. 🧵

One strait shut down energy, semiconductors, food systems, and AI supply chains — simultaneously. That's not a coincidence. It's a design flaw. The technologies are converging upward. The physical inputs they depend on are concentrating downward — into fewer geographies, fewer hands. New piece on why the physical layer of the convergence is where the next decade of strategic value sits. blog.josephnganga.com/every-…

The biggest bottleneck in global infrastructure isn't capital. It's intelligence. 73% of African energy projects never reach financial close. Not because the money isn't there — but because no one can reliably map which deals are real and which are PowerPoint. I've spent 18 years inside this gap. Now I'm building AI systems to close it. Today I'm launching The Nexus Brief — a weekly newsletter at the intersection of critical minerals, energy, infrastructure finance, geopolitics, and AI. First edition: "I Built a 16-Agent AI Team That Runs My Company." Subscribe free: blog.josephnganga.com

3/3: Today: DFI analysts spend weeks pulling data that should take hours. Due diligence stretches from months to years. The missing piece isn't more studies. It's intelligence infrastructure. AI can aggregate project data across 15 ECOWAS states and surface what matches an investor's mandate. $24B is on the table. The question is whether we build the infrastructure to reach it. #PAS2026 #PoweringAfrica

2/3: These 75 projects live in planning documents — WAPP Master Plan, national strategies, ECOWAS frameworks. But investors need project-level data: regulatory status, offtake agreements, risk profiles, financial metrics. They need to know which projects are actually bankable.

1/3 West Africa has 75 priority energy projects needing $24 billion. The money exists. IFC, DFC, AfDB, Africa50 all have mandates. So why is the pipeline stuck? The problem isn't capital. It's visibility.

Ethiopia 🇪🇹 is turning unused hydropower into digital gold. 600 MW ⚡ = $55 M (18% of revenue) in 10 months[1]💰. That forex funds grid upgrades & next-gen tech 🤖. Not a gamble—a blueprint for others. 👉 Full story: africacen.org/2025/08/25/why… #Energy #AI #Crypto