For a few years now, a group of us have tried to explain that the biggest challenge in African democracies like Ghana, Nigeria, and Kenya is not really the POLITICS. The real mess is in POLICY. We call this KATANOMICS. But we have done most of our explaining using essays. Everyone tells us, however, that reading nowadays is hard! People are too swamped and busy. So, we are trying our hands at using videos and photos. 😊 First video: piped.video/cj9pUzTr2vQ We are new to this so forgive the quality and focus on the message. But we are very welcome to all feedback. Keep them coming.

Essential reading (and digesting)

A “bad waist” story confronts you everywhere you go in Africa. In a matatu or in a trotro, at a church crusade or a Jumah service, on radio or on TikTok, waist pain is on every lip; herbal balms and potions are on every other billboard; and “sexual weakness” follows closely in tow. We need to have a family conversation about this crisis, but it starts with knowing that “bad waist problems” in these parts are often actually about “lower back pain.” Lower back disease is the leading cause of years lived with disability worldwide, and the World Health Organization expects the problem to grow fastest in Africa and Asia. Pooled African studies put the share of people suffering back pain in any given year at 57 per cent. In Ghana, KNUST researchers found it in 81.6 per cent of health workers surveyed in Greater Accra hospitals. In Obuasi, two in three underground miners had suffered an episode within twelve months, and among Volta farmers, the biggest pain complaint concerned the lower back. Like the mounting non-communicable disease (NCD) mess in general, the African health system, long primed for infectious diseases, struggles to deal with lower back pain. And it is not just because there is just about 1 physiotherapist per 100,000 people, the level of seriousness attached is much lower. But as individuals we can’t wait until these grand systemic issues get fixed. There is one simple thing each one of us can start doing right: posture and movement. Health professionals say that a healthy spine bends in a gentle S. The worst thing you can do, therefore, whether sitting, sleeping, or walking, is to slouch, crouch, or bend in ways that fold the spine into an unnatural shape for long stretches of time. Like prolonged C-spine stooping. Our culture has long celebrated that natural arched-back S curve - waist beads did not become an art form by accident - and evolutionary psychologists suggest the attraction may be rooted in its role in safe childbearing. But the world we have built folds our spines the wrong way: low stools in the market, short-handled hoes on the farm, cramped classroom benches, the long crawl through Accra traffic, hours hunched over phones, all wrong on so many levels. Our grandparents walked and squatted far more than we do. We sit and stoop. We need movement and variation. And, as much as possible, we must keep the spine in its natural, slightly arched, shape. Leaders have a role to play here, sure. Like setting ergonomic standards for school furniture and mattresses. Or following the lead of courts that outlawed the short-handled hoe in 1975 after years of ruined backs among farmworkers in the US. People will tell you to “grow a spine” in the face of hard decisions. Well, you can’t. You get only one in life. The hard choice is to take care of it.

Global progressives are celebrating Lula's stunning dash from behind to win the Presidency in Brazil. The huge sigh of relief is over the Amazon, inclusion & renewed hope in multilateralism. But expectations need to be tempered. Power in Brazil has tilted heavily to the Right.

New At Issue briefing from @BBSimons in #BWObserver Autumn26 shows how @WorldBankGroup's use of pre-launch data in #Mission300 accounting figures muddies water on initiative’s impact tinyurl.com/M300AtIssue

IMANI VP, @BBSimons, has proposed a lower-cost alternative to Ghana’s planned $270 million GhanaChat project, arguing that a secure AI assistant for the civil service could be a more practical starting point for the country’s AI ambitions. myjoyonline.com/bright-simon… #IMANI

IMANI VP, @BBSimons, has raised concerns about Ghana’s national artificial intelligence strategy, questioning its computing capacity targets, infrastructure projections and the proposed $270 million investment in GhanaChat. myjoyonline.com/bright-simon… #IMANI #ArtificialIntelligence

IMANI VP, @BBSimons, estimates Ghana’s proposed GhanaChat data centre could require a US$61m annual subsidy, with operating costs projected at US$76m a year. He questions whether revenue from computing services can sustain the facility. myjoyonline.com/ghanachat-da… #IMANI

India faced the same choice in 2024 and chose not to own the hardware. The IndiaAI Mission (about $1.2B over five years) empanelled private cloud providers instead. Around 34,000 GPUs sit on its portal, and approved users pay under ₹100 (about $1.10) a GPU-hour, with the state covering part. That moves the utilisation risk in your point 37 onto the providers, since the state only pays for hours someone uses. The open question is the one you raise: there's no public usage figure yet, only 9.3M GPU-hours sanctioned across 237 projects.

Shorter version of the previous post. 😀 Ghana's AI compute center and local model (GhanaChat) plans will blow your mind! See short summary below. 1. The AI datacenter and the "AI Fund" will cost about $270 million. Plus extra for registries and apps. Could be happy season for ICT contractors. 2. Unfortunately, the AI strategy's technical and financial projections are totally off. See attachments. 3. The target for how much computing capacity the country needs for next year (2027) is indicated as "1021 FLOPS" and the 2035 target as "1025 FLOPS". 4. Taken literally, it means the Ministry would like the entire computing capacity of the country in 2027 to be equivalent to a 1970s HP desktop computer. Or 2% of the capacity of the 1980 Intel math chip. 5. Howzat? 6. It's clear from the math that they meant to write 10²¹ operations a second. And 10²⁵ FLOPS (10 yottaFLOPS.) 7. How did the error survive civil service and cabinet review plus big presidential launch? I blame Katanomics. Policy matters simply fall through the cracks as big political banners are hoisted for public attention. 8. But this is not a typo matter. 9. Because read properly as 10²¹ FLOPS, we model 200k - 400k of Nvidia's newest chips & 600 - 1,100 MW of power all costing ~US$24 - 47 billion. 10. The high case is 2x the Ghanaian govt's entire programmed spending for 2026). Sounds realistic? 11. The US$250 million AI datacenter now being procured supplies about 1% of this 2027 target. 12. The strategy also projects "drastically increas[ing] Ghana's energy capacity by 0.5 to 1 OOM per year". "OOM" means "order of magnitude". 13. Literally means Ghana's electricity output expanding by 3x to 10x EVERY YEAR. 14. Did anyone in Cabinet not protest that this projection in 8 years would mean Ghana's electricity output growing by 10,000x? 15. The strategy also wants a "1010 compute" (no units) AGI lab. One translation: a lab with computing capacity of 10¹⁰ FLOPS. 16. For perspective, an iPhone 7 has about 350 gigaFLOPS of capacity. 17. Ghana's 2035 target of 10²⁵ is 160 times the entire global stock of AI compute today. Even if chips grow 1.4 times more energy-efficient every year, it would need 377 GW, 65x Ghana's current installed power generating capacity. 18. The AI compute center's budget, per our modelling, maps to about 2,100 GPUs and 4.3 MW of IT load. That alone adds the equivalent of 90% of Ghana's entire commercial data-centre load in 2024 (4.8 MW). Running it costs about US$76 million a year, which is about six times the revenue of the whole commercial industry. 19. A utilised GPU-hour costs about US$22 at base-case demand, against a market rate of US$6.71 today and roughly US$4.50 averaged over the hardware's life. The base case thus implies a subsidy of about US$61 million a year. 20. Can't sovereignty be achieved with a $13 million AI assistant for the civil service based on local hosting & orchestrated open-weight models?

1. Readers of this page know by now that JoyNews held a forum on Ghana's AI strategy and prospects. And the Digital Minister and I got into a serious debate about his Ministry's plan to build Ghana's competitor to Gemini, Kimi, ChatGPT, Qwen, Claude, etc. 2. Ghana's competitor to these global AI platforms is called GhanaChat. 3. It will be trained and hosted in a brand new AI datacenter. 4. 48 registries pooling data from across the public sector will make this AI smarter about Ghana than the foreign ones. 5. The AI datacenter and the "AI Fund" will cost about $270 million. Expect more to be spent on the registries and the apps. A lot of contracts. You can bet that a lot of ICT contractors are very happy. 6. The question is, if we have a couple of hundred million dollars to spend on making AI deliver great benefits for Ghana, should we spend it on GhanaChat, a top-tier AI data center, and the data-pooling system? And should we do it according to the strategy the Ministry has devised? 7. Or could we do much better by more critically assessing all our strategic options and comparing them to the Ministry's? 8. My argument on the show was that I am not convinced about GhanaChat offering the best value for our money. 9. It would take many pages to explain my position fully. But in this post I will highlight just the most important points. 10. Attached are extracts from page 13, 50, and 68 of Ghana's AI strategy. 11. If you have even the barest exposure to AI, you can see a massive error. Can you? 12. The target for how much computing capacity the country needs for next year (2027) is indicated as "1021 FLOPS" and the 2035 target as "1025 FLOPS". 13. If this was indeed correct, it would mean that the Ministry would like the entire computing capacity of the country in 2027 to be equivalent to a 1970s desktop computer. Or 2% of the capacity of the 1980 Intel math chip. 14. Howzat? 15. I'm sure you can agree that this would be totally ridiculous if that was the Ministry's intent. So, what happened? 16. Somebody who didn't fully understand how FLOPS work must have been involved in the drafting. What was meant to be 10 raised to the power 21 operations per second (10²¹ operations a second) was written out as "1021". Same for "1025". 17. Why wasn't the error caught? Because in Ghana when doing "consultations", governments rarely engage independent and critical voices who will point out absurdities without fear of offending big men and women. 18. I am not being petty. There is a serious strategic confusion beyond mere typos. 19. Read properly as 10²¹ FLOPS, the Ministry is telling us that by next year it will have some 200,000 of Nvidia's newest chips in use in Ghana, 500 MW of power devoted to AI, all costing about US$24 billion (close to the Ghanaian government's entire programmed spending for 2026). 20. The US$250 million AI datacenter now being procured supplies about 1% of this 2027 target. 21. Howzat? 22. Clearly, no serious critique of the plan would have okayed such a vision. Throughout the ministry's own review, cabinet approval, and a high-profile media launch, it is absolutely clear that not one critical reviewer was engaged. 23. My first major point here is that this is classic KATANOMICS: the lack of critical policy audiences. This is exactly how a sugar factory that was "designed from the start to fail" was allowed to be built. Many programs and initiatives in Ghana fail for precisely this absence of critical review. 24. In the strategy, you find statements like: "government aggressively building AI compute clusters" and by courting hyperscalers, "to leverage the scaling laws towards building general purpose AI and AGI". Then: "drastically increase Ghana's energy capacity 0.5 to 1 OOM per year". "OOM" means "order of magnitude". 25. When I challenged the Minister on TV, he confidently rebutted by claiming that the Energy Minister/Ministry has assured him of feasibility. Ghana's energy Ministry is committing to increase Ghana's power output by 3x to 10x EVERY YEAR so that the Digital Minister can deliver his AI strategy? Did anyone in Cabinet not protest that this projection in 8 years would mean expanding Ghana's electricity output by 10,000x? 26. I can bet you any amount that the Energy Minister will deny making any such commitments in Cabinet. Can we have clarity? Can JoyNews check? 27. Page 68 of the strategy defines an "Artificial General Intelligence Lab" as "an AI lab with access to more than 1010 compute". No unit is given at all. In fact, as written it is totally meaningless. What they likely mean is that Ghana will support the establishment of a lab with computing capacity of 10¹⁰ FLOPS (10 gigaFLOPS). 28. The problem is that such a lab would be totally useless. To give you perspective, an iPhone 7 has about 350 gigaFLOPS of capacity. 29. What would be required to meet the Ministry's 2027 AI target? If for frontier model development, then the following spec would be reasonable: 400,000 16-bit Nvidia B300 GPU units, 1,114 MW of dedicated power (which is 26% of Ghana's peak demand). And the budget could easily hit US$47.2bn. For less frontier use, we can halve these numbers. 30. Counted in the 64-bit arithmetic that scientific supercomputers are ranked by, Ghana's 2027 target is 455 times the world's fastest machine, China's LineShine at 2.2 exaflops, and 53 times all 500 machines on the list combined. 31. Anyone with the barest exposure to Ghana's fiscal context can tell the Minister for free that this won't happen next year. 32. Ghana's 2035 target of 10²⁵ is 160 times the entire global stock of AI compute today. Even if chips grow 1.4 times more energy-efficient every year, it would draw about 377 GW, 65 times Ghana's current installed generating capacity. 33. I will be charitable and propose that the strategy's drafters may have been thinking about training FLOP (note: without the "per second" at the end). For example, GPT-4's training run used about 2.1×10²⁵ FLOP, the same order of magnitude as Ghana's 2035 endpoint. 34. Unfortunately, that doesn't save the strategy. Recast this way, the target collapses the other way. 10²¹ FLOP is about 12 days' work for one H100 chip at 16-bit precision. 10²⁵ FLOP is one GPT-4-sized training run, which the US$250 million centre could finish in about four months. On that reading the centre meets the 2035 target before its first anniversary. Not meaningful. 35. Given how confusing the overall strategic plan is, would you bank on the first AI compute center and GhanaChat product already being procured at about $270 million? 36. Here are some bare numbers: the budget, per our modelling, maps to about 2,100 GPUs and 4.3 MW of IT load. That alone adds the equivalent of 90% of Ghana's entire commercial data-centre load in 2024 (4.8 MW). Running it costs about US$76 million a year, which is about six times the revenue of the whole commercial industry. 37. A utilised GPU-hour costs about US$22 at base-case demand, against a market rate of US$6.71 today and roughly US$4.50 averaged over the hardware's life. The base case thus implies a subsidy of about US$61 million a year. 38. If the concern is about Ministers and their Directors uploading sensitive documents into Gemini or Kimi, then why not a simpler sovereign AI assistant for the civil service first? 39. That would need roughly 144 GPUs: about 7×10¹⁷ operations a second at 8-bit, US$13 million up front and US$6.7 million a year. That is under a thousandth of the 2027 target, and a far more realistic initial-term strategy. Does Ghana really need to train its base model when orchestrators now allow one to host their own data and switch across multiple advanced open-weight models to serve an Assistant or Agentic system? 40. Howzat?

“Death is the solution to all problems. No life, no problem.” --- Joseph Stalin, Soviet Strongman & Overlord (per Rybakov)

“Who builds schools to teach people how to use WhatsApp?” @BBSimons says policymakers should focus on removing barriers that prevent Ghanaians from benefiting from digital technology, warning that excessive regulation can obstruct innovation & livelihoods myjoyonline.com/who-builds-s…

My blog now has a second dashboard for you to explore Ghana's fiscal data going back to 2008. Pick any year and you get a plain-language summary of what the government collected, spent and borrowed. You can chart any revenue or spending line by year, quarter or month, in cedis, as a share of GDP or as a share of revenue. You can also compare what was budgeted with what was actually collected or spent, see how much of the revenue went to wages, interest and statutory transfers before anything else was funded, and track which taxes carry the budget. You can also download all the data as a csv to explore in your own environment.

OP-ED | The Forgotten History of Africa’s Credit Rating Push, By Bright Simons @BBSimons examines the history of Africa’s homegrown credit rating agencies and the challenges facing the Africa Credit Rating Agency (AfCRA), arguing that transforming the continent’s borrowing costs will require more than a new institution, including stronger financial markets, improved disclosure rules and effective regulatory frameworks. 👉👉🔗tinyurl.com/yw7mu8r5 @afreximbank @_AfricanUnion @APRMorg @sifisof @Adam_Alqali #AfCRA #FinancialSovereignty #AfricanUnion #Agenda2063 #DevelopmentFinance #AfricaDevelopment

As an organizational leader building an AI first culture across my organization, i can definitely confirm first-hand that the biggest challenges are exactly as Bright describes. For AI to transform national organizations need to to build the building blocks properly : 1. How do we approach data purity and verification in Ghanaian organizations? 2. Do national organizations have clear SOPs for ALL processes that can be translated into agentic workflows? 3. Do we have full observability into the development of agentic workflows to prevent manipulation of data and results? 4. Do we have the necessary motivation to go digital ONLY across our key national registries? 5. Do we have the necessary expertise to create efficient agentic workflows rather than simply adding on extra steps to already inefficient processes? I would argue that Hon’s @samgeorgegh best legacy would be to solve these problems first before investing all attention to the infrastructure that may not translate into effective results without these building blocks in place.

I brought the policy-activist perspective to a discussion on Joy News about what AI means for Ghana. The discussion featured industry leaders and Ghana's Minister of Digital Tech & Innovations. 1. I emphasised that AI is primarily about learning and teaching. Not in the classroom sense, but in every sense. 2. Most of Ghana's policy challenges of the last many decades have been about an inability to learn from previous mistakes and change course effectively. 3. AI delivers what it is taught. 4. Whilst AI has immense "structural intelligence", its level of "cultural intelligence" tends to be capped by a society's ability to teach it well. Most complex problem solving involving AI is about "workflows" rather than "artefacts". Embedding powerful and transformative AI workflows into organisations, however, require major changes to the culture of doing things. 5. Countries are no different. 6. The question then is: can societies that struggle to learn really teach Artificial Intelligence to solve their problems more effectively than they have been doing with natural intelligence so far?

@BBSimons keeping it real as always.👌

Top 10 Must-Know: Bright Simons on Ghana’s Petroleum Hub wp.me/pfOUlK-DFV @BBSimons

Occasionally, I miss the old school Ghanaian motivational speakers. Where are they? I miss nuggets like: "Do you know why pizza is round but it is put in a square box and eaten in triangles?"