$40 Million for Water Wells. So Where Are They?

MrBeast raised $40 million for water projects. Good cause, massive number, lots of coverage.

The answer depends on a question most people don't think to ask.

The assumption everyone makes

Wells in Africa are cheap. Everyone knows this. You've probably seen the fundraiser pages - "$50 gives one person clean water for life," "$10,000 builds a well for 500 people."

So the math feels obvious. $40 million divided by $10,000 equals 4,000 wells. Maybe more if you're efficient. Some people throw around numbers like 10,000 or even 100,000.

I have an engineering background. Numbers like that make me want to open a spreadsheet.

So I did.

Two ways to spend the same $40M

The first thing the model shows is that "cost per well" isn't a number - it's a range that depends entirely on how you run the operation.

I modeled two approaches.

The first is what I call the system model. This is how serious NGOs actually operate: country offices, local staff, monitoring and evaluation, maintenance programs, supply chains for spare parts. High overhead, but the wells are designed to keep working.

The second is the lean model. Local crews, rented rigs, minimal overhead, no international staff, clustered drilling sites to cut logistics costs. The goal is to drill as many wells as possible with the money available.

Same $40 million. Very different machines.

What the model actually shows

Here's where it gets interesting.

Under a realistic system model, $40 million drills roughly 1,400 wells. After accounting for failure rates over five years, about 975 of those are still working.

Under lean execution - best case, favorable geology, existing local infrastructure - you can drill around 6,000 wells. After failures, roughly 2,400 remain functional.

And the lean model serves more people at a lower cost per person.

On pure output metrics, lean wins. It's not close.

The first thing that looks wrong but isn't

Most people expect the "responsible" NGO model to win. Spend more per well, get better results.

That's not what the numbers show - at least not at first.

The lean model drills 4x as many wells, ends up with 2.5x as many functional wells after five years, serves more people, and costs less per person. If someone showed you these results without context, you'd defund every NGO's overhead budget immediately.

So why doesn't everyone just do lean?

The number that changes everything

That's the failure rate for water wells in Africa that operate without a maintenance system. Within five years, 60% of them stop working. This isn't a pessimistic estimate - it's from WHO data and backed by field research across multiple NGOs. Some studies put it higher.

The lean model already accounts for this. The 2,400 functional-well figure is after applying the 60% failure rate.

But here's what the spreadsheet doesn't fully capture: failure isn't a switch that flips at year five. It's gradual. A pump starts struggling in year two. A part breaks in year three, and nobody replaces it because there's no supply chain within 100 kilometers. By year four, the well is technically "functional" but produces a fraction of its original output.

The system model's 30% failure rate isn't magic. It's what happens when someone is paid to drive out and fix things.

The metric that actually matters

Cost per well is the wrong number. Cost per person is better, but still incomplete.

The right metric is cost per person-year of service - not how many people got water once, but how long that access actually lasted.

A lean well that serves 250 people for three years and then fails is worth less than a system well that serves 500 people for fifteen years, even if the lean well costs half as much to drill.

The lean model's cost advantage shrinks significantly once you factor in lifespan. The high-control model - fewest wells, most expensive per unit, highest durability - starts to look competitive.

A note on the lean assumptions

The 6,000 well figure requires some context.

The lean model uses a 6% overhead rate. That is the floor - it assumes a single-district operation, pre-existing community relationships, shallow water tables, and local drilling infrastructure already in place. Realistically, even a stripped-down operation runs 12-15% overhead once you account for basic coordination and compliance.

Under more conservative lean assumptions - $9,000 per well instead of $5,800 - the number drops to roughly 3,500-4,000 wells.

The 6,000 figure is what lean produces under optimal conditions. Those conditions exist in parts of Kenya, Uganda, and Malawi. They don't exist everywhere, and they don't last forever as a program scales.

So what does $40M actually build?

Somewhere between 680 and 2,400 wells that are still working five years from now.

The wide range isn't uncertainty in the math. It's the difference between building a drilling operation and building a water system.

The money was never the constraint. It never is at this scale. The constraint is everything around the hole in the ground - the maintenance technician who shows up when the pump breaks, the spare parts warehouse 40 kilometers away, the community committee that actually owns the asset.

Drill without those things and you're not building infrastructure. You're building a monument that works for a few years.

Full model and all calculations are available here.

I built this in Excel (converted to Numbers) using data sourced from NGO field data, WHO failure-rate studies, and real-world well-cost benchmarks from Sub-Saharan Africa. If you find an error, I want to know.