Alex Odundo’s first machine was a disaster. Gears broke, bolts rattled off, and the wheel that was supposed to strip fiber from the sisal leaves he fed into it— the machine’s central component—turned unevenly, throwing the whole thing out of balance. Sisal is an agave relative smuggled into east Africa from the Yucatan peninsula in the nineteenth century. It is a large plant with sharp, sword-shaped leaves that splay out in a spiky rosette several feet across. It looks, as some cactuses do, like something you’d want to stay far away from. Kenyans plant sisal as livestock fencing, to deter their goats and cattle from breaking free of their intended range.

On paper, Odundo had designed a “sisal decorticator,” a machine to ease the tedium of transforming an unwieldy plant into fiber that could be used to make twine and rope. But the real thing, rendered in scrap metal by a high schooler with no formal training, did nothing but make noise.

Fifteen years and ten prototypes later, Odundo, now 30, has come up with a machine which he says could transform Kenya’s rural economy. “Sisal is a plant that can grow anywhere,” he told me on Skype, wide-eyed. It needs no labor between planting and harvest, and you can harvest the same plot for up to ten years. “You don’t need to apply fertilizer, and you don’t need to weed,” he said. At a time when Kenya has suffered from years of bad rainfall, sisal is a plant that flourishes without irrigation:

Usually, during winter, we are able to receive enough rainfall for our crops—maize, beans, millet, cassava, potato. In the past, these used to do well, but right now, the system is changing. We don’t get enough rainfall. If these crops cannot do well, what next?

Sisal is already traded in international commodities markets, and worldwide demand exceeds global supply by a large margin. So it seems like a natural cash crop: simply turning sisal leaves into fiber increases their value by a factor of ten. Car manufacturers use sisal to reinforce the plastics in car interiors; paper plants use it to strengthen the pulp for recycled paper; in the construction industry, sisal goes into roofing material and fiberboard. But in Kenya, save a few large processing plants on the coast, sisal simply grows, living barbed wire to keep livestock at bay.

The problem, for small farmers, is processing it. As a boy, Odundo recalled, sisal plants bounded his family’s land in the countryside outside Kisumu, Kenya’s third largest city, a dry, hilly area near Lake Victoria. From an early age, he was fascinated by all things mechanical. He fixed radios, watches, and bicycles for neighbors in the village of Nyabera, and drew blueprints for fun. Odundo paid for notebooks and school fees by helping his mother process sisal the old-fashioned way, pulling the leaves through scissor-like blades mounted on a tree next to their house.

Processing sisal by hand

In drought years, the family relied on sisal to make up the shortfall from bad harvests of maize and beans. They twisted the fiber into rope, or sold it directly to neighbors who wove it into hats, floor mats, and sacks for storing produce. But the process was incredibly labor intensive. Even with experience, a long day’s work yields only five dollars worth of fiber.

Until the rise of synthetic fibers in the 1960s, sisal was a mainstay of commercial agriculture throughout east Africa. At one point, Kenya, Tanzania, and Mozambique accounted for half the world’s sisal production. Demand for sisal plummeted with the rise of nylon, acrylic, and polyester. Prices fell, and most plantations went fallow. Now, as modern engineering has found new applications for the fiber, Brazil has emerged as the world’s leading producer, using large-scale farming and processing plants to produce nearly as much fiber as the rest of the planet combined. The market remains just out of reach for most Kenyan smallholders, who produce only 20% of the country’s sisal.

In 2010, Odundo participated in the first Maker Faire Africa in Nairobi, a gathering of inventors from around the continent. He exhibited the ninth prototype for his sisal decorticator, which caught the attention of the event’s organizers. Two years later, he came to the TED conference in Long Beach as a global fellow. There, standing in the bookstore, Odundo struck up a conversation with Jan Leeman, an investor based in Palo Alto who took an interest in his designs. “What he told me then—and it’s still true now—is that he has lots of orders and no means to fulfill them,” Leeman recalled. “As an investor, that’s exactly what you want to hear.”

Leeman said Odundo’s decorticator could serve as a bridge to the international market for sisal, giving farmers a way to participate in the global sisal industry without having a large plantation:

For a lot of countries throughout history, agricultural wealth is the starting point. You plant a crop that’s profitable and that money spreads to the rest of the economy. In Kenya, they’ve got some sisal, but they’d never plant a lot of it because they have no way to harvest it. How do you become a rich farmer if you can’t harvest?

With only six dollars worth of fuel, two people working with the diesel version of Odundo’s machine can produce up to thirty times as much fiber as someone working by hand, multiplying their output by a factor of fifteen. An electric version is even cheaper to operate, but that requires access to the grid.

Processing sisal with Odundo’s decorticator

One person works with a machete, cutting and stacking leaves to be processed, and another feeds them into a slot in the decorticator, which looks something like a small municipal mailbox on wheels. Inside, a toothed flywheel strips the fleshy green from the leaves, leaving only creamy yellow strands of fiber, like a horse’s tail. The byproduct, bright green crumbs of sisal flesh, can be used as fodder for cattle and goats, or as compost to boost water retention for other cash crops. Mounted on a two bicycle wheels, the 100 pound decorticator can be rolled from plot to plot wherever sisal is being grown.

Since 2011, Odundo has sold 35 machines to businessmen and cooperatives in Kenya and Tanzania. Two other machines he has developed — for turning sisal fiber into twine, and rolling twine into spools — have already sold in Rwanda, too. For those who can’t afford the machine itself, Odundo has two of his own decorticators that he rents out in a profit-sharing scheme with farmers near Kisumu. Large sisal companies near the coast pay 5-10 Kenyan shillings (6-12 cents) for a bundle of leaves that will yield a kilogram of fiber, and haul them by the truckload to their processing plants. By processing the fiber where it is grown, Odundo can afford to pay farmers 20 shillings for the same quantity of leaves, more than double the going rate. Leeman estimates the number of people whose livelihoods are improved as a result of Odundo’s decorticators in the “low hundreds,” a number that will climb with every machine sold. Odundo said that many farmers near Kisumu are already planting new sisal fields:

When you get some money out of something you thought was useless, you see the sense of planting more.

The hardest part of the equation is finding a way to build the machines locally. To do that requires what Leeman calls “middle tech,” moving away from a model of repurposing scrap metal by hand—where every part is different—to one in which you can build something twice, the exact same way. You need machines to build machines,” Leeman said. “If you look at pictures of Alex’s [first] machines, you’ll say, ‘oh, that looks like it came off a bicycle,’” oftentimes, because it did.

Over the years, Odundo’s prototypes have incorporated bicycle sprockets and frames, motorcycle chains, rims from car wheels, bevelled gears from truck transmissions, and motors for windshield wipers. “I was trying to make [parts] that didn’t exist,” he says. And so, at times, he had to re-design the decorticator around what was available in local scrap yards.

He built his prototypes in a succession of shops belonging to auto mechanics in Kisumu, and then, working off hours at the technical college where he took classes for two years. Every time a finished prototype didn’t work, Odundo says, he had to find a new person willing to take a chance on his design and offer him access to basic tools and electricity. After the fifth prototype, desperate to see some result, Odundo told one mechanic that he would give him the prototype he built in the man’s shop, if only it worked. Up through the eighth prototype, “that’s the way it went,” Odundo recalled, “until, one day, it worked. Then, I got a bunch of orders at once.”

Those early orders allowed Odundo to assemble a bare bones workshop in Kisumu, with a welding machine, an axle grinder, and tools for cutting metal. But there are five custom parts at the core of each decorticator that require a lathe and a mill, machines Odundo doesn’t have. For these, he still has to rely on the few metal shops around Kisumu that do piece work, making parts to order, one by one. It works, but it’s expensive, and it’s slow. If he were able to make these parts himself, Odundo said, the price of his sisal decorticator—now close to $1000—would fall by half.

In May, Leeman sponsored Odundo’s second trip to California, where he spent three weeks at Tech Shop, a prototyping studio in Menlo Park. “He was there 12 hours a day,” Leeman said.

The idea was to give Alex everything he might need if he were going to go out to a deserted island to build his machines.What kind of tools would you want? What skills would you need to use them?

While he was in Menlo Park, Odundo stayed in the home of Andrew Byrnes, social entrepreneur and PhD candidate in Stanford’s material science department, who told him about the concept of crowd funding. “The challenge for people like me,” Byrnes told me, “is that it’s really hard to know what people need in places where you’re not. Alex is trying to build the capabilities for Africans to build machines for Africans, and that’s a really good quality to have in a device built for Africa. Having access to cash—instead of taking 15 years to build a prototype, it could take a year or less.”

To illustrate the point, Byrnes told me about another prototype Odundo developed years ago, one of dozens in a life filled with tinkering. Odundo had shown him a video of a pulley system of weighted chains hanging on the end of a shop table in Kisumu: it was a gravity generator, Odundo explained. If you lifted the weight to a certain height, it would create electricity on the way down, enough to charge a cell phone, say, or power a light.

Later, Byrnes went online and showed Odundo the fundraising video for Gravitylight, a UK initiative that raised almost $400,000 from more than 6000 people on Indiegogo in January of this year. Odundo, predictably, was crestfallen—what might it have taken for him to bring that idea to the world when it first occurred to him?

As Byrnes put it, “How can you compete with someone who says—‘look what we did: we did this in 6 months, here’s a video, and here’s 100,000 eyes watching it—when you don’t have access to that?” But Odundo was also convinced of the potential for raising money online. His own campaign, to build a workshop that will manufacture his sisal decorticators, went live today.

He plans to call it the Victoria Innovation Centre, a place where young people—Odundos from a younger generation—might come to prototype their own designs. “From a development perspective, [that’s] the best thing that could happen,” Leeman said. “Alex will train a lot of people, and in a few years, you could have a whole machinist culture in Kisumu, where young kids come and say, ‘hey, this is a way to make money.’” But, he said, “It has to pay for itself. Because Alex has a sisal business, you’ll have a machine shop with a reason to exist.”