We sat down with Shingai Samudzi, CEO, Asoba, as part of the Xooglers in Climate Founder Series to explore his journey—from growing up between Zimbabwe and the US to building Asoba, an AI-driven virtual power plant platform tailored for African energy systems.

From decision science to data infrastructure

Born in the U.S. to Zimbabwean parents, his childhood was shaped by a diasporic movement and a scientific mindset. His father, a physicist who worked on the Human Genome Project, and his mother, an NIH epidemiologist, set the tone: systems thinking wasn’t a framework. It was a lifestyle.

Shingai holds degrees in Decision Science and International Relations from Carnegie Mellon University. He began his career at Looker (later acquired by Google), diving into the guts of AI data infrastructure. “Once we got into Google, this was 2019. It was before AI was in vogue,” he said. “People were still trying to figure out the basics—how to get pipelines and automation in place.”

His work evolved quickly. He helped Walmart improve supply chain forecasting, then joined Lumen Energy as their first business strategy hire. But while working on solar adoption in U.S. commercial real estate, Shingai saw the limitations of over-regulated markets. Tax incentives and utility monopolies distorted the real economics of solar. In South Africa, where the grid was underbuilt, he saw a different opportunity:

“You can actually design for flexibility from the start.”

Building Asoba: where AI meets infrastructure

At its core, Asoba is an AI-driven platform that orchestrates distributed energy resources (DERs)—solar panels, wind turbines, batteries—so they function like one intelligent, flexible power plant.

But that’s easier said than done.

“In Africa, up to 50% of the inverter data can be missing on a given day,” Shingai explained. “So before you even start forecasting, you need to build systems that can fill in the blanks—and do it accurately.”

Asoba’s platform begins with data interpolation, using digital twins and transfer learning to reconstruct missing inputs. Then it performs demand and production forecasting at the most granular level—down to individual inverters—before running optimization models based on the customer’s goals.

“Whether it’s CO₂ reduction, maximizing revenue, or grid balance, we tune our dispatch optimizer to that objective. Then we send recommendations hour-by-hour to a smart controller that executes them in real-time.”

The whole loop—from prediction to control—is fully automated.

Centralized vs decentralized: what a virtual power plant really does

To understand what makes Asoba’s platform transformative, it helps to contrast it with the traditional way power is managed.

“Traditionally, you had very centralized grids where you have maybe one, or a handful of power plants that are centrally controlled and managed,” Shingai explained. “You literally have a group of people in an operations room controlling electricity distribution for the entire country—or a municipal district—from a single control room.”

In this setup, power flows one way: from the grid to the consumer. And so does the money. Utilities generate the electricity, sell it to homes and businesses, and collect payments. It’s a top-down system—stable in some contexts, but rigid and vulnerable in others, especially where infrastructure is aging or thin.

Now contrast that with a virtual power plant, or VPP.

“In a virtual power plant type of scenario,” Shingai said, “the generation of electricity is now distributed.”

That means the power doesn’t just come from a few massive plants. It can come from solar panels on rooftops, wind turbines in remote fields, hydro stations, or even small-scale green hydrogen units. These are known as distributed energy resources (DERs), and they’re not centrally owned or controlled—they’re scattered, diverse, and often privately operated.

What ties them together is software. A VPP is not a physical plant—it’s a digital control system that orchestrates all these independent power sources to act like a coordinated whole.

“It’s a piece of software that connects to the devices via their sensors… so you might connect with inverters from a solar array or wind turbine. You need to be able to read the data in real-time.”

The VPP gathers this data, forecasts supply and demand in short intervals, and tells the system how to dispatch electricity and at what price. The energy—and the money—now flows in both directions.

“You might have someone with rooftop solar,” Shingai noted. “At times during the day, they’re buying electricity from the grid. At other times, they’re selling their spare electricity back. So the flow of electricity is now bidirectional. And so is the money.”

In this decentralized model, the grid becomes more flexible, more responsive.

*How distributed energy resources (DER) have sparked a need for the virtual power plant. Credit: Asoba*

Local energy, local language

In practice, most of Asoba’s customers don’t think in terms of VPPs or AI platforms. They think in terms of power purchase agreements (PPAs), grid access, and regulatory barriers.

“I know who the customers are who are gonna need this. But they don’t know that they need a VPP. And frankly, it doesn’t matter.”

This insight shaped Asoba’s go-to-market strategy. They stopped selling tech features and started selling outcomes—often in the form of demos powered by the customer’s own historical meter data.

“For our demo, we ask for a year’s worth of interval data. Then within 2–3 days, we come back with a forecast, and they can see the difference.”

The result? Better accuracy than their in-house tools—and a faster path to trust.

Navigating a fragmented grid

Africa’s energy landscape isn’t just decentralized—it’s politically fragmented. In South Africa, for instance, electricity rights are split between the national utility (Eskom) and dozens of municipalities.

“You want to sell electricity from Cape Town to George?” Shingai said. “You need three separate rights-of-way. Each one charges a toll. It’s like a network of little fiefdoms.”

That reality makes intelligent dispatch even more critical. Asoba’s system forecasts not only energy supply and demand, but also the economic viability of dispatching power across political boundaries. The goal: maximize returns while minimizing regulatory friction.

The policy bot

Beyond forecasting and dispatch, Asoba is also developing a policy engine—an AI-powered assistant designed to help navigate the complex and fast-changing world of electricity regulation, both in Africa and the United States.

“There’s a whole new regulatory regime coming around electricity trading,” Shingai explained. “Developers need help understanding it. That’s where the policy bot comes in.”

Originally built as an internal tool to help the Asoba team stay up to speed across different markets, the policy bot is now being launched. “We found so much demand for it—even from people who aren’t yet ready to use our platform but are deep in the climate space,” Shingai said.

Why the need? Because whether you're in South Africa or the U.S., keeping up with energy regulation is not just a legal challenge—it’s a strategic one.

“You can't build the best possible business model if you don't understand the policy like the back of your own hand.”

In the U.S., where each of the 50 states acts like its own regulatory body, the complexity is especially acute. “It’s physically impossible to carry all that knowledge in-house unless you’re a startup funded with billions of dollars,” Shingai noted.

The bot is trained on a wide range of regulatory data, and it’s designed to answer questions like:

How can I participate in the carbon credit market in California?
What are the interconnection requirements for commercial solar in New Jersey?
Which states are best positioned for community solar development in 2025?

But its ambition doesn’t stop there. Asoba is now evolving the tool toward version 1.5—where it doesn’t just answer static regulatory questions, but starts to anticipate change.

“We want to get to a point where the bot can help predict what might happen next—like how the California legislature could act on NEM 3.0 based on who’s coming into power.”

That means incorporating not just policy language, but also political dynamics, affiliations, funding coalitions, and ideology.

“It started out as just an energy policy chatbot. But then it turned into understanding the entire financial and political ecosystem that’s connected to energy,” Shingai said. “It’s about the players, the coalitions behind them, the philosophies, and how they inform the policies.”

Even with this sophistication, one of Shingai’s favorite features is the most honest one:

“For me, the most important part of the chatbot is that it can say: I don’t know.”

The human infrastructure behind the system

If Asoba is building an operating system for modern energy, it’s powered not just by AI—but by a deeply local, technically skilled, and mission-aligned team.

“The team is composed primarily of people based in South Africa,” Shingai shared. “We’ve got team members with PhDs in forecasting, operations leaders who transitioned from oil and gas to renewables, and we’re continuing to hire engineering talent locally.”

That local focus matters. It ensures the technology isn’t just exported—it’s embedded. The team’s expertise spans forecasting, infrastructure, and policy.

For founders looking to work in complex, global markets, Shingai’s journey offers a blueprint: design for the local context, build, learn, and surround yourself with a team that understands the terrain better than any imported playbook.

AI-powered virtual power plants for a smarter grid

Material innovation that transforms the fashion industry