Leaders from SunPower's recently closed power plant business have launched a company of their own, but they're not doing project development.

Instead, Terabase Energy will use technology to help utility-scale solar developers improve their project economics. As a software company, it can scale quickly and with minimal need for capital. But by attacking generally overlooked soft costs for the world's biggest solar plants, Terabase aims to save customers millions of dollars and accelerate global renewables adoption.

"These big power stations are giant, complicated exercises in managing material and managing people, and we think software can drive more efficiencies," said Terabase CEO and co-founder Matt Campbell.

To contextualize just how massive these operations are, Campbell described how the company is optimizing economics for a plant that involves 2 million panels, 4,000 shipping containers, 100 million pounds of steel and 500,000 hours of installation labor. Even small percentage gains in labor, shipping time, or the amount of steel required by the plant design translate into significant savings — or at least enough to justify a service fee.

Part physical, part digital

Achieving those improvements requires a mix of software engineering and physical analysis, said Campbell, who previously served as SunPower's VP of global power plants.

"Software by itself wasn’t enough to bring the efficiencies we think are required," he noted.

Thus, the team will put boots on the ground to fly drones for site inspection, take albedo measurements and otherwise assess the geotechnical characteristics of the location. It then feeds those data points into the software built with the company's collective experience developing large solar projects. It covers the full range of project evolution, from layout to materials choices to shipping logistics and the sequence of construction.

Terabase is not the first company to try squeezing cost out of solar projects. Panel and inverter manufacturers have had to do that to survive in an increasingly commoditized industry, as have construction contractors. But, Campbell argued, those improvements happen independently, leaving an opening for a holistic analysis across all the relevant activities.

In fact, it is precisely because hardware has gotten so cheap that it's become worthwhile to attack soft costs.

"Up until 18 months ago, the soft costs were a smaller percentage of the overall pie," Campbell said. "Now that the panel has gotten into the mid-20s [cents per watt], the contribution of the soft costs is much higher on a relative basis."

Soft costs are a more common topic of concern in residential solar, where they have exceeded hardware costs for several years. For home solar, the big hurdle is customer-acquisition costs, which have remained high despite years of attention from the industry. In massive solar power plants, soft costs include workflow on the site over months or years, financing costs, and even the seasonality of commodity prices and shipping fees.

"When you’re talking about 100 million pounds of steel, it might be worth it to optimize a little bit," Campbell said.

"Recovering developers"

SunPower started life as a solar manufacturer, but grew into many other parts of the value chain: utility-scale development, inverters, trackers and residential supply through a dedicated dealer network. At its height, SunPower's large-scale self-development business delivered one-third of the company's revenue.

Recently, the company decided to pare down its activities to focus on panel manufacturing and distributed generation. It will still supply panels to massive power plants, but will not compete with the biggest solar developers on scale and cost of capital.

Other developers have faced their own choices to refocus efforts in a competitive market; Cypress Creek Renewables cut its internal engineering, procurement and construction unit this summer, and First Solar announced it would restructure its EPC business after a first-quarter loss.

SunPower's exit from utility-scale left some industry veterans with time on their hands.

All six co-founders at Terabase held leadership roles at SunPower that translate fairly directly into their current dossiers. For instance, Thang Le, VP of project management and engineering, previously directed SunPower's customer-facing design engineering team. Pierre Gousseland, VP of business development and partnerships, managed SunPower's global sales and development engineering for power plants, and led construction of the 747-megawatt Solar Star plant in California.

Terabase plans to launch its basic platform for free online with the expectation that if people like what they see, they'll pay for advanced services like feasibility studies, performance engineering and more. A $1 million seed round closed in June keeps the lights on.

But the company already has revenue coming in from 10 design consulting clients, spanning several gigawatts across Australia, Vietnam, the U.S. and the Middle East. The software compares expected performance from different variables, like fixed tilt or tracker, space between rows, or bifacial versus mono. 

"Relationships are a huge asset," Campbell said of his team's collective experience in the industry. "We did projects in 20 countries, so we have this great global network and those relationships came with the team."

Though the founders are "recovering developers," they chose not to start a development shop. For one thing, they lacked hundreds of millions of dollars of cheap capital. Additionally, successful development requires local knowledge, and that's hard for a small startup to project around the world. 

"It’s a business about land and politics and infrastructure," Campbell said. "And we happen to all be in California, which is the most crowded development market of all."

Building a technology startup in Berkeley struck them as a more promising way to turn their development experience into a new business. And it means all those developers and EPCs working overseas can become customers instead of competitors.