Carbon Lighthouse CEO Brenden Millstein wants to stop climate change by tapping the cheapest resource there is: wasted energy. But in a crowded energy efficiency and services market, his company has found that data complexity — and business model simplicity — is an essential part of that mission.  

Since its founding in 2009, Carbon Lighthouse has steadily built up a roster of more than 500 buildings using its energy efficiency services, most achieving 20 to 30 percent utility bill reductions. This isn’t an exceptional figure in the world of big-building energy efficiency. Significant savings can come from simple retrofits like LED lighting replacements, HVAC system upgrades, or finding the simple yet expensive mistakes common in many buildings, like leaving things turned on overnight. 

But Carbon Lighthouse’s efficiency projects differ from many of the energy services contracts, shared savings agreements or no-money-down financing programs we’ve seen emerge from industry giants and startups alike over the past decade, Millstein said in an interview this summer. 

First, the company is often achieving its savings on top of the “low-hanging fruit” of lighting and HVAC upgrades, by using sensors and artificial intelligence to keep buildings fine-tuned for years after their upgrades. “We’re able to do it without replacing the big capital-intensive equipment — no cranes, no disruption,” he said. That's a big difference from many of the energy services giants like Siemens, Schneider Electric and Honeywell that are motivated by selling gear. 

Second, Carbon Lighthouse has simplified the contracts it writes with customers to a simple proposition, usually taking no more than 10 pages, he said. “We charge a fixed monthly fee, and in exchange, we guarantee we’ll deliver a dollar amount of savings that’s above a certain threshold,” he said. That could work out to something like $100,000 upfront for a guarantee of at least $250,000 per year in savings. 

A different kind of savings guarantee

To be sure, many energy services companies and startups alike offer different variations on the savings guarantee. But Millstein said those usually differ, in two important ways. 

First, Carbon Lighthouse doesn’t claim a piece of whatever upside emerges beyond its guarantee, a feature that makes sense economically, but can make contracts more complex, and thus more potentially risky for customers. 

Second, if a customer saves less than the guaranteed amount, Carbon Lighthouse writes them a check to cover the difference, Millstein said, thus taking on all the downside risk. So far, however, Carbon Lighthouse has only had to meet that obligation for about 3 percent of its customers, according to company data. And those instances were quickly corrected, as part of the continuous commissioning that helps avoid the inevitable “drift” away from efficiency gains in the years after they’re installed.

This level of accuracy — and the willingness to put its own money behind it — has been a part of Carbon Lighthouse’s contracts since about 2012. But the artificial intelligence platform, dubbed Carbon Lighthouse Unified Engineering System (CLUES), that provides the data behind those estimates was not planned, Millstein said. "It pretty much developed out of panic."

“Our very first customers, when they signed the contract, had already done the low-hanging fruit, the medium-hanging fruit — the orchard had been harvested,” he said. He and his co-founder, Carbon Lighthouse President Raphael Rosen, decided to “get $2,000 worth of sensors and install them, and pray that somebody missed something in the data.” 

The data discovered that a large fan was running improperly, a fix that saved the customer about 3 percent per year on energy bills. “It was enough,” he said. The startup was able to land more clients — a private school in Oakland, California and an office owner in Sacramento — and buy the sensors to find the missing efficiency potential.

“Sure enough, there’s a pump running when it’s not supposed to run, and we get 4 percent energy savings," said Millstein. "A tiny number, but it was enough.” 

Hunting for efficiency improvements

This software-driven approach starts with a preassessment, with two engineers deploying “a couple hundred sensors” in a building's mechanical rooms, on the roof, and a small sampling of occupied areas over a day or two. Once those sensors are in, all the data streams to the cloud go into the company’s CLUES platform, which can ingest and analyze tens of millions of data points per building to create custom simulations of individual buildings, rather than use more generic building-type models more common in industry-standard software. 

CLUES looks for two things, Millstein said. First, it looks for one-time changes to make. Things like lighting retrofits, changing valves and other easy efficiency fixes that may be unrealized in its customers’ properties. Of course, these aren’t hard to find, even using spreadsheets and manual entry, and “a lot of our clients have already done a lot of that,” he noted. 

“The second thing we look for is all of the ongoing optimization,” he said. And this is where Carbon Lighthouse gets, on average, about 80 percent of its savings over time. This concept of continuous commissioning is also a common one in the building energy efficiency field, but it’s not always a priority for vendors who are primarily interested in selling new HVAC or lighting equipment, Millstein noted. 

Even for the efficiency services companies that tie their revenues to ongoing performance, keeping buildings tuned to optimal energy efficiency isn’t easy. Traditional approaches, built on alarms and alerts for building equipment that’s operating out of spec, can often go unnoticed by busy facility managers, particularly if they’re not tied in real time to an understanding of what the energy costs of improperly functioning equipment may be.

But the CLUES platform, by combining its sensor-driven building systems expertise with its understanding of energy costs for different systems, can take an active role in managing them as a whole to save energy, said Millstein. “Think of it as getting your car tuned up, but automatically and every five minutes,” through data delivered from the CLUES platform to the individual building systems and operators assigned to manage them. 

A successful deal with Tesla

Millstein acknowledged that Carbon Lighthouse’s use of sensors and cloud computing are becoming standard fare in the energy efficiency industry, as evidenced by companies such as FirstFuel,  Ecova's Retroficiency, BuildingIQ, and NRG’s Station A spinout. It's also apparently in the big-data platforms that underlie the building energy services offerings from industry giants like Schneider’s EcoStruxure platform. 

Carbon Lighthouse doesn’t offer its software as a subscription or a service, keeping its use in-house. The company also works with a network of contractors to deploy the equipment it monitors on behalf of its clients. Since 2010, it has begun to include rooftop solar as a potential factor in its building energy analysis.

This spring, Carbon Lighthouse landed a $27 million round of investment led by GRC SinoGreen, with participation from JCI Ventures, Ulupono Initiative, SV Tech Ventures, and Tesla co-founder and CTO JB Straubel. Tesla is one of Carbon Lighthouse’s premiere customers, having engaged it in a six-year contract that’s helped reduce utility bills by $90,800 per year — or by 106 tons of carbon dioxide annually — across its Palo Alto, Calif. manufacturing, office and R&D facilities. 

That project represented a real challenge in going beyond the efficiencies Tesla had already gotten from its in-house building engineering team and two different leading efficiency firms, he said. “When our engineers got to the site, they were pretty scared they wouldn’t find anything,” he said. 

But the sensors they deployed helped provide tons of information previously lacking to the company — electrical flows, chiller fluid flows, occupancy rates, and the like. That information helped find additional efficiencies, Millstein said. When Tesla replaced about 10,000 square feet of machine shop with dense office space about eight months after the project started, “we were able to reoptimize the building, and ended up increasing savings by 7 percent per year,” he said.