Palmetto AI: Electrifying America with AI-Powered Software Developers

Determining Solar Panel and Heat Pump Needs – A Shift in Approach

Individuals considering solar panels often ask: “How many panels are required?” Similarly, those exploring heat pumps frequently inquire about appropriate sizing. Traditionally, answering these questions necessitated an on-site visit from a qualified contractor.

However, over the past ten years, solar installers have significantly streamlined this process. Some companies, notably Tesla, have even removed the need for a site visit altogether. This advancement is largely due to sophisticated software.

This software, when integrated with diverse data sources, enables installers to accurately model the number of panels a specific home will require for optimal energy production.

Palmetto’s Expansion into Whole-Home Electrification

Like many in the solar industry, Palmetto utilizes an internal tool for project assessment. This tool was developed following the acquisition of Mapdwell, a company specializing in solar mapping.

As the growth rate of the residential solar market has moderated, Palmetto has broadened its focus to encompass other electrification projects, including backup batteries and heat pumps.

Developing software solutions for these products, which are installed within the home and inaccessible to satellite imagery, presents unique challenges.

Creating a Digital Twin of the U.S. Housing Stock

To overcome these challenges, Palmetto initiated a project to compile data and create a virtual representation of every residential building in the United States. According to Michael Bratsafolis, Palmetto’s president of energy intelligence, the company is effectively “simulating digital twins of the entire U.S. residential building stock.”

The company’s system leverages both publicly available and proprietary data sources. When precise information is unavailable, its artificial intelligence algorithms utilize contextual clues to infer characteristics hidden within the building’s structure.

For instance, a home constructed in the 1950s likely features 2x4 lumber in its exterior walls, which restricts the amount of insulation that can be installed.

“This technology is capable of dissecting a home into over 60 distinct characteristics and attributes,” Bratsafolis explained.

Validating Models with Real-World Data

Palmetto validates its models by referencing data collected from homes that have purchased or are leasing solar panel systems through the company.

“This allows us to anonymize and utilize actual home data to assess model performance and further refine its accuracy,” he stated.

Opening the Platform to Third-Party Developers

Palmetto has also decided to make its tool accessible to external developers, a decision inspired by Bratsafolis’s previous experience at Twilio.

“My background includes a developer-focused approach, utilizing micro-services to provide API building blocks that empower partners and developers,” he said.

Developers receive 500 free API calls each month, with subsequent calls priced at five cents each. Customized pricing options are available for larger clients.

Driving Electrification and Market Insights

Bratsafolis anticipates that developers will create tools to accelerate the electrification of U.S. buildings. Palmetto, in turn, will gain valuable insights into emerging market trends: “The potential use cases are difficult to predict, but an API product provides invaluable insight into demand and market direction.”