JPMorgan Chase Could Keep a Closer Eye on Scope Three Emissions

JPMorgan Chase may be tracking its ESG investments even closer. 

The financial firm filed a patent application for a system for “carbon impact tracking” of applications, software suites or consumer products. JPMorgan’s system generates carbon impact summaries that offer a number of metrics that help clients understand whether or not the product is environmentally healthy, as well as strategies to fix it if it’s not.  

“Accurate reporting of energy and carbon usage is essential for understanding potential climate impacts of platforms and software running on those platforms,” the firm said in its filing. 

JPMorgan’s system takes in a variety of data related to a program, including the amount of energy consumed within a predetermined time, what kind of energy the system runs on, and the location at which the energy is consumed. From this, the system calculates different metrics of impact, including kilograms of carbon dioxide emissions, kilowatt-hours of energy, or a dollar equivalent of a measurement of “carbon social cost.”

The system also recommends mitigation actions, such as ways to increase energy efficiency, reduce or offset energy consumption, or switch the sources it gets its energy from. 

For example, if an application relies on training a machine learning mode that sucks up a lot of energy, the system may recommend ways to limit the size of a dataset to reduce energy waste. Or if a cloud platform is burning through lots of energy, it may recommend more efficient container and virtual machine placements for a more “sustainable cloud,” the firm said. 

Software is used all day, every day, by millions of people. It doesn’t have as intense of a carbon footprint as industries such as aviation or automotive, but the sheer amount of time we spend with software applications can create a significant carbon footprint, said Lauren Densham, head of ESG & Impact at Energize Capital. The rapid onset of AI has only worsened software’s energy problem, as training AI models can suck up a significant amount of energy. 

Within companies, software-related emissions are often part of what’s called the Scope 3 category, or those that are the result of activity that isn’t controlled by the organization reporting their carbon emissions. Estimating these can often be difficult, said Densham, as it’s hard to track down what energy use comes from what applications with so many users. 

Companies often use spending data to estimate emissions in this way by “multiplying a dollar amount by an emissions factor.” she said. “But that’s not the data that you actually need. You need the actual activity data of a specific location, specific process, specific data center, to actually use that information to decarbonize.” JPMorgan’s patent may offer a solution, she said. 
While it’s unclear how exactly JPMorgan could implement this, ESGs are a focus for the financial institution. The bank has more than 300 analysts focused on integrating ESG factors into their research and a team dedicated to sustainable investing. Patenting this kind of tech could help the firm offer its clients a more granular look at the companies they want to invest in, as well as the capability to keep a closer eye on its own emissions.