The investment choices of National Grid’s corporate venture capital division suggest the company places significant importance on automation as a key component of the future utility landscape.

The substantial focus on automation and machine learning by one of the largest investor-owned utilities in the country, serving approximately 20 million customers, is noteworthy and indicates a potential direction for the broader industry.

Since its inception, National Grid Partners, National Grid’s venture capital arm, has directed funding towards 16 startups whose core offerings center around machine learning. Recently, the firm invested in AI Dash, a company utilizing machine learning algorithms to analyze satellite imagery and identify vegetation encroaching on National Grid’s power lines, thereby helping to prevent service interruptions.

A further recent investment, Aperio, employs data gathered from sensors monitoring vital infrastructure to forecast potential data loss resulting from deterioration or cybersecurity threats.

To date, approximately $135 million of the $175 million in total investments made by the firm has been allocated to businesses that incorporate machine learning into their services.

“Artificial intelligence will be essential for the energy sector to successfully meet ambitious goals related to decarbonization and decentralization,” stated Lisa Lambert, National Grid’s chief technology and innovation officer, and the founder and president of National Grid Partners.

While the beginning of the year saw a slower pace of investment due to the COVID-19 pandemic, National Grid’s investment activity has accelerated, and the company remains on course to achieve its yearly investment objectives, according to Lambert.

Modernization is crucial for an industry that largely relies on spreadsheets and the accumulated knowledge of a workforce that is aging, with limited plans to address knowledge gaps upon employee retirements, Lambert explained. This situation is driving National Grid and other utilities to increase their adoption of automation.

“The majority of companies within the utility sector are currently pursuing automation to improve efficiency and reduce costs. Currently, much of the industry’s operational knowledge is documented in manuals, and networks are primarily managed using spreadsheets, alongside the skills and experience of their operators. The potential retirement of these experienced personnel presents significant challenges. Automating and digitizing processes is a primary concern for all the utilities within the Next Grid Alliance.”

Currently, much of the automation implemented has focused on streamlining fundamental business processes. However, Lambert indicated that emerging capabilities will enable the automation of more complex activities further along the value chain.

“Machine learning represents the next advancement—enabling predictive maintenance of assets and enhancing customer service. For instance, Uniphore allows organizations to learn from every customer interaction, integrating that knowledge into its algorithms to improve future engagements. This represents the next generation of service,” Lambert said. “Once all operations are digitized, insights can be gained from every interaction, whether with an asset or a customer.”

Lambert also foresees growing demand for new machine learning technologies as utilities strive to rapidly decarbonize their operations. The transition away from fossil fuels will require fundamentally new approaches to operating and managing the power grid, with a reduced reliance on human intervention.

“Over the next five years, utilities must successfully implement automation and analytics if they hope to achieve a net-zero emissions target—requiring a different approach to asset management,” Lambert said. “Wind turbines and solar panels differ from traditional distribution networks. Many established engineers may not prioritize innovation, as their expertise lies in technologies relevant to assets built decades ago—whereas these renewable energy sources have emerged in the age of OT/IT convergence.”