The Data Layer Part 3: Data Analytics

Introducing Part 3 of our deep dive, Data Analytics technologies enabling the sustainability transition

Mark Tomasovic
May 6, 2024

The Data Layer: The Role of Digital Infrastructure in Climate

Part 3: Data Analytics

New climate assets are fitted with digital sensors and systems that generate millions of datapoints every minute. With this volume of data, purpose-built tools are needed for analyzing these large datasets to visualize patterns and optimize asset performance. These machine learning platforms ingest and streamline data analysis for complex workloads so that the user can make time-sensitive, impactful decisions. For example, physics simulation, computer vision, and mineral exploration can require the processing of terabytes of data to derive insights. In this section of the Digital Infrastructure deep dive, we cover Data Analytics – machine learning and AI technologies that enable the manipulation and visualization of climate data at scale.

Defining Data Analytics Platforms

Climate assets produce large datasets that need sophisticated processing methods to derive insights. The ability to consume large datasets enables engineering required for the energy transition. For example:

- Cloud-Based Simulation: New engineering simulation products enhance the design of equipment and simulate equipment performance under rigorous operating conditions.

- Computer Vision: Computer vision technologies can process enormous image volumes associated with monitoring equipment and can automate the visual inspection of infrastructure.

- Mineral Exploration: New machine learning platforms can process terabytes of geologic data to identify untapped mineral resources that have been previously overlooked.

Today, entrepreneurs are building cloud-based machine learning platforms to analyze data and run simulations at scale. In this section, we cover the software tools needed to streamline engineering and drive sustainability forward.

Cloud-based Simulation

Cloud-based simulation software is a horizontal category that spans several verticals: from power flow analysis to mechanical equipment design. Software in this category runs machine learning models to simulate physics so that the user can understand how equipment responds under certain conditions. Let’s take power grid modeling for example. The U.S. power grid is a complex web of transmission and distribution lines, each rated for a specific capacity. As more distributed energy resources are added to the grid, regulators, utilities, and engineering design firms need a way to understand if the grid can support these new additions. Simulation software can help engineers easily understand if the grid can support new energy resources at a specific location, streamlining the design and permitting of new projects.

As today’s infrastructure continues to age, engineers need the ability to model and simulate the mechanical integrity of assets. Recent advances in compute power have unlocked the ability to run sophisticated digital twin models that accurately represent how physics will affect the performance of equipment in the field. For example, today, operators are running simulations to identify where equipment may fail and are taking proactive measures to make upgrades where needed. With next gen simulation tools, users can pinpoint maintenance across existing assets and deploy new assets with engineering certainty.  

Computer Vision

Industrial operations require an extraordinary volume of visual inspections to ensure compliance and safe operating condition. Fortunately, computer vision can automate the human eye to process enormous image volumes accurately.

Before computer vision, wind energy technicians would fly drones or use long distance cameras to capture images of wind turbine blades. The technicians would review hours of drone footage and tag images where turbine blades looked damaged. An engineer would then classify the damage based on subject matter expertise and determine if maintenance is required.

With computer vision, hours of manual image review can now be automated. Computer vision software can automatically scan through hours of footage and categorize turbine blades based on damage type. Energize estimates that by automating key pain points in the wind inspection workflow, wind operators could save 6,500 hours of effort, equating to over $400K in cost savings.

Mineral Exploration

The energy transition requires an unprecedented amount of critical minerals. A typical electric vehicle requires 6x the minerals compared to a combustion vehicle. An onshore wind plant requires 9x the minerals compared to a natural gas fired plant. Software tools can streamline the discovery of these critical materials and remove the soft costs associated with the extraction of minerals for use in the energy transition.

Today, cloud-based machine learning models can process large datasets to identify untapped resources. Companies like KoBold Metals aggregate multiple data sources and use cloud-based artificial intelligence to determine where to continue exploration. Because of declining cost of computation, cloud-based software tools now have the capability to ingest terabytes of information and run more sophisticated analysis than previously available with on-prem resources.


Once large volumes of data are routed from climate assets, cloud-based machine learning platforms can process the data to identify insights. From running physics simulations to automating visual inspection to identifying untapped minerals, new AI tools streamline engineering and maintenance of assets. In the attached report, we dive deeper into other data analytics technologies and highlight other macro trends driving entrepreneurs to build the digital infrastructure for the sustainability transition. Think there is an area we missed or a technology you’d like to explore more? Reach out!