Environmental, Social, and Governance (ESG) criteria are used to evaluate how a company impacts and is impacted by its commitment and exposure to social responsibility, environmental stewardship, and ethical business practices. Companies increasingly focus on ESG for a variety of reasons, including enjoying a competitive operational advantage and stakeholder preference for responsible sustainability practices.
The landscape of ESG reporting requirements for US companies is evolving with a mix of regulations and investor pressure driving more transparency. ESG frameworks identify technology and innovation as key tools for mitigating risks and securing opportunities. Serverless compute and data-sharing technologies are effective solutions that help organizations transition to long-term sustainability.
While ESG reporting rules are not intended to restrict a company’s carbon footprint directly, they provide decision-relevant information. Companies may be evaluated by customers, employees, investors, creditors, etc., based in part on their climate impact. Providing this information through public disclosure allows these important stakeholders to make informed decisions and incentivizes companies to seek environmental opportunities to address stakeholder pressure. Disclosure also requires companies to track critical KPIs that they can use to mitigate risks and/or seize opportunities strategically.
In March 2024, the Securities and Exchange Commission (SEC) passed requirements for all publicly traded companies to disclose in their 10-K annual report any climate-related risks and opportunities that might materially impact their financial well-being. Large public companies must also disclose certain emissions information. While disclosure was already in place in Europe as of 2024, California and the broader United States disclosure rules will be phased in over the coming years. Companies with operations in either jurisdiction will be subject to these more stringent reporting requirements.
How to Report Company Emissions
Scope 1: direct emissions (under the company's control, e.g., generated from using its own servers)
Scope 2: indirect emissions (purchased by the company, e.g., electricity purchased to keep servers cool)
Scope 3: all other emissions up and down the value chain, and usually the largest source of emissions (suppliers and customers) ( e.g., serverless compute vendors)
While California and the European Union require disclosure of all three, the SEC has excluded Scope 3 emissions from its recent Climate-Related Risk Disclosure Regulation because they are the hardest to track. For example, a Life Sciences organization’s data is Scope 1 & 2, while their Databricks/Wayfinder usage falls under Scope 3.
While it may feel as if processes in the cloud have no physical impact, they actually require significant physical resources. Data storage and analytics have emerged as sources of environmental concern, with a carbon footprint exceeding the airline industry’s (Forbes). Computing and storage require electricity and generate a tremendous amount of heat that is managed through air conditioning (electricity) and water. The environmental impact of AI and data analytics can be determined not only by these factors but also by CPU usage when the system is idle or inefficient.
Cloud computing offers many benefits for sustainability. Unlike on-premises IT infrastructure, which can be energy-hungry, cloud computing consolidates resources and uses efficient data centers that can be powered using renewable energy sources. This consolidation and efficiency can significantly reduce the overall carbon footprint of IT operations. McKinsey research indicates that cloud-powered solutions can fast-track the implementation of nearly half of the decarbonization initiatives required to achieve a 1.5°C pathway by 2050. “As companies gear up to comply with increasingly complex sustainability regulations, strategic use of cloud-powered technologies can be a crucial differentiator.”
Determining the carbon footprint of data analytics involves assessing the energy and resources consumed throughout the lifecycle of data processing, storage, and transmission.
- Efficiency of energy sources (renewable)
- Overall energy consumption (computational processing, data storage, and transmission)
- The type and efficiency of hardware used (e.g., servers, storage devices, networking equipment, and data processing units)
- Efficiency of algorithm and code execution, parallel processing, and optimization techniques that can reduce computational overhead and energy use
Serverless compute is a cloud-based, on-demand architecture without the burdens of managing machine resources and servers. Instead of renting or owning servers, users can focus on writing and deploying their code while the cloud provider takes care of automatically scaling the resources needed to run these functions and only charges based on the actual amount of computing resources used during execution. This approach allows users to be more agile, reduce costs, and focus more on building and analyzing rather than managing infrastructure.
When using serverless, companies are already achieving an important step towards sustainability – reducing their greenhouse gas emissions.
Serverless compute pools resources, whereby CPU and memory are shared amongst users and can be assigned as needed while keeping data isolated and secured from unauthorized access. This multi-tenant deployment allows overhead resources to be spread among multiple applications or functions, thereby reducing the number of physical servers needed to support the same workload and energy consumption. If your company were to be responsible for collecting, storing, maintaining, and analyzing data on your own server, you would be responsible for all resource usage and the resulting emissions that occur even when you are not actively using the server. With a multi-tenant system, your carbon usage is determined by your active usage.
Serverless platforms dynamically allocate resources based on demand. Traditional servers often run idle or underutilized, consuming power inefficiently. Serverless architecture allows for precise allocation of computing resources only when needed, thereby reducing overall energy consumption and environmental impact.
Power consumption increases with utilization; one node running at 80% CPU consumes less electricity and has greater efficiency than two computers running at 40% each.
Specifically, serverless compute allows a reduction in virtual hardware that has positive spillover effects in minimizing physical hardware, electrical usage, and emissions. Serverless technology automatically scales resources up or down based on traffic. This dynamic scaling ensures that no idle servers are running, which reduces the energy wasted on powering and cooling underutilized servers.
Due to efficiency gains, serverless compute can accomplish the same analytical functions without imposing the environmental impact of on-premise infrastructure. Moreover, by facilitating data storage, serverless warehouses allow companies to avoid unused or oversized storage to maximize the benefit of their utilization, generating both cost and sustainability gains. Many serverless providers operate large-scale data centers designed for energy efficiency. These data centers often incorporate advanced cooling systems, renewable energy sources, and energy-efficient hardware to minimize environmental impact.
Databricks and Kythera Labs’s Wayfinder’s serverless compute capabilities are examples of technology that can save business analysts valuable time, costs, and cloud waste. With serverless, companies rely on the server capabilities of Databricks and Wayfinder, which enables customers to avoid the costs and time associated with managing, maintaining, and running hardware and servers. As an on-demand service, companies avoid using and paying for unneeded cloud resources and capacity, eliminating the constant energy drain of underutilized physical servers.
Businesses can contribute to sustainability and cost goals by optimizing their cloud resource usage.
By following these practices, you can leverage the potential of serverless computing for a more sustainable and cost-effective IT environment.
Inefficient data replication can also have significant environmental impacts as it often involves transferring enormous datasets across networks or duplicating them in storage systems. This procedure can increase energy consumption due to greater processing and data transfer requirements. Additionally, storing multiple copies of data requires more physical storage space, which may lead to increased demand for data centers and associated energy consumption for cooling and maintenance.
As an alternative to copying data, companies can use technology features like Delta Sharing, a feature of Databricks and Wayfinder, which allow users to share data across data platforms, clouds or regions without copying. This reduces the need for data duplication across different environments, which can lower overall data storage and transfer energy consumption. By enabling direct access to data without replication, Delta Sharing optimizes data retrieval processes, reducing the computational resources and energy required for data processing. Simplifying data lifecycle management reduces storage costs and optimizes resource utilization, thereby enhancing operational efficiency.
Delta Sharing's scalability also allows organizations to adapt quickly to changing business needs without compromising performance or sustainability goals. This flexibility supports agile operations and resource allocation strategies.
Delta Sharing: Social and Governance Opportunities
The practice of copying data may impact ESG risks around data security and governance. Each copy of data represents an additional potential point of vulnerability. If not managed properly, copied data can increase the risk of unauthorized access and breaches, with potential subsequent social impacts if sensitive information is compromised. Furthermore, managing multiple copies of data requires additional storage, processing, and management resources, which can affect operational efficiency and increase costs associated with data governance.
Copying data introduces challenges such as increased storage costs, complexity in data management, and potential security risks if not properly controlled. Therefore, organizations must implement robust data management strategies, including data lifecycle management, encryption, access controls, and compliance measures, to maximize the benefits of data copying while mitigating associated risks.
Delta Sharing can have a positive impact on reported ESG factors by promoting environmental sustainability through reduced data movement and optimized resource utilization, enhancing social outcomes through improved collaboration and data security, and strengthening governance practices through centralized access controls and compliance efficiency. These benefits collectively contribute to a more sustainable and responsible approach to data management and analytics within organizations.
Companies should prepare to address ESG reporting standards and to meet stakeholder sustainability expectations. Companies that use healthcare data can support environmental sustainability through optimized resource utilization and reduced data movement, enhance social outcomes through improved collaboration and data security, and strengthen governance practices through centralized access controls and compliance efficiency. These benefits collectively contribute to a more sustainable and responsible approach to data management and analytics within organizations.
