Introduction: Sustainable IT Monitoring & Grid Signals

Introduction

The rising electricity consumption of the IT sector and the associated emissions mean sustainable IT is no longer a “nice-to-have” or a side project of isolated developer teams. It becomes an operational necessity for all companies managing a large IT infrastructure and a growing customer demand for IT or Cloud service providers.

In 2024, data centers accounted for 1.5% of global electricity consumption (crypto excluded). Data transmission networks accounted for roughly the same part. Combined, this represents one-fifth of the US electricity generation in 2024. It shows that the rise of electricity consumption today is not only true for data centers but extends to all IT infrastructure and services.

This electricity demand is significantly increasing. In the US, the expansion of data centres has been one of the main drivers of electricity demand growth over the last year and is projected to account for half of electricity demand growth between now and 2030. Data centers could account for one-third of the electricity demand in Ireland next year. As the electricity demand of data centers is set to double by 2030, it will exceed heavy industries' electricity consumption in some parts of the world.

This surge in power demand also raises multiple challenges, especially for building and integrating more data centers to the grid. This calls for more management of the IT infrastructure as a whole.

The rise of electricity consumption and the associated emissions are not limited to data centers but extend to all IT infrastructure and services. This urges companies to more closely monitor IT emissions at a time when regulations are strengthening. It also calls for more transparency from cloud and IT services providers, who must develop energy and emissions monitoring features or run the risk of losing customers and prospects to competition.

FinOps and GreenOps are known for being complementary, making operations more efficient and sustainable, reducing at the same time operating costs and emissions. By more closely monitoring your IT electricity consumption and emissions, your company and your clients leverage both financial and sustainability benefits.

In Today’s lesson

In the rest of our introductory lesson today, we will cover the following:

Some basics of electricity grids’ dispatch, and what this means for electricity consumers
What electricity grids' signals are and why they are crucial in Sustainable IT Monitoring

[WIP] Vocabulary

Electricity grid:

Signal:

Electricity mix:

Electricity flows:

Carbon-free and renewable energy percentages:

Electricity load:

Net load:

Carbon intensity:

Electricity prices:

Electricity grids

Electricity grids are complex systems that connect electricity consumers and producers. Grid operators are responsible for ensuring demand equals generation at every second of the year, leveraging available production sources.

Because electricity demand keeps fluctuating, and available production sources as well (think about wind and solar generation, but these are not the only ones), a multitude of parameters are changing every second on the electricity grid.

As an image is worth a thousand words, here are two gifs that help better understand how things change over time and space on an electricity grid (here, the carbon intensity of electricity - how many grams of CO2 are emitted for every kilowatt-hour of electricity consumed).

The first gif illustrates how carbon intensity fluctuates in time, between quarters, months, days, and hours on the Dutch grid.

The second gif illustrates how carbon intensity fluctuates in space for the same hour, depending on geographic granularity - here, the view switches from the grid level to the national level in the US and Canadian grids.

Introduction

The rising electricity consumption of the IT sector and the associated emissions mean sustainable IT is no longer a “nice-to-have” or a side project of isolated developer teams. It becomes an operational necessity for all companies managing a large IT infrastructure and a growing customer demand for IT or Cloud service providers.

In 2024, data centers accounted for 1.5% of global electricity consumption (crypto excluded). Data transmission networks accounted for roughly the same part. Combined, this represents one-fifth of the US electricity generation in 2024. It shows that the rise of electricity consumption today is not only true for data centers but extends to all IT infrastructure and services.

This electricity demand is significantly increasing. In the US, the expansion of data centres has been one of the main drivers of electricity demand growth over the last year and is projected to account for half of electricity demand growth between now and 2030. Data centers could account for one-third of the electricity demand in Ireland next year. As the electricity demand of data centers is set to double by 2030, it will exceed heavy industries' electricity consumption in some parts of the world.

This surge in power demand also raises multiple challenges, especially for building and integrating more data centers to the grid. This calls for more management of the IT infrastructure as a whole.

The rise of electricity consumption and the associated emissions are not limited to data centers but extend to all IT infrastructure and services. This urges companies to more closely monitor IT emissions at a time when regulations are strengthening. It also calls for more transparency from cloud and IT services providers, who must develop energy and emissions monitoring features or run the risk of losing customers and prospects to competition.

FinOps and GreenOps are known for being complementary, making operations more efficient and sustainable, reducing at the same time operating costs and emissions. By more closely monitoring your IT electricity consumption and emissions, your company and your clients leverage both financial and sustainability benefits.

In Today’s lesson

In the rest of our introductory lesson today, we will cover the following:

Some basics of electricity grids’ dispatch, and what this means for electricity consumers
What electricity grids' signals are and why they are crucial in Sustainable IT Monitoring

[WIP] Vocabulary

Electricity grid:

Signal:

Electricity mix:

Electricity flows:

Carbon-free and renewable energy percentages:

Electricity load:

Net load:

Carbon intensity:

Electricity prices:

Electricity grids

Electricity grids are complex systems that connect electricity consumers and producers. Grid operators are responsible for ensuring demand equals generation at every second of the year, leveraging available production sources.

Because electricity demand keeps fluctuating, and available production sources as well (think about wind and solar generation, but these are not the only ones), a multitude of parameters are changing every second on the electricity grid.

As an image is worth a thousand words, here are two gifs that help better understand how things change over time and space on an electricity grid (here, the carbon intensity of electricity - how many grams of CO2 are emitted for every kilowatt-hour of electricity consumed).

The first gif illustrates how carbon intensity fluctuates in time, between quarters, months, days, and hours on the Dutch grid.

The second gif illustrates how carbon intensity fluctuates in space for the same hour, depending on geographic granularity - here, the view switches from the grid level to the national level in the US and Canadian grids.

Introduction

The rising electricity consumption of the IT sector and the associated emissions mean sustainable IT is no longer a “nice-to-have” or a side project of isolated developer teams. It becomes an operational necessity for all companies managing a large IT infrastructure and a growing customer demand for IT or Cloud service providers.

In 2024, data centers accounted for 1.5% of global electricity consumption (crypto excluded). Data transmission networks accounted for roughly the same part. Combined, this represents one-fifth of the US electricity generation in 2024. It shows that the rise of electricity consumption today is not only true for data centers but extends to all IT infrastructure and services.

This electricity demand is significantly increasing. In the US, the expansion of data centres has been one of the main drivers of electricity demand growth over the last year and is projected to account for half of electricity demand growth between now and 2030. Data centers could account for one-third of the electricity demand in Ireland next year. As the electricity demand of data centers is set to double by 2030, it will exceed heavy industries' electricity consumption in some parts of the world.

This surge in power demand also raises multiple challenges, especially for building and integrating more data centers to the grid. This calls for more management of the IT infrastructure as a whole.

The rise of electricity consumption and the associated emissions are not limited to data centers but extend to all IT infrastructure and services. This urges companies to more closely monitor IT emissions at a time when regulations are strengthening. It also calls for more transparency from cloud and IT services providers, who must develop energy and emissions monitoring features or run the risk of losing customers and prospects to competition.

FinOps and GreenOps are known for being complementary, making operations more efficient and sustainable, reducing at the same time operating costs and emissions. By more closely monitoring your IT electricity consumption and emissions, your company and your clients leverage both financial and sustainability benefits.

In Today’s lesson

In the rest of our introductory lesson today, we will cover the following:

Some basics of electricity grids’ dispatch, and what this means for electricity consumers
What electricity grids' signals are and why they are crucial in Sustainable IT Monitoring

[WIP] Vocabulary

Electricity grid:

Signal:

Electricity mix:

Electricity flows:

Carbon-free and renewable energy percentages:

Electricity load:

Net load:

Carbon intensity:

Electricity prices:

Electricity grids

Electricity grids are complex systems that connect electricity consumers and producers. Grid operators are responsible for ensuring demand equals generation at every second of the year, leveraging available production sources.

Because electricity demand keeps fluctuating, and available production sources as well (think about wind and solar generation, but these are not the only ones), a multitude of parameters are changing every second on the electricity grid.

As an image is worth a thousand words, here are two gifs that help better understand how things change over time and space on an electricity grid (here, the carbon intensity of electricity - how many grams of CO2 are emitted for every kilowatt-hour of electricity consumed).

The first gif illustrates how carbon intensity fluctuates in time, between quarters, months, days, and hours on the Dutch grid.

The second gif illustrates how carbon intensity fluctuates in space for the same hour, depending on geographic granularity - here, the view switches from the grid level to the national level in the US and Canadian grids.

Introduction

The rising electricity consumption of the IT sector and the associated emissions mean sustainable IT is no longer a “nice-to-have” or a side project of isolated developer teams. It becomes an operational necessity for all companies managing a large IT infrastructure and a growing customer demand for IT or Cloud service providers.

In 2024, data centers accounted for 1.5% of global electricity consumption (crypto excluded). Data transmission networks accounted for roughly the same part. Combined, this represents one-fifth of the US electricity generation in 2024. It shows that the rise of electricity consumption today is not only true for data centers but extends to all IT infrastructure and services.

This electricity demand is significantly increasing. In the US, the expansion of data centres has been one of the main drivers of electricity demand growth over the last year and is projected to account for half of electricity demand growth between now and 2030. Data centers could account for one-third of the electricity demand in Ireland next year. As the electricity demand of data centers is set to double by 2030, it will exceed heavy industries' electricity consumption in some parts of the world.

This surge in power demand also raises multiple challenges, especially for building and integrating more data centers to the grid. This calls for more management of the IT infrastructure as a whole.

The rise of electricity consumption and the associated emissions are not limited to data centers but extend to all IT infrastructure and services. This urges companies to more closely monitor IT emissions at a time when regulations are strengthening. It also calls for more transparency from cloud and IT services providers, who must develop energy and emissions monitoring features or run the risk of losing customers and prospects to competition.

FinOps and GreenOps are known for being complementary, making operations more efficient and sustainable, reducing at the same time operating costs and emissions. By more closely monitoring your IT electricity consumption and emissions, your company and your clients leverage both financial and sustainability benefits.

In Today’s lesson

In the rest of our introductory lesson today, we will cover the following:

Some basics of electricity grids’ dispatch, and what this means for electricity consumers
What electricity grids' signals are and why they are crucial in Sustainable IT Monitoring

[WIP] Vocabulary

Electricity grid:

Signal:

Electricity mix:

Electricity flows:

Carbon-free and renewable energy percentages:

Electricity load:

Net load:

Carbon intensity:

Electricity prices:

Electricity grids

Electricity grids are complex systems that connect electricity consumers and producers. Grid operators are responsible for ensuring demand equals generation at every second of the year, leveraging available production sources.

Because electricity demand keeps fluctuating, and available production sources as well (think about wind and solar generation, but these are not the only ones), a multitude of parameters are changing every second on the electricity grid.

As an image is worth a thousand words, here are two gifs that help better understand how things change over time and space on an electricity grid (here, the carbon intensity of electricity - how many grams of CO2 are emitted for every kilowatt-hour of electricity consumed).

The first gif illustrates how carbon intensity fluctuates in time, between quarters, months, days, and hours on the Dutch grid.

The second gif illustrates how carbon intensity fluctuates in space for the same hour, depending on geographic granularity - here, the view switches from the grid level to the national level in the US and Canadian grids.

Carbon intensity fluctuates in time, between quarters, months, days, and hours on the Dutch grid.

Carbon intensity fluctuates in space for the same hour, depending on geographic granularity

Grid signals

There are a multitude of signals that can be followed about the grid. The fundamental ones are:

Load

Electricity mix per technology in MW
Electricity prices in €/MWh

But signals derived from the above can provide additional insights, such as:

Percentage of renewables or low-carbon sources on the grid
Carbon intensity of electricity in gCO2/kWh
Net load

The generation mix is very different from one grid to another (nuclear or not, penetration of renewables, phased out coal or not, …) as well as load profiles (electric heating/cooling or not, penetration of electric mobility and/or battery storage, behind-the-meter generation, …). This means all of the above signals have great fluctuations in space.

The same applies to time fluctuations with renewable energy generation, times of peak consumption, seasonality in heating and cooling demand, lower demand over the weekend and at night, etc. This means these signals fluctuate in space and time.

As a first example, let’s see how the electricity load fluctuates within a day. You can see these fluctuations in figures 1 and 2, and how these fluctuations differ between two grids (Germany and PJM in the US), but also between a weekday and a weekend on the same grid (a Sunday and a Monday on the German grid).

Grid signals

There are a multitude of signals that can be followed about the grid. The fundamental ones are:

Load

Electricity mix per technology in MW
Electricity prices in €/MWh

But signals derived from the above can provide additional insights, such as:

Percentage of renewables or low-carbon sources on the grid
Carbon intensity of electricity in gCO2/kWh
Net load

The generation mix is very different from one grid to another (nuclear or not, penetration of renewables, phased out coal or not, …) as well as load profiles (electric heating/cooling or not, penetration of electric mobility and/or battery storage, behind-the-meter generation, …). This means all of the above signals have great fluctuations in space.

The same applies to time fluctuations with renewable energy generation, times of peak consumption, seasonality in heating and cooling demand, lower demand over the weekend and at night, etc. This means these signals fluctuate in space and time.

As a first example, let’s see how the electricity load fluctuates within a day. You can see these fluctuations in figures 1 and 2, and how these fluctuations differ between two grids (Germany and PJM in the US), but also between a weekday and a weekend on the same grid (a Sunday and a Monday on the German grid).

Grid signals

There are a multitude of signals that can be followed about the grid. The fundamental ones are:

Load

Electricity mix per technology in MW
Electricity prices in €/MWh

But signals derived from the above can provide additional insights, such as:

Percentage of renewables or low-carbon sources on the grid
Carbon intensity of electricity in gCO2/kWh
Net load

The generation mix is very different from one grid to another (nuclear or not, penetration of renewables, phased out coal or not, …) as well as load profiles (electric heating/cooling or not, penetration of electric mobility and/or battery storage, behind-the-meter generation, …). This means all of the above signals have great fluctuations in space.

The same applies to time fluctuations with renewable energy generation, times of peak consumption, seasonality in heating and cooling demand, lower demand over the weekend and at night, etc. This means these signals fluctuate in space and time.

As a first example, let’s see how the electricity load fluctuates within a day. You can see these fluctuations in figures 1 and 2, and how these fluctuations differ between two grids (Germany and PJM in the US), but also between a weekday and a weekend on the same grid (a Sunday and a Monday on the German grid).

Grid signals

There are a multitude of signals that can be followed about the grid. The fundamental ones are:

Load

Electricity mix per technology in MW
Electricity prices in €/MWh

But signals derived from the above can provide additional insights, such as:

Percentage of renewables or low-carbon sources on the grid
Carbon intensity of electricity in gCO2/kWh
Net load

The generation mix is very different from one grid to another (nuclear or not, penetration of renewables, phased out coal or not, …) as well as load profiles (electric heating/cooling or not, penetration of electric mobility and/or battery storage, behind-the-meter generation, …). This means all of the above signals have great fluctuations in space.

The same applies to time fluctuations with renewable energy generation, times of peak consumption, seasonality in heating and cooling demand, lower demand over the weekend and at night, etc. This means these signals fluctuate in space and time.

As a first example, let’s see how the electricity load fluctuates within a day. You can see these fluctuations in figures 1 and 2, and how these fluctuations differ between two grids (Germany and PJM in the US), but also between a weekday and a weekend on the same grid (a Sunday and a Monday on the German grid).

Same day load fluctuations in Germany, compared to PJM (US)

German grid load on two consecutive days

These fluctuations are also true on the electricity mix side. Again, between Germany and PJM, the resources available to the grid operator to meet electricity demand are very different in nature and in proportion.

Electricity Mix in Germany on the 4th of August, 2025

Electricity Mix in PJM (US) on the 4th of August, 2025

You can explore how electricity demand and electricity mix fluctuate in space and time, and how these affect other characteristics of the grid such as the carbon intensity, or the share of renewables on our two freely available tool below:

Live map
Data explorer

This already highlights one key consideration when choosing the right data to ingest into your products and solutions: granularity matters! The granularity you need for your solution also depends on other factors, such as: how granular is your electricity consumption data? How frequently and granularly can your monitoring be?

Insights increase with granularity. Moving from a yearly to hourly granularity improves accuracy by 20% on average worldwide, and by more than 40% in several grids. Value is already achieved with small improvements in granularity; switching from a yearly to a monthly view already improves accuracy by more than 10% worldwide.

You will learn more about how to source these grid signals and the key considerations to keep in mind for a successful data ingestion in the second lesson of this course.

The types of insights unlocked are not only a function of the granularity, but also of the time perspective considered:

Historical data powers trend analysis between years, accurate carbon accounting of IT emissions, dashboards of monthly activity with electricity consumption, and associated emissions.
Real-time data unlocks live insights on electricity carbon intensity and price, raises awareness among teams, and paves the way for optimization in time and space

Forecasted data represents the cutting edge. Predictions over multiple days empower advanced and automated decision-making to optimize usage, increase efficiency, reduce costs, and emissions.

Going further

Let’s illustrate the importance of granularity when sourcing grid signals. We take as an example a data center located in Germany, for which we approximate the load profile by the curve below:

You can explore how electricity demand and electricity mix fluctuate in space and time, and how these affect other characteristics of the grid such as the carbon intensity, or the share of renewables on our two freely available tool below:

Live map
Data explorer

This already highlights one key consideration when choosing the right data to ingest into your products and solutions: granularity matters! The granularity you need for your solution also depends on other factors, such as: how granular is your electricity consumption data? How frequently and granularly can your monitoring be?

Insights increase with granularity. Moving from a yearly to hourly granularity improves accuracy by 20% on average worldwide, and by more than 40% in several grids. Value is already achieved with small improvements in granularity; switching from a yearly to a monthly view already improves accuracy by more than 10% worldwide.

You will learn more about how to source these grid signals and the key considerations to keep in mind for a successful data ingestion in the second lesson of this course.

The types of insights unlocked are not only a function of the granularity, but also of the time perspective considered:

Historical data powers trend analysis between years, accurate carbon accounting of IT emissions, dashboards of monthly activity with electricity consumption, and associated emissions.
Real-time data unlocks live insights on electricity carbon intensity and price, raises awareness among teams, and paves the way for optimization in time and space

Forecasted data represents the cutting edge. Predictions over multiple days empower advanced and automated decision-making to optimize usage, increase efficiency, reduce costs, and emissions.

Going further

Let’s illustrate the importance of granularity when sourcing grid signals. We take as an example a data center located in Germany, for which we approximate the load profile by the curve below:

You can explore how electricity demand and electricity mix fluctuate in space and time, and how these affect other characteristics of the grid such as the carbon intensity, or the share of renewables on our two freely available tool below:

Live map
Data explorer

This already highlights one key consideration when choosing the right data to ingest into your products and solutions: granularity matters! The granularity you need for your solution also depends on other factors, such as: how granular is your electricity consumption data? How frequently and granularly can your monitoring be?

Insights increase with granularity. Moving from a yearly to hourly granularity improves accuracy by 20% on average worldwide, and by more than 40% in several grids. Value is already achieved with small improvements in granularity; switching from a yearly to a monthly view already improves accuracy by more than 10% worldwide.

You will learn more about how to source these grid signals and the key considerations to keep in mind for a successful data ingestion in the second lesson of this course.

The types of insights unlocked are not only a function of the granularity, but also of the time perspective considered:

Historical data powers trend analysis between years, accurate carbon accounting of IT emissions, dashboards of monthly activity with electricity consumption, and associated emissions.
Real-time data unlocks live insights on electricity carbon intensity and price, raises awareness among teams, and paves the way for optimization in time and space

Forecasted data represents the cutting edge. Predictions over multiple days empower advanced and automated decision-making to optimize usage, increase efficiency, reduce costs, and emissions.

Going further

Let’s illustrate the importance of granularity when sourcing grid signals. We take as an example a data center located in Germany, for which we approximate the load profile by the curve below:

You can explore how electricity demand and electricity mix fluctuate in space and time, and how these affect other characteristics of the grid such as the carbon intensity, or the share of renewables on our two freely available tool below:

Live map
Data explorer

This already highlights one key consideration when choosing the right data to ingest into your products and solutions: granularity matters! The granularity you need for your solution also depends on other factors, such as: how granular is your electricity consumption data? How frequently and granularly can your monitoring be?

Insights increase with granularity. Moving from a yearly to hourly granularity improves accuracy by 20% on average worldwide, and by more than 40% in several grids. Value is already achieved with small improvements in granularity; switching from a yearly to a monthly view already improves accuracy by more than 10% worldwide.

You will learn more about how to source these grid signals and the key considerations to keep in mind for a successful data ingestion in the second lesson of this course.

The types of insights unlocked are not only a function of the granularity, but also of the time perspective considered:

Historical data powers trend analysis between years, accurate carbon accounting of IT emissions, dashboards of monthly activity with electricity consumption, and associated emissions.
Real-time data unlocks live insights on electricity carbon intensity and price, raises awareness among teams, and paves the way for optimization in time and space

Forecasted data represents the cutting edge. Predictions over multiple days empower advanced and automated decision-making to optimize usage, increase efficiency, reduce costs, and emissions.

Going further

Let’s illustrate the importance of granularity when sourcing grid signals. We take as an example a data center located in Germany, for which we approximate the load profile by the curve below:

Grid carbon intensity (hourly & monthly), and an example of a data center load curve in Germany

We consider two possible scenarios for the data center operator for monitoring their emissions: using a monthly carbon intensity of the German grid (282 gCO2/kWh in August), or using the hourly carbon intensity of the grid on the 4th of August 2025, already used in the example above.

To calculate the data center’s emissions from electricity consumption, the operator must multiply the load for each hour by the monthly emission factor (the same factor for each hour) or the hourly emission factor for each hour. With a monthly emission factor, the emissions computed per hour have the same profile as the load. Hours of high load are considered hours of high emissions, and inversely, overseeing how clean the grid might be at each hour, and how that fluctuates over time.

We consider two possible scenarios for the data center operator for monitoring their emissions: using a monthly carbon intensity of the German grid (282 gCO2/kWh in August), or using the hourly carbon intensity of the grid on the 4th of August 2025, already used in the example above.

To calculate the data center’s emissions from electricity consumption, the operator must multiply the load for each hour by the monthly emission factor (the same factor for each hour) or the hourly emission factor for each hour. With a monthly emission factor, the emissions computed per hour have the same profile as the load. Hours of high load are considered hours of high emissions, and inversely, overseeing how clean the grid might be at each hour, and how that fluctuates over time.

We consider two possible scenarios for the data center operator for monitoring their emissions: using a monthly carbon intensity of the German grid (282 gCO2/kWh in August), or using the hourly carbon intensity of the grid on the 4th of August 2025, already used in the example above.

To calculate the data center’s emissions from electricity consumption, the operator must multiply the load for each hour by the monthly emission factor (the same factor for each hour) or the hourly emission factor for each hour. With a monthly emission factor, the emissions computed per hour have the same profile as the load. Hours of high load are considered hours of high emissions, and inversely, overseeing how clean the grid might be at each hour, and how that fluctuates over time.

We consider two possible scenarios for the data center operator for monitoring their emissions: using a monthly carbon intensity of the German grid (282 gCO2/kWh in August), or using the hourly carbon intensity of the grid on the 4th of August 2025, already used in the example above.

To calculate the data center’s emissions from electricity consumption, the operator must multiply the load for each hour by the monthly emission factor (the same factor for each hour) or the hourly emission factor for each hour. With a monthly emission factor, the emissions computed per hour have the same profile as the load. Hours of high load are considered hours of high emissions, and inversely, overseeing how clean the grid might be at each hour, and how that fluctuates over time.

Electricity emissions of data center computed using both hourly and monthly emission factors

With hourly emissions factors, however, more insights are uncovered: the highest emissions are calculated at 8 am, even though the load is not at its peak yet, and inversely, emissions remain low at 2 pm, even though the data center remains at its peak consumption. The reason behind this is that the grid carbon intensity reached a peak at 8 am in the morning when consumption was high on the grid, and reached a minimum at the beginning of the afternoon when solar generation was high.

This example illustrates the importance of more granular signals in Sustainable IT Monitoring.

All numbers are available in the table below, where low values are highlighted in green and high values in red.