Sustainability with Google Cloud

Study guide for Cloud Digital Leader, Unit 6 · Topic 3. This is the topic's lecture in reading form — every slide's teaching, figures and worked examples, in order — followed by the official Google Cloud pages its claims rest on.

What the exam guide asks. Discuss how Google Cloud helps organizations meet sustainability goals and reduce environmental impact.

Objectives, quoted from the exam guide:

1. Describe Google Cloud’s commitment to sustainability and reducing environmental impact.
2. Discuss how Google Cloud provides products to support organizations’ sustainability goals.

Sustainability with Google Cloud

What Google commits to, and what it hands the customer

Google's side — a 24/7 carbon-free energy goal, a metric that measures it hour by hour, and accounting that shows the work. The customer's side — tools to see its own footprint, choose cleaner regions and times, and stop paying energy for idle resources.

The last topic of the exam has two objectives, and they are the two sides of one arrangement. The first is Google's side: its commitment to sustainability and to reducing environmental impact. The second is the customer's side: the products Google provides so an organization can pursue its own sustainability goals. The guide's summary puts them together — how Google Cloud helps organizations meet sustainability goals and reduce environmental impact. The useful way to hold this topic is that Google controls how clean the electricity in a region is, and the customer controls where, when and how much it runs. Neither side can finish the job alone, and every tool on the second half of this deck exists to connect the two.

Google Cloud's commitment

Carbon-free energy every hour, in every region, by 2030

• Regions draw electricity from the local grid
• Goal: match consumption with carbon-free energy, hourly, per region, by 2030
• Google adds carbon-free generation on top of what the grid supplies
• Efficiency in the hardware: TPUs and newer machine types use less energy per unit of work

Worked example (synthetic). A retailer's sustainability officer asks whether "running on Google Cloud" makes its footprint zero. The honest answer is the goal's own wording: matching consumption hour by hour is a 2030 target, and the grid under each region still varies today.

Start with the fact everything else follows from. To power each Google Cloud region, Google uses electricity from the grid where the region is located. So a region is only as clean as its grid plus whatever Google adds to it. Google's commitment is stated as a goal with three qualifiers that each matter: to match its energy consumption with carbon-free energy, or CFE, every hour, and in every region, by 2030. Every hour means it is not an annual average that lets a sunny afternoon offset a windless night. Every region means a clean region cannot offset a dirty one. And by 2030 means it is a target, not a present fact — the exam will reward the candidate who reads it that way. The mechanism is added supply: in addition to the carbon-free energy already supplied by the grid, Google has added carbon-free energy generation in a location to reach its 24/7 carbon-free energy objective. The other half of reducing impact is using less energy for the same work, and Google's framework states it at the hardware level: tensor processing units, or TPUs, are engineered for optimal energy efficiency, and when machine types are updated they are often designed to be more energy-efficient, with higher performance per watt.

How a region's CFE% is built

Two inputs every hour, one number a customer can act on

Figure: A flow chart: grid generation in a given hour and Google-attributed clean energy on that grid combine into an hourly carbon-free energy percentage for the region, which is averaged over the year. The yearly figure tells a customer the share of time its application runs on carbon-free energy, and a region at 75 percent or above is labelled low carbon.

Worked example (synthetic). Two candidate regions meet a team's latency needs. The team does not need to know how either grid is built — the higher CFE% is the one where its application spends more of its time on carbon-free energy.

Google measures its commitment with one metric, and it is worth knowing how it is made, because the exam asks what it means. CFE%, the carbon-free energy percentage, is calculated for every hour. It tells Google what percentage of the energy it consumed during that hour was carbon-free, and it depends on two elements: the generation feeding the grid at that time — which power plants are running — and the Google-attributed clean energy produced onto that grid during that time. Google then aggregates the average hourly CFE percentage for each region over a year; at retrieval, the published figures were for 2025, and they are replaced annually, which is why this deck teaches the metric and not any region's number. For a customer the figure has a direct reading: it represents the average percentage of time your application will be running on carbon-free energy. And it drives a label. For a location to be called low carbon, it must belong to a region with a Google CFE% of at least 75 percent, or, where CFE% is not available, a grid carbon intensity below a stated ceiling.

Showing the work: two ways to count the same electricity

Market-based credits Google's clean-energy purchases; location-based does not

Worked example (synthetic). A finance team sees two different emissions totals for the same month and suspects an error. Neither is wrong: one leaves Google's clean-energy purchases out, the other credits them.

A commitment is only as credible as the accounting behind it, and this is where Google shows its work. Google's Carbon Footprint methodology reports the same electricity two ways. Location-based emissions do not take into account Google's renewable power purchase agreements or other contracts for carbon-free electricity — they show the physical grid. Market-based emissions do include the impact of Google's carbon-free electricity purchases for the appropriate data centers, following the Greenhouse Gas Protocol's market-based method. Underneath both, the calculation is built from the bottom up, relying heavily on machine-level power and activity monitoring inside Google data centers, and overhead energy — power systems, cooling and lights — is allocated hourly to every machine and its users. That last point matters for the exam's phrase reducing environmental impact: the data-center overhead is counted, not hidden. One honest limit belongs on the same slide. Google states that the customer-specific greenhouse gas data in Carbon Footprint has not been third-party verified or assured.

The products that support a customer's goals

See the footprint, choose where and when, remove the waste

• Carbon Footprint: emissions per billing account, computed automatically
• Region carbon data, Region Picker, Cloud Location Finder: choose a cleaner place
• Time-shifting batch work to the hours the grid is cleanest
• Autoscaling and the unattended project recommender: stop powering idle resources

Worked example (synthetic). A media company must report its cloud emissions next quarter and cut them next year. Carbon Footprint gives it the figure for the report; the region and scheduling tools are how it cuts the next one.

The second objective is the customer's toolkit. Google frames it as the other half of its own goal: as it works towards 2030, it wants to empower customers to use its 24/7 carbon-free energy efforts and consider the carbon impact of where they locate their applications. The toolkit maps onto three things a customer actually controls. First, visibility. Carbon Footprint provides visibility for each customer into the climate impacts of the products it buys from Google Cloud, so that it can report on and act to reduce them. It needs no setup: the data is computed automatically for the billing account, with no application programming interface to enable. Second, place and time. Google's framework calls the choice of region an important architectural decision for a workload's carbon footprint, and it names the tools: the Google Cloud Region Picker helps select regions based on carbon footprint, cost and latency, and Cloud Location Finder finds locations based on proximity, carbon-free energy usage and other parameters. Batch work adds the time dimension — techniques like time-shifting and carbon-aware scheduling run batch workloads in the hours when the grid's carbon intensity is lowest. Third, waste. Automated and dynamic scaling prevents energy waste from idle or over-provisioned infrastructure, and the unattended project recommender finds projects nobody is using so they can be reclaimed or removed. Underneath all of it sits the Architecture Framework's sustainability pillar, which gives recommendations for workloads that are energy-efficient and carbon-aware.

Which tool answers which question

Name the question in the scenario, then the tool

Worked example (synthetic). An item asks which tool a company uses to include its Google Cloud emissions in its annual greenhouse gas report. The graded answer is Carbon Footprint — its data goes in as Scope 3.

Exam items on this objective arrive as a question someone in the business is asking, so rehearse the mapping in that direction. "What are our cloud emissions?" is Carbon Footprint. Its reports follow the Greenhouse Gas Protocol, and because Google allocates its own Scope 1, 2 and 3 emissions to customers by usage, a customer can put that data into its own reports as Scope 3 — indirect emissions in its value chain. The dashboard shows the figure two ways, on a market-based tab and a location-based tab. "Can we analyze it alongside everything else?" is the export: Carbon Footprint data can be exported to BigQuery for analysis or custom dashboards and reports. "Which region is cleanest for us?" is the region carbon data and the Region Picker. "When should this batch job run?" is time-shifting. And "what are we running that nobody uses?" is the unattended project recommender, because unattended resources are hard to identify and tend to cause unnecessary waste as well as security risk. The trap in these items is picking the tool you have heard of rather than the one that answers the question asked.

Three levers the customer controls

Where, when, and how much

Figure. Three cards. Where: run long-lived applications in the region with the highest carbon-free energy percentage available, after data residency requirements. When: schedule batch work for the hours when grid carbon intensity is lowest. How much: scale with demand and remove unused resources, since less wasted energy lowers both cost and emissions.

Worked example (synthetic). A bank must keep customer data in one country. "Pick the cleanest region" still applies — but only among the regions residency allows. The levers work inside constraints, not instead of them.

Every tool in this objective moves one of three levers, and naming them is the fastest way through a scenario. Where: Google's region carbon page says that if you run an application over time, running in the region with the highest CFE% — the carbon-free energy percentage — will emit the lowest carbon emissions. The framework adds the constraint that comes first — data residency and sovereignty is a foundational factor that dictates the choice of region — so the cleanest region is chosen among the ones you are allowed to use. When: batch workloads can be time-shifted to hours when the grid is cleanest, because you can predict when they must run. How much: the framework says the reduction in wasted energy translates to lower costs and lower carbon emissions. That sentence is why sustainability is a business case and not only an ethical one, and Google's pillar says it outright — sustainability is not a trade-off against other business objectives; its practices help accelerate them.

Measure, act, verify — then measure again

A tool used once is a report; used in a loop it is a programme

Figure: A loop of four steps: quantify emissions with Carbon Footprint, identify carbon hotspots, implement targeted workload optimizations, verify the outcome, and return to quantifying.

Worked example (synthetic). A retailer moves one batch pipeline to a cleaner region and schedules it overnight. Next month's Carbon Footprint figure is how it knows whether that worked — the verify step, not the move, closes the loop.

The framework's measure-and-improve principle turns the toolkit into a practice. Its recommendations use Carbon Footprint to quantify carbon emissions, identify carbon hotspots, implement targeted workload optimizations, and verify the outcomes of those optimizations. Drawn as a loop, the point is the arrow back to the start: verification is the next measurement. The framework also says why a finance leader should care — this approach lets you align cost optimization goals with verifiable carbon reduction targets. The same resource you stop wasting shows up in both the bill and the footprint, so a sustainability programme and a cost programme can share one set of numbers.

What this topic actually tests

Two discriminations, two objectives

A goal or a fact? 24/7 carbon-free energy in every region is a 2030 goal, and CFE% measures progress hour by hour. Which tool for which question? Carbon Footprint to see and report; region data and Region Picker to choose where; time-shifting to choose when; scaling and the unattended project recommender to stop the waste.

Close the topic, and the exam, on two discriminations. First, read Google's commitment exactly: matching consumption with carbon-free energy every hour and in every region is a goal for 2030, and CFE% is the hourly measure of how close each region is today. An answer that says Google Cloud already runs entirely on carbon-free energy everywhere misreads the goal. Second, match the tool to the question. Carbon Footprint answers how much, and feeds a customer's own reporting as Scope 3. Region carbon data, the Region Picker and Cloud Location Finder answer where. Time-shifting answers when. Scaling and the unattended project recommender answer how much is wasted. And the reason all of this belongs in a business exam is the framework's own claim: less wasted energy is lower cost and lower emissions at the same time.

Official sources for this topic

• Cloud Digital Leader exam guide — Section 6
• Carbon-free energy for Google Cloud regions
• Well-Architected Framework — Optimize resource usage for sustainability
• Carbon Footprint reporting methodology
• View Carbon Footprint data
• Well-Architected Framework — Use regions that consume low-carbon energy
• Unattended project recommender
• Google Cloud Well-Architected Framework — Sustainability
• Export your carbon footprint
• Well-Architected Framework — Continuously measure and improve sustainability