Google has released new data showing the energy, carbon, and water use linked to its Gemini AI system. The report is one of the most detailed disclosures from a major tech company about the environmental footprint of artificial intelligence.
The numbers per query seem small, but the rise of AI worldwide makes these findings crucial. They help us grasp the larger sustainability challenge. Let’s examine Google’s report findings.
From Queries to Carbon: Measuring AI’s True Cost
Artificial intelligence systems require powerful data centers to process user prompts. These data centers run on large amounts of electricity and water for cooling. To provide more transparency, Google calculated the average environmental cost of a single Gemini AI text query.
The company reported that one prompt:
- Uses about 0.24 watt-hours of electricity (similar to watching TV for less than nine seconds)
- Produces about 0.03 grams of CO₂ equivalent (CO₂e)
- Consumes about 0.26 milliliters of water (roughly five drops)
Google looked at the energy used to run AI and also considered the electricity used when the system is idle. Additionally, it factored in the extra infrastructure that supports data centers.
The chart below shows how much energy different AI models use for each prompt. The results come from two types of data: estimates (gray) and direct measurements (black, red, and blue).
It also shows that results can vary a lot depending on how the energy use is measured. For example, the model Llama 3.1 (70B) was found to handle anywhere from about 580 prompts to 3,600 prompts per kilowatt-hour, depending on the method used.
By comparison, Google’s Gemini Apps prompts had a narrower and more consistent range of results. By releasing these figures, the company hopes to create a standard way of reporting AI’s environmental impact.
Smarter, Faster—But Still Energy Hungry
Google reported that Gemini has become much more efficient compared to earlier versions. On a per-query basis, the AI now uses about 33 times less energy than it did a year ago. This improvement comes from advances in hardware, optimized algorithms, and better data center operations.
However, efficiency gains do not necessarily mean lower overall emissions. The demand for AI services is growing rapidly, leading to more total queries. As a result, even though each prompt is cheaper in energy terms, the combined usage continues to increase.
This trend is an example of the “Jevons paradox,” where greater efficiency can lead to higher total consumption when demand rises quickly. Google’s own environmental report shows this effect. The company’s total greenhouse gas emissions have risen 51% since 2019, with AI being a key driver.
Data Centers and Their Rising Power Needs
Google’s data centers are at the heart of AI operations. In 2024, these facilities consumed 30.8 million megawatt-hours of electricity, more than double the amount in 2020. This sharp rise highlights the scale of resources required to support AI growth.
At the same time, Google has made efforts to reduce the climate impact of its facilities. Even as electricity demand increased by 27%, the company cut its direct data center emissions by 12%. This was achieved through clean energy contracts, efficiency upgrades, and improved cooling technologies.
Google has made deals with utilities in Indiana and Tennessee. These agreements enable Google to lower data center power use when grid demand is high. This strategy, known as demand response, helps prevent blackouts and lowers stress on local power systems.
Beyond Wind and Solar: Google’s Nuclear Bet
While renewable energy remains central to Google’s strategy, the company is also exploring new approaches to meet the constant power needs of AI. Key actions include:
- Advanced nuclear power: The company partners with Kairos Power and the Tennessee Valley Authority. This will support molten salt nuclear reactors. They can deliver reliable, low-carbon energy.
- Demand-response agreements: Reducing electricity use during peak times in states like Indiana and Tennessee to ease grid strain.
- Expanded clean energy contracts: Securing renewable sources to match rising data center demand.
These steps show that Google is pursuing a mix of solutions beyond traditional renewables. Nuclear power, in particular, is seen as a stable complement to solar and wind for 24/7 operations.
Transparency or Greenwashing? The Debate Over Metrics
Google’s decision to share detailed per-query metrics has been praised as a step toward industry-wide accountability. Few tech companies have provided such clear data. This transparency helps policymakers, researchers, and the public see the real costs of AI.
At the same time, experts have raised concerns about what the report leaves out. Some argue that Google’s calculations do not fully account for indirect emissions or the impact of where electricity is sourced. Others note that per-query figures, while helpful, may downplay the large-scale effects of billions of queries worldwide.
The contrast between small individual costs and large overall emissions illustrates the complexity of the issue. It shows why companies need efficient technology and broad strategies. They have to manage total demand and align with climate targets.
Can AI Innovation Outpace Emissions?
Google’s disclosure highlights the balancing act facing the entire AI industry. On one hand, new technology can drive efficiency, reduce per-query energy use, and open pathways to sustainable power. On the other hand, the sheer scale of AI adoption risks outpacing these improvements.
For AI to grow in a sustainable way, companies will need to combine efficiency gains with renewable energy, nuclear solutions, and smarter grid management. Transparency will also play a central role in building trust and creating common standards across the sector.
As more companies adopt AI and integrate it into daily life, the question of energy and carbon costs will become even more urgent. Google’s report is an early attempt to measure and address this challenge.
Whether the industry can keep total emissions in check while meeting growing demand will shape the future relationship between AI and the environment.
Google’s new data provides a clearer picture of what it takes to run AI systems like Gemini. One prompt uses just a few drops of water and a tiny bit of carbon. But when billions of people interact, the environmental impact grows massively.
By being transparent about these numbers, Google has set a benchmark for the industry. The next challenge will be to turn these insights into broader changes that ensure AI grows without driving emissions sharply higher. The path forward will require innovation, investment, and cooperation across the technology and energy sectors.
The post Google Reveals the Environmental Cost of Gemini AI Query appeared first on Carbon Credits.
Insetting program design starts with your supplier map. The carbon claim comes much later.
Consider a coffee procurement team that funds a farm intervention before checking its sourcing data. The supplier aggregates coffee lots from several regions. The buyer lacks farm-level traceability. Another purchaser expects to report the same outcome. A more attractive project brochure solves none of these problems.
If the distinction between a value-chain intervention and a purchased credit is still open, read CCC’s Insetting vs Offsetting guide before you set the brief.
Insetting program design begins with the purchasing relationship and works outward. You need to decide 1) where an intervention belongs, 2) what business problem it addresses, 3) who pays for it, 4) how results will be measured, and 5) what the evidence will let you say about them. This five-step playbook turns that sequence into a procurement plan.
Step 1: Insetting program design starts with supplier mapping
Start with the commodities that matter most to your Scope 3 inventory and sourcing continuity. Ask finance for spend by supplier, sustainability for the emissions baseline, and category managers for origin and volume records. Then identify which suppliers can trace material to a jurisdiction, sourcing region, producer group, or individual land-management unit.
The level of traceability affects what you can account for. Under the GHG Protocol Land Sector and Removals Standard, the spatial boundary for purchased land-based products depends on the traceability associated with each volume. Farm-specific factors or removals require physical traceability through an acceptable chain-of-custody model. A contract with a tier-one trader, by itself, does not establish a link to the farms behind its deliveries.[1]
Build a first-pass table with five columns: commodity and annual volume; supplier and origin; current traceability level; land-use and production risks; and the business owner who can change the purchasing terms. Mark missing data plainly. A blank cell is cheaper to fix before a program starts than after a claim is drafted.
If your company sets an SBTi Forest, Land and Agriculture target (FLAG), keep its land-related target boundary separate from energy and industry emissions. The current FLAG guidance also requires relevant companies to assess primary deforestation-linked commodities in their value chains.[2]
Step 2: Choose a commodity and jurisdiction for a pilot
Select one supply corridor where the sourcing connection can be established and procurement has a reason to stay. A useful pilot has repeat purchases, suppliers willing to share records, an intervention matched to a diagnosed problem, and a defined plan to measure the result. Carbon potential alone is a poor selection rule if the supplier cannot deliver the crop your business needs.
Compare candidate regions on land-use risk, supplier concentration, local rights and tenure, the availability of baseline data, and exposure to drought, erosion, or forest conversion. Check the relevant legal requirements separately. A nature project does not satisfy a due-diligence obligation merely because it has a carbon method attached.
Consider a coffee buyer sourcing from two districts. In the first, producer-group records identify the farms, while recent dry seasons have increased rejected green coffee lots. In the second, deliveries pass through several intermediaries and farm origins are unknown. Pilot shade management in the first district, with local agronomic advice and quality checks built into the plan. The buyer can trace purchases there and test the intervention against a defined sourcing problem.
Use a short decision memo to record why the selected corridor beats the alternatives, what remains uncertain, and what would stop the pilot. That memo is the first control against a project drifting toward whatever happens to photograph well.
Step 3: Design the supplier engagement and commercial model
The intervention needs an owner on each side of the contract. Decide who supplies planting material or training, who maintains the practice, who bears the cost if a harvest disappoints, and how long the buyer expects to purchase from participating suppliers. Where smallholders participate, budget cash flow and technical support for practices that offer a later return.
Put these terms in a supplier agreement or pilot addendum. It should cover eligible farms and volumes, consent to data collection, access for monitoring, any land or carbon rights that matter, benefit sharing, and what happens if the purchasing relationship ends. Ask local counsel and producer representatives to review the arrangement. The Accountability Framework is a useful reference for supply-chain work involving forests, ecosystems, and human rights.[3]
Avoid assigning an automatic carbon benefit to the buyer. The contract should say which party has the right to report each outcome, how overlapping claims will be checked, and how any issued credits will be handled. A reduction reported in a buyer’s inventory cannot simply be sold for someone else’s offset claim without an accounting review. SBTi FLAG guidance excludes purchased credits used as offsets from near-term FLAG target achievement.[2]
Set commercial measures beside environmental ones: delivered volume, rejection rate, supplier retention, and procurement cost. They tell you whether the program works as a supply-chain investment, even if carbon accounting takes longer to resolve.
Step 4: Build monitoring, reporting, and verification before rollout
Agree on the baseline and data flow before paying for a full rollout. Identify the land units or sourcing regions in scope, the practices that change, the records suppliers already keep, and the additional measurements the program requires. Decide which data can come from field visits, transaction records, or remote sensing, and who checks each source.
For carbon, separate three questions. Did the practice happen? Did emissions or removals change under an applicable method? Can the buyer attribute a defensible share to its purchased goods without double counting? The GHG Protocol’s 2026 guidance includes dedicated chapters on spatial boundaries, data and methods, evaluating actions, credited outcomes, and assurance.[1] These are distinct jobs. A photograph of a planted tree answers only part of the first.
CCC’s MRV and Additionality article explains the separate evidence questions a project reviewer will ask.
Create an evidence file for every reporting period: dated supplier and farm records, purchase volumes, methods and emission factors, uncertainty, quality checks, and changes from the baseline. Specify who can challenge the data and how corrections are recorded. If a project also seeks credit certification, keep project accounting and corporate inventory accounting separate. The GHG Protocol says its land standard focuses on entity-level inventories and does not set the rules for credit certification.[4]
Start with the minimum viable measurement design that can survive review. Add complexity when the pilot tells you which data actually changes a decision.
Step 5: Decide what can be reported and claimed
Insetting program design needs reporting rules before the marketing copy. Procurement, sustainability, finance, and legal should agree on four distinct outputs: operational performance; observed land or nature outcomes; inventory effects, if the accounting criteria are met; and any separately issued credits. One dashboard can display them, but they should not share a single unqualified headline.
SBTi FLAG guidance allows certain land-based removals within a company’s operations or supply chain to contribute to FLAG pathways, subject to criteria including permanence, traceability, data quality, and allocation. It does not allow purchased offsets to meet near-term FLAG targets.[2] Eligibility therefore depends on evidence and the applicable target rules, not on calling an activity “insetting.”
Draft a claims register. For each proposed sentence, record the period, boundary, data owner, calculation method, reviewer, and whether the result is measured, estimated, or still a goal. If the evidence supports only “we funded a pilot with 200 farms,” say that. Do not turn a funding fact into a verified emissions reduction.
What good looks like by year three
Treat year three as a review point, not a promised payback date. A mature pilot should have mapped participating supply volumes, a stable supplier agreement, a repeatable data collection process, and a comparison of commercial and environmental results against the baseline. It should also have a documented answer for outcomes that cannot be included in the buyer’s inventory.
Scale only where the pilot improves a sourcing decision and the evidence can follow the product. If neither condition holds, redesign the intervention or change the corridor. A disciplined program can conclude that a site is unsuitable. That is useful procurement intelligence.
Discuss your insetting program with CCC
Carbon Credit Capital designs nature-based supply chain investments around sourcing needs and measurable sustainability outcomes. Schedule a consultation with our carbon and sustainability experts to map your tier-one supplier network, FLAG exposure, and first pilot decision.
Sources
[1] GHG Protocol, Land Sector and Removals Standard, version 1.1, and companion Guidance, June 2026: https://ghgprotocol.org/land-sector-and-removals-standard ; https://ghgprotocol.org/land-sector-and-removals-guidance
[2] Science Based Targets initiative, Forest, Land and Agriculture Science-Based Target-Setting Guidance, version 1.2, March 2026, especially pp. 16–17, 37, and 41: https://sciencebasedtargets.org/resources/files/SBTiFLAGGuidance.pdf
[3] Accountability Framework initiative, Accountability Framework and guidance for companies: https://accountability-framework.org/
[4] GHG Protocol, Frequently Asked Questions: Land Sector and Removals Standard, questions 10 and 18: https://ghgprotocol.org/blog/frequently-asked-questions-land-sector-and-removals-lsr-standard
Nature-based carbon investments deserve consideration when they address a specific business dependency and produce outcomes you can measure. Imagine Moonbucks, a global coffee roaster planning next season’s espresso blend. It is considering natural Grade 4 coffee from Jimma, in southwestern Oromia, Ethiopia. Its sourcing decisions affect the roasting schedule and the cash tied up in coffee purchases. Moonbucks needs to understand which measures could protect that supply, what they would cost and how it would judge the results.
1. Define the sourcing problem before choosing a project
Moonbucks needs a dependable component for its espresso blend. Its procurement team wants continuity of volume and an agreed cup profile. Replacing a shipment at short notice could require new sample approval and changes to the blend, and given the schedules for cultivation, processing, and freight delivery, is practically impossible. Moonbucks needs to get this right, and its initial concern is inconsistent deliveries. It commissions an assessment of the supplying farms and post-harvest chain. The questions cover land condition, harvest availability and whether drying or storage problems are affecting saleable lots.
This distinction changes the investment brief. A tree-planting proposal cannot be assumed to fix drying problems. Better storage cannot restore eroded land. The team must establish which problem exists before selecting the response.
CCC’s guide to finding nature-based solutions in your supply chain explains where to look. Here, the next task is to turn a broad concern into a testable proposal.
2. Develop nature-based carbon investments around the actual supply chain
The assessment would trace Moonbucks’ purchases back through the exporter and supplying producer groups. Which farms supply the contracted lots? Who controls land-management decisions? How long does the buyer expect to maintain the purchasing relationship?
Furthermore, Jimma natural Grade 4 cannot be treated as interchangeable with every Ethiopian coffee. Moonbucks would need to test any substitute against its own blend and quality requirements.
Suppose Moonbucks’ assessment identifies an opportunity to maintain shade trees and restore vegetation along exposed streambanks. The IPCC describes agroforestry as a way to store carbon in vegetation and soil, with potential benefits for erosion control and water quality.1 Moonbucks would need local advice on species, growing conditions and maintenance before deciding how these measures fit its sourcing plan.
The proposal now has a defined sourcing connection and a candidate intervention. The next steps are to test its effect on production, assess carbon eligibility and build the financial case.
3. Compare the proposal with other ways to protect supply
WBCSD’s Nature-based Solutions Blueprint recommends comparing nature-based options with other responses across their benefits and costs.2 For Moonbucks, that means evaluating the landscape proposal alongside drying improvements, storage controls and alternative sourcing arrangements.
The product’s natural-process designation matters. Moonbucks should examine the drying and dry-milling chain alongside land conditions. Its team would check storage practices, contamination risks and incoming-lot quality to see whether problems arise before or after the coffee leaves the farms.
The revised proposal might combine land stewardship with separate post-harvest controls. Each component needs its own purpose and budget. A sustainability label should not obscure which activity is expected to solve which problem.
The project develops as the evidence improves. Its scope should follow the diagnosis rather than preserve the first idea unchanged.
4. Establish what the carbon outcome could count toward
Moonbucks’ sustainability team now tests the proposed carbon case separately from the sourcing case. Is the objective a measurable change in Moonbucks’ value-chain emissions, verified credits for separate voluntary use, or distinct outcomes under carefully defined arrangements?
Scope 3 covers value-chain emissions. A purchasing relationship alone does not prove that a farm intervention can be included in Moonbucks’ inventory. The team needs an applicable accounting method, defensible boundaries and evidence supporting attribution.
SBTi’s FLAG Guidance (forest, land and agriculture), version 1.2, says purchased credits used as offsets cannot meet near-term FLAG or energy and industry targets. Eligible land-based removals are subject to accounting requirements.3 Moonbucks needs to test its proposal against those rules before including an outcome in its targets.
For Moonbucks’ Jimma proposal, maintaining existing trees and establishing new vegetation could involve different carbon questions. Specialists must examine the baseline and additionality, meaning what changes because the project proceeds. They must also assess measurement and the risk that stored carbon is lost. CCC’s MRV and additionality article introduces those tests.
Until that work is complete, Moonbucks should not budget an assumed Scope 3 reduction or credit-sale receipt. Direct emissions reductions remain the priority.
5. Build a financial case that survives weaker assumptions
Moonbucks’ finance team would now build a multi-year budget. UNEP’s State of Finance for Nature 2023 compares current finance with the investment needed to meet climate, biodiversity and land-restoration goals.4 The company would cost its proposed activities, farmer participation and ongoing monitoring. Verification, where required, is a separate expense. Maintenance and responsibility for replacing failed plantings must remain visible after the initial establishment budget ends.
Operational benefits need equally careful treatment. If fewer rejected lots reduce emergency purchasing, count that saving once. Do not add “supply resilience” as another financial benefit when it represents the same avoided cost.
Moonbucks should compare a base case with slower benefits, higher implementation costs and no carbon revenue. It should also distinguish production outcomes from quality and shipment outcomes. A single good harvest would not prove that the landscape intervention caused better delivery performance.
A project can remain worth funding even when its carbon case needs more work. That decision must rest on the operational evidence and the company’s objectives, without unsupported environmental claims.
6. Fund a pilot with a clear expansion decision
Moonbucks’ first approval would fund a bounded pilot, rather than commit the entire sourcing program. Before spending, Moonbucks would agree on participating farms, responsibilities and the evidence needed at review.
Its scorecard could track farmer participation, implementation costs and plant survival, alongside accepted coffee volumes and lot-quality results. Carbon measurements would follow the selected method. Farm-level traceability must connect the intervention to the actual sourcing relationship.
Suppose Moonbucks’ review later finds that storage changes explain most quality improvements, while the land measures need longer observation. Those findings would shape the next decision: expand the proven component, adjust the landscape work or pause further spending. They would not justify crediting all improvements to tree planting.
This gives nature-based carbon investments a practical place in a multi-year strategy. Funding increases as the evidence supports it, and the project can change direction before a larger commitment.
Discuss a sourcing opportunity with CCC
Carbon Credit Capital’s nature-based supply-chain investment service uses its Dual-Value Model to examine supply-chain economics alongside potential sustainability outcomes.
Schedule a consultation to discuss your sourcing exposure and the questions to resolve before allocating capital.
Sources
- Intergovernmental Panel on Climate Change (IPCC), Climate Change 2022: Mitigation of Climate Change, Chapter 7, Agriculture, Forestry and Other Land Uses, section 7.4.3.3 (Agroforestry).
- World Business Council for Sustainable Development (WBCSD), Nature-based Solutions Blueprint, 5 June 2024.
- Science Based Targets initiative (SBTi), Forest, Land and Agriculture Science-Based Target-Setting Guidance, version 1.2, March 2026, section 3.1.4, pp. 37–38.
- United Nations Environment Programme (UNEP), State of Finance for Nature 2023, 2023.
A plastic credit funds the collection of plastic waste in places where no formal system exists to collect it. Each credit represents a measured, independently audited quantity of plastic taken out of the environment or returned to use, tracked on a public registry.
