Skip to content

#5 of 30 · Enterprise AI product

Salesforce (Agentforce)

Created the AI Energy Score; runs its agents on other people's models.

San Francisco, California·salesforce.com ·Verified Oct 16, 2026
BGood confidence
Index score
Good confidence

In plain words

How scores work

Salesforce scores 64 out of 100 — middle of the pack. Some disclosure or clean-energy commitments, offset by rising emissions or blind spots. Confidence in this score is good confidence (B).

What helps

  • Efficiency · Agentforce routes to small xGen/xLAM models where possible and publishes AI Energy Score ratings for them.
  • Carbon · 100% renewable procurement; Scope 1+2 market-based down 50% ahead of FY2031; net-zero residual emissions maintained through credits.
  • Transparency · Annual Stakeholder Impact Report with assured metrics; first proprietary-model disclosures under the AI Energy Score.

What hurts

  • Efficiency · Frontier calls go to partner models with no per-prompt figures.
  • Transparency · No per-prompt figures for Agentforce.

What nobody knows

  • Per-prompt water disclosed
  • Training emissions in model cards
  • Water-use effectiveness reported
  • Hardware reuse or e-waste reported

Salesforce's Agentforce rides on Amazon Bedrock and partner frontier models plus its own small xGen and xLAM models. The company maintains net-zero residual emissions across its value chain through carbon credits, procures renewable certificates equal to 100% of electricity, cut Scope 1+2 by 50% ahead of its FY2031 target, and co-launched the AI Energy Score — becoming the first model developer to publish efficiency ratings for proprietary models.

Products
Agentforce · Einstein · xGen / xLAM models · Slack AI
Hosting
Hyperforce on AWS (plus Google Cloud and Azure regions); frontier models via Amazon Bedrock, OpenAI, Anthropic and Google.

Key figures

Per-prompt and fleet metrics. Hover a title for what it means; open “How we got this number” for the working.

How to read the labels:Company figureCompany figure. Published by the company itself. We take it as given and record the scope next to it.Measured independentlyMeasured independently. Measured or estimated by researchers or analysts with no stake in the result.ecogpt.co estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Not disclosedNot disclosed. The company publishes nothing on this and no defensible estimate exists.— hover any label.
Energy per promptElectricity used to answer one typical request. Under 0.3 Wh is good. A 10 W LED bulb uses 0.3 Wh in about two minutes.
0.30
Wh / typical prompt
≈ 25.0 smartphone charges per 1,000 prompts
ecogpt.co estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.Techstrong.aiEpoch AILuccioni, Jernite, Strubell (arXiv 2311.16863)

ecogpt.co estimate; no company figure.

How we got this number

Most Agentforce turns call a frontier partner model through Amazon Bedrock (≈0.3 Wh per Epoch's GPT-4o estimate); Salesforce's own xGen and xLAM models are small and sit in the 0.04 Wh class. Uncertainty roughly ±3×.

Water per promptWater evaporated cooling the servers for one request. Scope matters: on-site cooling only, or also the water used to generate the electricity (which adds ≈1.8 mL per Wh).
0.05
mL / typical prompt
≈ 0.09 500 mL bottles per 1,000 prompts
ecogpt.co estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.Techstrong.aiEpoch AILuccioni, Jernite, Strubell (arXiv 2311.16863)Amazon

On-site cooling only, so comparable with Google and OpenAI's figures.

How we got this number

0.3 Wh per prompt × 0.15 L/kWh (AWS's disclosed fleet WUE (Hyperforce runs on AWS)). Counting the water evaporated to generate the electricity adds ≈0.54 mL.

Carbon per promptGreenhouse gas from one request on the grid where it actually runs. Below 0.1 g is clean. A kilometre in a petrol car is about 170 g.
0.117
gCO₂e / typical prompt
≈ 0.69 km driven (petrol car) per 1,000 prompts
ecogpt.co estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.Techstrong.aiEpoch AILuccioni, Jernite, Strubell (arXiv 2311.16863)

Location-based: what the grid actually emitted. Renewable certificates would bring the market-based figure near zero.

How we got this number

0.3 Wh × 390 gCO₂/kWh (US grid, location-based).

Annual emissionsThe company's whole yearly footprint, in million tonnes of CO₂e. Clouds and device makers dwarf pure labs because they also count shops, logistics or products. Compare the trend, not the size.
0.08
MtCO₂e (FY2025, Scope 1+2 market-based)
Company figureCompany figure. Published by the company itself. We take it as given and record the scope next to it.Salesforce

Scope 3 ≈2.1 Mt before credits; net-zero residual via offsets.

Emissions trendDirection of travel against last year or the company's own baseline. Negative is good. Positive means the AI build-out is outrunning clean energy.
-50.0
% Scope 1+2 vs FY2019 baseline
Company figureCompany figure. Published by the company itself. We take it as given and record the scope next to it.Salesforce

Target for FY2031, met early.

Water per kWh (WUE)Litres of water consumed per kilowatt-hour of server electricity, across the fleet. 0.15 is excellent, 0.3 is good, 1.8 is the US data-centre average.
0.15
L/kWh (inherited from AWS)
ecogpt.co estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.Amazon
How we got this number

No disclosure; AWS's fleet figure applies to Hyperforce capacity.

Clean electricityShare of electricity matched with renewable or carbon-free sources. Annual matching counts certificates bought anywhere; hourly matching on the same grid is far stricter.
100
% renewable procurement (certificates and PPAs)
Company figureCompany figure. Published by the company itself. We take it as given and record the scope next to it.Salesforce

Hardware & materials

The chips, power, copper and future e-waste behind the fleet — the part sustainability reports leave out

Factors we use
AI chips in service
5,000
GPU-equivalents (order of magnitude)
≈ 69 racks of 72 GPUs
ecogpt.co estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.Techstrong.ai

GPUs, TPUs or custom accelerators the company runs. Analyst figures are in H100-equivalents, company figures in physical chips.

How we got this number

Salesforce runs small models of its own on cloud GPUs and routes frontier calls to partners, so its dedicated fleet is a few thousand GPU-equivalents at most. Uncertainty ±3×.

Power draw
7.00
MW
≈ 5,250 US homes' worth of power
DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Techstrong.aiSemiAnalysis

Megawatts the AI fleet pulls. One megawatt runs roughly 750 US homes.

How we got this number

5,000 accelerators × 1.3 kW each, all-in. Fleet power implied by the chip count.

Electricity a year
0.05
TWh / year
≈ 4,762 US homes for a year
DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Techstrong.aiSemiAnalysis

Terawatt-hours per year. One TWh is a year of electricity for about 95,000 US homes.

How we got this number

7 MW × 8,760 h × 80% utilisation. Compute fleet only.

Carbon baked into the hardware
2.6
ktCO₂e embodied in chips and servers
≈ 565 petrol cars driven for a year
DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Techstrong.aiarXiv 2607.05408NVIDIA developer blog

CO₂e emitted making the chips and servers before they answer a single prompt.

How we got this number

5,000 accelerators × 514 kg CO₂e each (chip plus share of the host server). Emitted in manufacturing, before the first prompt is served.

Future e-waste
85.0
tonnes of servers to retire within ~5 years
≈ 0.5 empty Boeing 747s by weight
DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Techstrong.aiNVIDIAThe Register

Tonnes of server hardware that will be retired in three to five years.

How we got this number

5,000 accelerators × 17 kg of server hardware each. Tomorrow's e-waste unless reused; generative AI could produce 2.5 Mt of it a year by 2030.

Copper in the build
189
tonnes of copper in the build
≈ copper for 2,277 electric cars
DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Techstrong.aiSemiAnalysisBHPS&P Global Market Intelligence

Tonnes of copper in cabling, busbars and transformers.

How we got this number

7 MW × 27 t of copper per MW (cabling, busbars, transformers).

Fossil generators on site
3.00
backup diesel generators (implied)
≈ 7.8 MW of backup capacity
DerivedDerived. Arithmetic on a disclosed or independent input using a published conversion factor.Techstrong.aiSemiAnalysisLatitude MediaLatitude Media / Floodlight

Diesel backup units or gas turbines at the data centres, permitted or observed.

How we got this number

7 MW ÷ 2.6 MW per generator, the average size permitted at US data centres. Permit counts are not published fleet-wide; this is the number a fleet this size normally carries.

Hardware reused or recycled
—
No hardware reuse or e-waste figure published.
Not disclosedNot disclosed. The company publishes nothing on this and no defensible estimate exists.

Share of retired servers and parts given a second life.

Mineral depletion
—
Only Mistral has published a mineral-depletion figure.
Not disclosedNot disclosed. The company publishes nothing on this and no defensible estimate exists.

Kilograms of antimony-equivalent, the life-cycle unit for scarce minerals consumed.

Score breakdown

Each pillar is 0–100; weights are fixed across the index

Energy efficiency — How much electricity one answer costs.weight 20% · confidence B
65
  • Agentforce routes to small xGen/xLAM models where possible and publishes AI Energy Score ratings for them. Salesforce
  • Frontier calls go to partner models with no per-prompt figures. Techstrong.ai
Carbon intensity — How clean that electricity is, and whether total emissions are rising.weight 25% · confidence B
65
  • 100% renewable procurement; Scope 1+2 market-based down 50% ahead of FY2031; net-zero residual emissions maintained through credits. Salesforce
  • Scope 3 of roughly 2.1 Mt dominates and depends on suppliers' progress — above all AWS. Salesforce
Water — How much water the data centres consume, and whether it is replaced.weight 15% · confidence C
55
  • No water figures of its own; AWS's fleet WUE of 0.15 L/kWh applies to hosted capacity. Amazon
Transparency — How much the company actually publishes.weight 20% · confidence B
62
  • Annual Stakeholder Impact Report with assured metrics; first proprietary-model disclosures under the AI Energy Score. SalesforceSalesforce
  • No per-prompt figures for Agentforce. Techstrong.ai
Commitments — Whether it has dated, measurable climate and water targets.weight 10% · confidence A
72
  • SBTi targets: 50% Scope 1+2 by FY2031 (met early), 90% by FY2041, 97% Scope 3 intensity cut by FY2041; 100% renewable maintained. Salesforce
Conduct — How it behaves: permits, lawsuits, gas turbines, community water.weight 10% · confidence B
70

Models

What this company ships, and what each prompt costs

ModelWeightsEnergy / promptWaterTraining CO₂e
Agentforce (typical agent turn)
Partner frontier models + xGen small models
Closed0.30 Wh Estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.0.045 mL Estimateecogpt.co estimate. Our own estimate from published factors. The working is shown next to the number; treat it as an order of magnitude, not a measurement.— NoneNot disclosed. The company publishes nothing on this and no defensible estimate exists.

Where the compute lives

Named sites from company disclosures and press reporting

OperationalConstructionPlanned
  • San Francisco HQCalifornia, USA · Hyperforce control plane.
  • AWS us-east-1Virginia, USA
  • AWS FrankfurtGermany

Timeline

Incidents and developments involving this company

PositiveResearch·Feb 10, 2025

AI Energy Score launches star ratings for 166 models across 10 tasks

Hugging Face, Salesforce, Cohere and Carnegie Mellon publish an Energy Star-style leaderboard: GPU watt-hours per task, with five relative star bands. Open models are tested automatically; closed models through a sealed sandbox.

Salesforce, Hugging Face, Cohere, Meta, Mistral, Alibaba, DeepSeek, IBM