2023

Environmental
Data Index

2023 Meta Sustainability Report 1

 Forward looking statements

This report covers only Meta’s accepted accounting principles, topics that reflect Meta’s This report contains forward- statements largely on our Any one of those factors,
business and does not address continue to evolve, and may be significant economic, social looking statements. All current expectations and including as the result of changes
the performance or operations based on assumptions believed and environmental impacts or statements contained in this projections about future events in circumstances, estimates
of our suppliers, contractors or to be reasonable at the time of that substantially influence the report other than statements and trends that we believe may that turn out to be incorrect,
partners. Statements regarding preparation but may be subject to assessments and decisions of a of historical fact, including affect our financial condition, standards of measurement that
targets, goals and commitments revision. This report has not been diverse set of stakeholders. We statements regarding our future results of operations, business change over time, assumptions
are aspirational and may also externally assured or verified by are not using these terms as they results of operations and financial strategy, short-term and long- not being realized, or other risks
be based on estimates and an independent third party unless are used under the securities or position, our business strategy term business operations and or uncertainties, could cause
assumptions under developing otherwise noted. This report other laws of the United States or and plans, and our objectives objectives, and financial needs. our actual results, including the
standards that may change in the represents Meta’s current policy any other jurisdiction or as these for future operations, as well achievement of targets, goals or
Especially with respect to the
future. As such, no guarantees and intent and is not intended to terms are used in the context statements regarding targets, commitments, to differ materially
matters discussed in this report,
or promises are made that create legal rights or obligations. of financial statements and goals and commitments, are from those expressed or implied
many factors and uncertainties
they will be met or successfully financial reporting. This report forward-looking statements. in writing in any forward-looking
In this report, our use of the relating to our operations and
executed, and actual results is not comprehensive, and for The words “believe,” “may,” statements made by Meta or on
terms “material,” “materiality” business environment, all of
may differ, possibly materially. that reason, should be read in “will,” “estimate,” “continue,” its behalf.
and other similar terms is which are difficult to predict and
In addition, data, statistics and conjunction with our most recent “anticipate,” “intend,” “expect,”
consistent with that of GRI, many of which are outside of our
metrics included in this report Annual Report on Form 10-K, our and similar expressions are
SASB, TCFD and other standards control, influence whether any
are non-audited estimates, subsequent reports on Forms intended to identify forward-
referenced in the preparation forward-looking statements can
not necessarily prepared in 10-Q and 8-K and other filings looking statements. We have
of this report, or refers to or will be achieved.
accordance with generally made with the Securities and based these forward-looking
Exchange Commission (SEC).

2023 Meta Sustainability Report 2

 Forward looking statements

We describe these risks and We also caution that the Such links or references are not
uncertainties in our SEC filings, important factors referenced incorporated by reference to this
including our most recent therein may not include all of report, and we can provide no
Annual Report on Form 10-K the factors that are important assurance as to their accuracy.
and our subsequent reports on to readers. Our forward-looking The use or inclusion of the
Forms 10-Q and 8-K, as well as, statements speak only as of information is also not intended
with respect to targets, goals the date of this report or as of to represent endorsements of
and commitments outlined in the date they are made, and we any apps and services.
this report or elsewhere, the undertake no obligation to update
challenges and assumptions this report to reflect subsequent
that are either identified in this events or circumstances, except
report or that we are unable to as required by law. Given these
foresee at this time. We cannot risks and uncertainties, readers
assure that the results reflected are cautioned not to place undue
or implied by any forward-looking reliance on such forward-looking
statement will be realized or, statements.
even if substantially realized,
This report may contain links
that those results will have
to other internet sites or
the forecasted or expected
references to third parties.
consequences and effects.

2023 Meta Sustainability Report 3

 Environmental footprint

1.1 GHG emissions 1,2,3,4,5 Greenhouse gas intensity

Total GHG emissions Market-based Scope 1 & 2 emissions (in metric tons CO2e/unit of key performance indicators)

Market-based (in metric tons CO2e) 2017 2018 2019 2020 2021 2022

2017 2018 2019 2020 2021 2022 GHG intensity per

0.00029 0.00015 0.00008 0.00001 0.00002 0.00002
monthly active person
Net total 1,096,000 1,008,000 4,330,000 4,984,000 5,740,244 8,453,471
GHG intensity per
million USD of revenue
- - - - 0.49 0.58
Carbon removal (carbon
- - - 145,000 90,000 80,000
credits applied) 7
GHG intensity per MWh - - - - 0.0061 0.0058

Total 1,096,000 1,008,000 4,330,000 5,129,000 5,830,244 8,533,471

1. Prior to 2021, values were rounded and totals were calculated before rounding throughout this report.
Scope 1 25,000 42,000 44,000 29,000 55,173 66,934
2. “Other data center-related facilities” includes facilities where Meta used less than 100,000 MWh of electricity in the reporting year, such as
Percent of total GHG warehouses or colocation facilities. Owned, online data centers are always reported by site, even if they were below this threshold.
2% 4% 1% 1% 1% 1%
emissions (Scopes 1-3) 3. Meta’s methodology for calculating greenhouse gas emissions can be found on page 15.
4. Prior to 2018, Scope 3 emissions included only business travel, employee commute and construction. Meta includes emissions from all relevant
Scope 2 591,000 314,000 208,000 9,000 2,487 273
categories in Scope 3 for reporting years 2019 to the present.
Percent of total GHG 5. In the 2022 reporting year, several updates to reporting were applied to the 2021 and later inventories.
54% 31% 5% <1% <1% <1%
emissions (Scopes 1-3)
(a) Data from life cycle assessments for our hardware and sold products were used to calculate our Scope 3 emissions.
(b) 2021 category 1, 2, 8, & 11 emissions were recalculated with higher quality data inputs to improve accuracy.
Scope 3 480,000 652,000 4,078,000 5,091,000 5,772,583 8,466,264
(c) All Scope 3 Categories were broken out individually to improve transparency and eliminate the previously reported “Other Applicable
Percent of total GHG Categories”
44% 65% 94% 99% 99% 99% (d) Emissions associated with third-party construction-related energy usage were recategorized into Category 1 instead of Category 3 to better
emissions (Scopes 1-3)
align with the GHG Protocol Scope 3 Category Boundaries.
Location-based (in metric tons CO2e) (e) Emissions associated with overhead electricity load at leased data centers was recategorized into Category 8 Instead of Category 3 to better
align with the GHG Protocol Scope 3 Category Boundaries.
2017 2018 2019 2020 2021 2022 (f) 2021 Category 6 emissions were recalculated to incorporate more accurate and transparent methodologies for applying sustainable aviation fuel
emissions reductions.
Total 1,387,000 1,983,000 6,295,000 8,559,000 10,163,476 14,007,222 (g) 2021 Total Fuel and Energy Consumption were recalculated to eliminate third-party party construction-related fuel use outside of Meta’s
Operational Control.

2023 Meta Sustainability Report 4

 Environmental footprint

Operational GHG emissions Market-based Scope 1 & 2 emissions (in metric tons CO2e)6 (Continued)

Market-based Scope 1 & 2 emissions (in metric tons CO2e)6 2017 2018 2019 2020 2021 2022

2017 2018 2019 2020 2021 2022 Prineville, OR 239,000 137,000 1,000 3,000 3,862 4,501

Total operational GHG Leased data center

616,000 356,000 252,000 38,000 57,661 67,207 98,000 102,000 188,000 - 25 72
emissions facilities

Data centers total 568,000 314,000 207,000 14,000 25,240 22,163 Other data center-
40,000 17,000 4,000 2,000 40 166
related facilities
Altoona, IA 1,000 1,000 2,000 1,000 2,118 920
Offices total 48,000 42,000 44,000 24,000 32,421 45,044

Clonee, Ireland <500 <500 <500 1,000 1,364 264

Dekalb, IL - - - - 0 1,859 6. In the 2019 reporting year, three updates to reporting were applied to 2017 (baseline year) and later inventories:

(a) Vehicles operated by the Transportation Team in support of commuting and inter-campus travel were previously counted in Scope 3 – Employee
Eagle Mountain, UT - - - - 3,250 3,609
commute. After re-visiting Meta’s operational control of these vehicles, it was determined that they should be accounted for in Scope 1.
(b) It was determined that Meta overestimated natural gas emissions by including estimates for offices that do not in fact use natural gas.
Forest City, NC 136,000 53,000 9,000 <500 1,401 587
Recalculations have been applied to the inventory to remove these inaccuracies.
(c) Fugitive emissions from refrigerant losses at offices not under Meta operational control were moved from Scope 2 to Scope 3.
Fort Worth, TX 1,000 1,000 1,000 <500 779 625

Gallatin, TN - - - - - 138

Richmond, VA - - <500 <500 4,822 821

Huntsville, AL - - - - 261 1,788

Los Lunas, NM - 1,000 1,000 <500 1,067 1,298

Luleå, Sweden <500 <500 <500 <500 374 79

New Albany, OH - - <500 2,000 408 2,605

Newton County, GA - - - - 300 535

Odense, Denmark - - <500 <500 2,824 655

Papillion, NE - <500 <500 3,000 2,348 1,642

2023 Meta Sustainability Report 5

 Environmental footprint

Market-based vs. Location-based

Scope 2 emissions (in metric tons CO2e)

2018 2019 2020 2021 2022

Market-based Location-based Market-based Location-based Market-based Location-based Market-based Location-based Market-based Location-based

Total facilities GHG emissions 314,000 1,241,000 205,000 1,885,000 9,000 2,718,000 2,487 3,080,194 273 3,921,611

Data centers total 308,000 1,181,000 197,000 1,813,000 2,000 2,650,000 2,487 2,987,964 273 3,821,450

Altoona, IA - 346,000 - 483,000 - 555,000 - 425,377 - 474,826

Clonee, Ireland - 82,000 - 143,000 - 159,000 - 187,475 - 178,367

Dekalb, IL - - - - - - - 2,122 - 8,087

Eagle Mountain, UT - - - - - - - 62,962 - 145,985

Forest City, NC 52,000 201,000 8,000 208,000 - 202,000 - 165,026 - 143,754

Fort Worth, TX - 212,000 - 295,000 - 399,000 - 378,198 - 355,696

Gallatin, TN - - - - - - - - - 2,664

Richmond, VA 137,000 - - 3,000 - 69,000 - 146,396 - 204,494

Huntsville, AL - - - - - - - 32,464 - 156,885

Los Lunas, NM - 12,000 - 135,000 - 266,000 - 276,795 - 347,033

Lueleå, Sweden - 7,000 - 6,000 - 7,000 - 3,917 - 2,782

New Albany, OH - - - 20,000 - 157,000 - 229,785 - 335,561

Newton County, GA - - - - - - - 84,402 - 258,773

Odense, Denmark - 1,000 <500 18,000 - 57,000 2,487 51,171 273 49,198

Papillion, NE - 3,000 - 101,000 - 294,000 - 329,674 - 458,460

2023 Meta Sustainability Report 6

 Environmental footprint

Scope 2 emissions (in metric tons CO2e) (Continued)

2018 2019 2020 2021 2022

Market-based Location-based Market-based Location-based Market-based Location-based Market-based Location-based Market-based Location-based

Prineville, OR - 145,000 - 167,000 - 200,000 - 245,996 - 284,462

Leased data center facilities 102,000 128,000 188,000 193,000 - 223,000 - 272,848 - 323,060

Other data center-related facilities 17,000 44,000 1,000 41,000 2,000 62,000 - 93,354 - 91,364

Offices total 6,000 60,000 8,000 72,000 7,000 68,000 - 92,230 - 100,160

2023 Meta Sustainability Report 7

 Environmental footprint

Value chain GHG emissions Scope 3 emissions (in Metric Tons CO2e) (Continued)

Scope 3 emissions (in Metric Tons CO2e) 1, 5, 7, 8 2017 2018 2019 2020 2021 2022

2017 2018 2019 2020 2021 2022 Category 9: Downstream

Transportation and - - 5,000 10,000 37 16
Total 480,000 652,000 4,078,000 5,091,000 5,772,583 8,466,264 Distribution 5

Category 1: Purchased Of Total (in %) - - <1% <1% <1% <1%

- - 1,428,000 1,846,000 2,956,909 2,545,466
Goods & Services 5, 8
Category 11: Use of Sold
- - 5,000 390,000 106,232 62,306
Of Total (in %) - - 35% 36% 51% 30% Products 5

Of Total (in %) - - <1% 8% 2% <1%

Category 2: Capital Goods 5, 8 - - 1,671,000 2,516,000 2,466,041 5,346,583

Category 12: End-of-Life

Of Total (in %) - - 41% 49% 43% 63%
Treatment of - - <500 <500 1,267 3,775
Sold Products 5
Category 3: Fuel & Energy-
- - 264,000 56,000 10,483 12,658
Related Activities 5
Of Total (in %) - - <1% <1% <1% <1%

Of Total (in %) - - 6% 1% <1% <1%

Category 4: Upstream 1. Prior to 2021, values were rounded and totals were calculated before rounding throughout this report.
Transportation and - - 65,000 49,000 180,183 176,636 5. In the 2022 reporting year, several updates to reporting were applied to the 2021 and later inventories.
Distribution
(a) Data from life cycle assessments for our hardware and sold products were used to calculate our Scope 3 emissions.
Of Total (in %) - - 2% 1% 3% 2% (b) 2021 Category 1, 2, 8, & 11 emissions were recalculated with higher quality data inputs to improve accuracy.
(c) All Scope 3 categories were broken out individually to improve transparency and eliminate the previously reported “Other Applicable
Category 5: Waste Categories.”
- - 4,000 10,000 18,430 18,519
Generated in Operations 5, 8 (d) Emissions associated with third-party construction-related energy usage were recategorized into Category 1 instead of Category 3 to better
align with the GHG Protocol Scope 3 Category Boundaries.
Of Total (in %) - - <1% <1% <1% <1% (e) Emissions associated with overhead electricity load at leased data centers was recategorized into Category 8 Instead of Category 3 to better
align with the GHG Protocol Scope 3 Category Boundaries.
Category 6: Business Travel
246,000 397,000 529,000 129,000 8,653 251,807 (f) 2021 Category 6 emissions were recalculated to incorporate more accurate and transparent methodologies for applying sustainable aviation fuel
5, 7
emissions reductions.
(g) 2021 Total Fuel and Energy Consumption were recalculated to eliminate third-party construction-related fuel use outside of Meta’s Operational
Of Total (in %) - - 13% 3% <1% 3%
Control.
Category 7: Employee 7. Sustainable Aviation Fuel was purchased in 2022 and associated emissions reductions are reflected in the inventory.
43,000 71,000 90,000 61,000 23,163 45,054
Commuting 8
8. In the 2022 reporting year, the following updates to the methodology were applied:
Of Total (in %) - - 2% 1% <1% <1% (a) A new Category 5 estimation methodology was developed to improve completeness across all operations.
(b) Employee commuting now includes emissions calculations on a well-to-tank basis.
Category 8: Upstream
- - 16,000 24,000 1,185 3,444 (c) a new Category 1 and Category 2 methodology was developed to improve the completeness, accuracy and reliability of the underlying activity
Leased Assets 5
and financial data.
Of Total (in %) - - <1% <1% <1% <1%

2023 Meta Sustainability Report 8

 Environmental footprint

2.1 Electricity Electricity consumption by facility (in MWh) (Continued)

Electricity consumption 2017 2018 2019 2020 2021 2022

Electricity consumption by facility (In MWh) Papillion, NE - 5,000 178,000 519,000 736,810 1,007,635

2017 2018 2019 2020 2021 2022 Prineville, OR 426,000 488,000 573,000 686,000 898,409 982,177

Total electricity Leased data center facilities 359,000 432,000 647,000 795,000 964,650 1,105,834
2,462,000 3,427,000 5,140,000 7,170,000 9,420,839 11,508,131
consumption
Other data center-related
Electricity from grid (%) 100% 100% 100% 100% 100% 100% 135,000 133,000 113,000 206,000 249,843 256,939
facilities

Data centers total 2,360,000 3,245,000 4,918,000 6,966,000 9,117,122 11,167,416 Offices Total 102,000 181,000 222,000 204,000 303,717 340,657

Altoona, IA 500,000 612,000 853,000 980,000 950,705 1,043,606 Electricity intensity (in MWh/unit of key performance indicators)

Clonee, Ireland 1,000 200,000 382,000 487,000 634,648 668,290 2017 2018 2019 2020 2021 2022

Dekalb, IL - - - - 4,724 16,934 Electricity intensity per

- - - - 0.0026 0.0031
monthly active person
Eagle Mountain, UT - - - - 229,946 504,049
Electricity intensity per
- - - - 79.9 98.7
million USD revenue
Forest City, NC 433,000 547,000 614,000 595,000 580,842 492,786
Electricity mix (in % of total electricity used)
Fort Worth, TX 189,000 461,000 695,000 941,000 1,014,447 959,419
2017 2018 2019 2020 2021 2022
Gallatin, TN - - - - 0 6,264
Renewable 51% 75% 86% 100% 100% 100%
Richmond, VA - - 10,000 204,000 515,270 701,003
Non-renewable 49% 25% 14% 0% 0% 0%
Huntsville, AL - - - - 85,286 368,841
2.2 Total energy consumed
Los Lunas, NM - 26,000 289,000 571,000 717,932 929,488
Energy consumption (in GJ)5
Luleå, Sweden 301,000 337,000 373,000 369,000 306,054 267,471
2017 2018 2019 2020 2021 2022
New Albany, OH - - 38,000 270,000 511,414 702,694
Total energy consumption - - - 27,075,000 34,882,163 42,560,221
Newton County, GA - - - - 215,279 636,266
Direct energy consumption - - - 438,000 853,042 1,138,794
Odense, Denmark - 4,000 128,000 343,000 500,863 517,718
Indirect energy
- - - 26,638,000 34,029,121 41,421,428
consumption

2023 Meta Sustainability Report 9

 Environmental footprint

2.3 Fuels 5. In the 2022 reporting year, several updates to reporting were applied to the 2021 and later inventories

(a) Data from life cycle assessments for our hardware and sold products were used to calculate our Scope 3 emissions.
Fuel consumption 5
(b) 2021 Category 1, 2, 8, & 11 emissions were recalculated with higher quality data inputs to improve accuracy.
(c) All Scope 3 categories were broken out individually to improve transparency and eliminate the previously reported “Other Applicable Categories”
2017 2018 2019 2020 2021 2022 (d) Emissions associated with 3rd party construction related energy usage were recategorized into Category 1 instead of Category 3 to better align
with the GHG Protocol Scope 3 Category Boundaries
Natural gas (therms) - - - - 6,153,856 7,539,592 (e) Emissions associated with overhead electricity load at leased data centers was recategorized into Category 8 Instead of Category 3 to better
align with the GHG Protocol Scope 3 Category Boundaries
Diesel — diesel fuel (gal) - - - - 363,082 1,376,871 (f) 2021 Category 6 emissions were recalculated to incorporate more accurate and transparent methodologies for applying sustainable aviation fuel
emissions reductions
Diesel — distillate fuel oil
- - - - 842,460 724,151 (g) 2021 Total Fuel and Energy Consumption were recalculated to eliminate 3rd party construction-related fuel use outside of Meta’s Operational
No.4 (gal)
Control

Gasoline (gal) - - - - 52,375 119,955

Propane (gal) - - - - 0 0

Renewable fuels

Hydrotreated vegetable oil

- - - - 0 0
(gal)

2.4 Data center operations and design

Power usage effectiveness (PUE)

2017 2018 2019 2020 2021 2022

PUE (data center energy

1.10 1.11 1.11 1.10 1.09 1.08
efficiency)

Sustainable design

Green building standards for data centers and offices (% of sq ft covered by green building standards and/or EnMS)

2017 2018 2019 2020 2021 2022

Total - - - - 98% 99%

Data centers (LEED Gold or

- - - - 100% 100%
above, or ISO 50001)

Offices (LEED Gold or

- - - - 97% 98%
above, or ISO 50001)

2023 Meta Sustainability Report 10

 Environmental footprint

3.1 Water withdrawal 9 Water withdrawal by facility (in cubic meters)

Water withdrawal 2017 2018 2019 2020 2021 2022

Water withdrawal by facility (in cubic meters) Leased data center facilities 473,000 533,000 1,011,000 645,000 603,629 772,921

2017 2018 2019 2020 2021 2022 Other data center-related

85,000 264,000 54,000 42,000 197 0
facilities
Total water withdrawal 1,609,000 2,367,000 3,430,000 3,726,000 5,042,564 4,893,023
Offices total 470,000 631,000 699,000 726,000 1,624,773 1,275,021

Data centers total 1,139,000 1,730,000 2,731,000 3,000,000 3,417,791 3,618,003

Water withdrawal by source
Altoona, IA 106,000 139,000 145,000 151,000 140,231 199,378
Water withdrawal by source (in cubic meters)
Clonee, Ireland 10,000 188,000 395,000 615,000 927,914 838,654
2017 2018 2019 2020 2021 2022

Dekalb, IL - - - - 0 29,659
Total water withdrawal 1,609,000 2,367,000 3,430,000 3,726,000 5,042,564 4,893,023

Eagle Mountain, UT - - - - 57,701 89,366

From surface water - - - - 0 0

Forest City, NC 129,000 99,000 85,000 68,000 64,053 62,853

From groundwater - - - 37,000 33,285 37,343

Fort Worth, TX 98,000 269,000 322,000 300,000 253,520 346,115

From seawater - - - - 0 0

Gallatin, TN - - - - 0 0
From produced water - - - - 0 0

Richmond, VA - - - 42,000 80,478 54,994

From third-party water (e.g.
- - - 3,689,000 5,009,279 4,855,680
municipal water supply)
Huntsville, AL - - - - 38,520 103,501
Water usage effectiveness (WUE)
Los Lunas, NM - 25,000 92,000 140,000 152,666 161,436
2017 2018 2019 2020 2021 2022
Luleå, Sweden 66,000 53,000 58,000 49,000 38,922 25,358
Annual data center WUE 0.24 0.27 0.27 0.30 0.26 0.20
New Albany, OH - - 33,000 35,000 121,194 87,413

Newton County, GA - - - - 105,121 77,203 9. Not included in Meta’s 2022 water withdrawal numbers are an additional 1,780,000 cubic meters of water withdrawn for the construction of Meta
data centers.
Odense, Denmark - - 266,000 360,000 373,355 427,937

Papillion, NE - - 62,000 108,000 106,339 100,912

Prineville, OR 172,000 160,000 208,000 445,000 353,951 240,302

2023 Meta Sustainability Report 11

 Environmental footprint

Water withdrawal intensity (in cubic meters/unit of key performance indicators) Water consumption from areas with water stress (in cubic meters)

2017 2018 2019 2020 2021 2022 2017 2018 2019 2020 2021 2022

Water withdrawal per Total water consumption 838,000 1,279,000 1,971,000 2,202,000 2,568,849 2,638,188
0.000755 0.001020 0.001200 0.001130 0.001405 0.001308
monthly active person
From areas with high or
Water withdrawal per
- - - - 42.8 42.0 extremely high baseline - - - - 162,477 443,150
million USD revenue
water stress

Water withdrawal from areas with water stress (in cubic meters) From areas without water
- - - - 2,406,372 2,195,038
stress
2017 2018 2019 2020 2021 2022
3.3 Water discharge
Total water withdrawal 1,609,000 2,367,000 3,430,000 3,726,000 5,042,564 4,893,023
Water discharge by source (in cubic meters)
From areas with high or
extremely high baseline - - - - 1,390,166 1,130,181 2017 2018 2019 2020 2021 2022
water stress
Total water discharge - - - 1,524,000 2,473,716 2,254,835
From areas without water
- - - - 3,652,398 3,762,843
stress
To surface water - - - - 0 0
Recycled water (in cubic meters)
To groundwater - - - - 0 0
2017 2018 2019 2020 2021 2022
To seawater - - - - 0 0
Total water recycled 469,000 673,000 854,000 643,000 580,223 265,906
To third-party water (e.g.
- - - 1,524,000 2,473,716 2,254,835
municipal sewers)
3.2 Water consumption
Water discharge to areas with water stress (in cubic meters)
Water consumption (in cubic meters)
2017 2018 2019 2020 2021 2022
2017 2018 2019 2020 2021 2022

Total water discharge - - - 1,524,000 2,473,716 2,254,835

Total water consumption 838,000 1,279,000 1,971,000 2,202,000 2,568,849 2,638,188

To areas with water stress - - - - 863,836 687,031

Data centers total - - - 2,197,000 162,477 2,510,686
To areas without water
Offices total - - - 73,000 2,406,372 127,502 - - - - 1,609,879 1,567,804
stress

2023 Meta Sustainability Report 12

 Environmental footprint

3.4 Water stewardship

Water restoration (in cubic meters)

2017 2018 2019 2020 2021 2022

Volumetric water restoration

- 132,000 145,000 2,250,000 2,335,672 2,351,562
benefits

Progress on 2030 net positive water goal (in cubic meters)

2017 2018 2019 2020 2021 2022

Total water consumption 838,000 1,279,000 1,971,000 2,202,000 2,569,000 2,638,000

Total water restored - 132,000 145,000 2,250,000 2,335,672 2,351,562

Water use embedded in purchased electricity (In cubic meters)

2017 2018 2019 2020 2021 2022

Embedded consumption
in purchased electricity - - - - - 31,923,969 41,172,356
location-based

Embedded consumption
in purchased electricity - - - - - 3,312,616 2,894,787
market-based

Avoided water consumption - - - - 28,611,342 38,277,569

2023 Meta Sustainability Report 13

 Environmental methodology

At Meta, our sustainability work helps us to operate efficiently and responsibly in our mission to build
SCOPE 1 EMISSIONS • Stationary combustion (e.g., natural gas consumed at our Menlo
community and bring the world closer together. As a global company, we recognize the tech industry’s Direct emissions from our Park campus for heating)
environmental impact and role to play in addressing climate change. We embrace the responsibility to data centers, offices and
transportation fleet • Mobile combustion (e.g., diesel emissions from our intercampus
understand the full scope of our footprint and be transparent and accountable in our mission to reduce shuttles)
our emissions.
• Fugitive emissions (e.g., refrigerant losses)
Identifying the source of our emissions on an annual basis enables us to prioritize emissions reduction SCOPE 2 EMISSIONS • Purchased electricity
Indirect emissions from
where we can make the most meaningful progress on our path to net zero emissions across our value • District heating
purchased energy for our data
chain in 2030. Similarly, minimizing our water use, being transparent with our water data, and restoring centers and offices • Stationary combustion from leased sites
water in the same watersheds where our data centers are located are vital to reach our commitment to
restore more water than we use by 2030.
In 2020, Meta reduced our operational emissions by 94% from a 2017 baseline and addressed the residual
emissions with high-quality carbon removal projects. As a result, Meta’s operations have produced net zero
Meta’s GHG emissions
emissions since then.
Meta’s GHG footprint includes the emissions associated with running our business and data centers, as
well as the indirect emissions upstream and downstream of our operations. These emissions correspond FULL VALUE CHAIN EMISSIONS

to Scope 1, Scope 2 and Scope 3 emissions as defined by the World Resources Institute (WRI) Greenhouse Scope 3 emissions come from sources within our full value chain beyond our operations and comprise the

Gas Protocol↗. Meta uses the operational control approach when calculating our GHG footprint, in which largest component of our footprint. Scope 3 includes:

we account for 100% of the GHG emissions over which we have operational control.
1. Upstream emissions, such as the emissions from manufacturing our data center servers or emissions
from employee commutes; and
OPERATIONAL EMISSIONS
Scope 1 and 2 emissions are considered our operational emissions. Scope 1 emissions come from our direct 2. Downstream emissions, such as the emissions associated with people using our Portal or Quest devices.
operations, such as combustion of natural gas to heat our offices and the fuel burned in our employee
shuttles. Scope 2 includes indirect emissions from purchased energy, such as the electricity powering our
data centers. We consider purchased electricity for construction and overhead electricity within leased
data centers outside of our operational control and therefore report these in Scope 3.

2023 Meta Sustainability Report 14

 Environmental methodology

SCOPE 3 EMISSIONS Upstream: How we calculate our GHG emissions

Our value chain emissions
upstream and downstream of • Purchased goods and services (e.g., upstream emissions from Meta is aligning our emissions reduction targets with the Science Based Targets initiative↗ and takes a
our operations purchased office supplies) scientific, standardized approach to calculating its GHG emissions in accordance with the GHG Protocol↗.
• Capital goods (e.g., server hardware) Furthermore, Meta’s GHG emissions data and methodologies undergo third party verification each year.
This is completed annually to ensure that only the most accurate and up-to-date data is publicly reported.
• Fuel and energy-related activities

• Upstream transportation and distribution (e.g., emissions We quantify our GHG emissions via activity data, LCAs and financial data. We prioritize calculating our
associated with the transportation of AR/VR-related consumer emissions through activity data that directly measures an activity that results in GHG emissions, such
hardware)
as kilowatt hours (kWh) of electricity. Due to the complex nature of our business and value chain, we use
• Waste generated from our operations other methods to help calculate our emissions when activity data is not available.
• Business travel

• Employee commuting (including telecommuting) We measure our emissions by metric tons of carbon dioxide equivalent,
• Upstream leased assets (Including leased data center overhead
or CO2e, units. CO2e is used to standardize the emissions from different
electricity use) GHGs based on their global warming potentials.

Downstream:
ACTIVITY DATA
• Downstream transportation and distribution For activity data, we take the quantity of a specific measured activity and multiply it by an associated
• Direct use of our AR/VR-related consumer hardware emissions factor to calculate the total emissions from that activity. For example, the kWh of electricity
• End-of-life treatment of our AR/VR-related consumer hardware consumed at a Meta site is multiplied by the appropriate country-specific or regional-specific, publicly
available emissions factor to calculate the total emissions from that site’s electricity use. We use activity
data to calculate:

• Scope 1 and 2 emissions • Business travel (including radiative forcing)

• Fuel and energy-related activities • Employee commuting

• Waste generated in operations • Direct use of our AR/VR-related consumer

hardware
• Upstream transportation and distribution where
supplier specific data is available

2023 Meta Sustainability Report 15

 Environmental methodology

Where activity data is incomplete or unavailable for an operation that results in GHG emissions, existing MARKET-BASED INSTRUMENTS
activity data is used as a proxy to estimate these emissions. This ensures we are reporting a complete We have publicly committed to supporting its global operations with 100% renewable energy. We procure
GHG inventory across all of our operations. For example, the weight of waste at several Meta sites is used and retire one Energy Attribute Certificate (EAC) for every MWh of electricity used to power our global
as a proxy to estimate waste at other sites in the same region that do not have final waste weight data.
operations. Meta also procures and retires one EAC for every MWh of electricity use in select Scope

LCAs 3 categories.A Additionally, Meta procures Sustainable Aviation Fuel (SAF) and applies the associated

To understand cradle-to-gate emissions and/or upstream emissions that are released before certain assets emissions reductions from SAF allocated in the reporting year as a market-based instrument to Category

are used (e.g., the emissions released from the production of concrete before it is poured), we conduct 6: Business Travel.
third-party LCA studies or utilize LCA tools to measure our impact. This is applicable in our 2022 inventory
A core focus of Meta’s renewable energy program is adding new renewable energy projects to the
for the following emissions:
electricity grids that support our data centers to drive the transition to renewable energy in our
• Upstream emissions associated with the materials used in the construction of our data centers communities. In alignment with these principles, Meta adheres to the following EAC market boundaries:
• Upstream emissions of materials in office renovations and new construction
1. Owned data centersB: EACs from the same grid regionC
• Cradle-to-gate emissions of our augmented and virtual reality related consumer hardware, such as
Portal and Quest devices 2. Leased data centersD: EACs from the same grid region or same geographic regionE

• Cradle-to-gate emissions in key data center hardware components, such as hard drives 3. Other Scope 2 loads (offices, points-of-presence): EACs from same grid region or same geographic region

• End-of-life treatment of our AR/VR-related consumer hardware 4. Scope 3 loads: EACs from same grid region; once exhausted, EACs from same geographic region

Meta’s methodology aligns with the market boundaries set forth by the GHG Protocol for over 95% of our
FINANCIAL Scope 2 emissions, including for all Scope 2 emissions from our owned data centers. A small portion of our
Our Environmentally Extended Input Output (EEIO) method utilizes financial spend data and applies Scope 2 emissions are not covered by EACs within the GHG Protocol’s market boundaries set forth, but
industry-specific emission factors (e.g., kg CO2e per dollar spent on electronic manufacturing) published by are instead covered by EACs from within the same geographic region.
the U.S. Environmental Protection Agency (EPA)↗ to calculate “cradle-to-gate” emissions. We apply the
A. This includes data center construction in Category 1: Purchased Goods & Services, transmission and distribution loss in Category 3: Fuel & Energy
EEIO method to the following: Related Activities, employee work from home in Category 7: Employee Commuting, leased data center overhead electricity use in Category 8:
Upstream Leased Assets, and United States-based electricity consumption from our products in Category 11: Use of Sold Products.
• Purchased goods and services
B. Owned data centers include all completed data centers owned and operated by Meta. Data center loads while under construction are treated in line
with leased data centers.
• Capital goods not related to data center and office construction, AR/VR-related consumer hardware,
and key data center hardware components C. Grid Regions: WECC, ERCOT, MISO/SPP, PJM/NC, SERC, Nordpool (Europe), Singapore/Southeast Asia

D. For reporting year 2022, all leased data center load was in the United States and covered by EACs generated in-country.
• Upstream transportation and distribution where supplier specific data is unavailable
E. Geographic Regions: Americas (AMER); Europe, Middle East, and Africa (EMEA); Asia Pacific (APAC)

• Upstream leased assets

2023 Meta Sustainability Report 16

 Environmental methodology

Improving our GHG methodology Meta’s water withdrawal

As Meta decarbonizes our value chain over the next decade, the data and methodology that drives our The water that we use in our offices and at our data centers are withdrawn from our local water utilities
climate work will evolve and improve each year. We have disclosed our Scope 1 and 2 emissions for the or local aquifers. We report our water withdrawals based on data from our local water utilities or meter
last decade. We began reporting on some Scope 3 categories in 2015 and have reported on every relevant data, where available. We also report our water withdrawal during construction, based on reported data
category defined by the GHG Protocol since 2019. As techniques to calculate our emissions improve, we from our construction partners. Not included in Meta’s 2022 operational water withdrawal numbers are an
will apply those methods to previous years to refine our GHG footprint. For example, in 2020 we used the additional 1,780,000 cubic meters of water withdrawn for the construction of Meta data centers.
EPA’s updated EEIO emission factors for our Scope 3 calculations and updated our 2019 data accordingly.
Meta’s water consumption
Going forward, we will focus on increasing accuracy and granularity of our data. For example, we re-
baselined our 2020 data based on updated LCA data for key data center hardware and our AR/VR-related For our data centers, we determine our water consumption via two methods:
consumer hardware. We will use activity data for more emissions categories as methods to do so become
1. Calculating the difference between water withdrawal and wastewater discharge
available. We will continue reporting and updating our emissions boundaries as our business grows on our
path to net zero emissions. 2. Calculating consumption based on cycles of concentration from our cooling systems

For our offices, we estimate our water consumption based on industry averages. All of our wastewater is
PUE/WUE
discharged to local wastewater facilities.

Each year, we calculate the Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE) of our
Water risk
data centers. PUE measures how efficiently our data centers consume the energy to operate our servers
and network infrastructure. It is calculated by dividing the energy consumed at the data center by IT We use water stress metrics in the WRI’s Aqueduct tool↗ to conduct initial assessments of our water
electricity load. The closer our annual PUE is to “1” indicates how efficient our data centers are designed to risks. When appropriate, we increase the level of water risk based on additional local knowledge.
consume electricity.

Annual WUE is calculated by dividing our water withdrawal, in liters, by IT electricity load, in kWh. The
closer WUE is to “0”, the more efficient consumption of water to cool our IT-related infrastructure.

These metrics are calculated based on best available data, including internal meters, design estimates, and
utility bills where applicable.

2023 Meta Sustainability Report 17