Environmental Performance Indicators
AMD collects environmental data from AMD locations worldwide as shown on the following table including AMD energy use, electricity consumption, greenhouse gas (GHG) emissions, water use, waste generation and compliance. In addition, AMD collects GHG emissions data for direct manufacturing suppliers representing approximately 85% of total spend and over 70% of total Scope 3 Category 1 emissions. For the remaining manufacturing and non-manufacturing suppliers, AMD uses spend-based estimates derived from CEDA factors. AMD also discloses data for other applicable Scope 3 categories, including total energy consumption outside the organization, water and waste metrics for contract wafer manufacturing and total product weight metrics.
We obtained external limited-level assurance based on the International Standard on Assurance Engagements (ISAE) 3000 Revised for selected 2025 data, including our Scope 1 and 2 GHG emissions and performance against the related goal; Scope 3 GHG emissions for business travel (Category 6); and CR goals, excluding the 2030 goal related to supply chain carbon intensity. We also obtained external limited-level assurance for 2024 base-year data on Scope 3 GHG emissions from purchased goods and services (Category 1). We plan to obtain assurance for 2025 Category 1 data and performance against the related carbon intensity goal after completing supplier validation.
Beginning in fiscal year 2025, AMD reports environmental data on a fiscal-year basis to align with regulatory reporting requirements and the AMD financial reporting calendar and support greater consistency across disclosures. Where fiscal-year data are not readily available, calendar-year data are used as a proxy. Because the AMD fiscal year ends on the last Saturday of December, AMD considers calendar-year data representative. Data for prior years were reported on a calendar-year basis from Jan. 1 through Dec. 31.
n/a = not available
Values shown in italics represent adjusted data and are different from values shown in previous Corporate Responsibility Reports.
Energy
Greenhouse Gas Emissions
Operational Emissions (Scope 1 and 2)6
Scope 3 Emissions
| 2022 | 2023 | 2024 | 2025 | |
| Estimated Scope 3 Emissions (mtCO2e)10 | 26,401,535 | 23,126,774 | 24,622,646 | 24,078,447 |
| Category 1: Purchased goods and services | 4,019,620 | 3,909,347 | 2,669,26611 | 3,369,70712 |
| Category 3: Fuel-and energy-related activities (not in Scope 1+2) | 19,750 | 17,004 | 12,346 | 22,560 |
| Category 4: Upstream transportation and distribution | 80,947 | 90,702 | 102,409 | 85,663 |
| Category 5: Waste generated in operations | 79 | 260 | 331 | 332 |
| Category 6: Business travel | 10,779 | 13,985 | 33,308 | 44,601 |
| Category 7: Employee commuting | 33,247 | 35,052 | 39,071 | 44,014 |
| Category 8: Upstream leased assets13 | n/a | n/a | n/a | 157 |
| Upstream Scope 3 Emissions | 4,164,422 | 4,066,350 | 2,856,731 | 3,567,034 |
| Category 11: Use of sold products | 22,236,48414 | 19,059,598 | 21,764,55715 | 20,502,075 |
| Category 12: End of Life treatment of sold products | 630 | 825 | 1,35816 | 2,09317 |
| Category 15: Investments | n/a | n/a | n/a | 7,244 |
| Downstream Scope 3 Emissions | 22,237,114 | 19,060,423 | 21,765,915 | 20,511,412 |
| Total Estimated Scope 3 Emissions Per Revenue (mtCO2e / USD $ Millions)1 | 1,116 | 959 | 955 | 695 |
Other Emissions
| 2022 | 2023 | 2024 | 2025 | |
| NOx Emissions (metric tons) | n/a | 26 | 29 | 22 |
| SOx Emissions (metric tons) | n/a | 0.6 | 0.7 | 0.2 |
| GHG-HCFC-22 (metric tons) | n/a | 384 | 122 | 32 |
| F-Gas Emissions (HFEs) (metric tons)18 | n/a | n/a | n/a | 126 |
Water
| 2022 | 2023 | 2024 | 2025 | |
| Total Water Withdrawal (million liters) | 183 | 225 | 267 | 266 |
| Austin | 5 | 14 | 38 | 38 |
| Bengaluru | 3 | 5 | 5 | 6 |
| Dublin | 9 | 13 | 10 | 2 |
| Hyderabad | 8 | 9 | 10 | 11 |
| Hyderabad 2 (former Xilinx site) | 4 | 11 | 11 | 12 |
| Longmont | 72 | 46 | 76 | 65 |
| Markham | 19 | 40 | 26 | 28 |
| San Jose | 20 | 26 | 26 | 25 |
| Santa Clara | 12 | 18 | 21 | 33 |
| Shanghai | 14 | 18 | 19 | 19 |
| Singapore | 10 | 9 | 8 | 8 |
| Singapore 2 (former Xilinx site) | 6 | 8 | 9 | 8 |
| Other sites combined | 1 | 9 | 8 | 11 |
| Total Water Discharge (million liters) | n/a | n/a | n/a | 234 |
| Total water Consumption (million liters) | n/a | n/a | n/a | 32 |
| Water Withdraw/Revenue (million liters/$USD millions)1 | 0.008 | 0.010 | 0.010 | 0.010 |
| Contract Wafer Manufacturing Water use / Revenue (million liters/USD millions)1 |
0.54 | 0.55 | 0.3¹¹ | 0.26 |
Waste
|
2022 |
2023 |
2024 |
2025 |
Total Non-Hazardous Waste (NHW) Generated (metric tons) |
785 |
1,268 |
1,455 |
1,453 |
Atlanta |
n/a |
2019 |
6 |
10 |
Austin |
182 |
190 |
186 |
193 |
Bengaluru |
21 |
21 |
49 |
63 |
Dublin |
29 |
45 |
59 |
64 |
Hyderabad |
15 |
46 |
115 |
107 |
Longmont |
55 |
57 |
57 |
47 |
Markham |
211 |
468 |
526 |
538 |
San Jose |
180 |
285 |
303 |
315 |
Santa Clara |
37 |
34 |
59 |
49 |
Singapore 1 |
31 |
38 |
54 |
20 |
Singapore 2 (former Xilinx site) |
15 |
25 |
21 |
20 |
Other sites combined |
2 |
39 |
21 |
27 |
NHW Recycled (metric tons) |
622 |
755 |
784 |
788 |
NHW Landfilled (metric tons) |
135 |
461 |
596 |
600 |
NHW Incinerated (metric tons) |
28 |
29 |
34 |
25 |
NHW Treated (metric tons) |
n/a |
22 |
41 |
41 |
NHW Landfill Diversion Rate (%) |
83% |
63% |
57% |
56% |
Total Hazardous Waste (HW) Generated (metric tons) |
24 |
80 |
108 |
93 |
Austin |
2 |
5 |
7 |
3 |
Bengaluru |
n/a |
1 |
7 |
25 |
Hyderabad |
n/a |
20 |
20 |
11 |
Hyderabad 2 (former Xilinx site) |
12 |
2 |
12 |
3 |
Markham |
<1 |
1 |
1 |
2 |
Santa Clara |
2 |
<1 |
1 |
<1 |
Singapore 1 |
8 |
50 |
55 |
44 |
Singapore 2 (former Xilinx site) |
<1 |
1 |
1 |
1 |
Other sites combined |
<1 |
<1 |
6 |
3 |
HW Recycled/Reused (metric tons) |
12 |
24 |
33 |
40 |
HW Treated Off-Site (metric tons) |
8 |
50 |
56 |
45 |
HW Incinerated (metric tons) |
2 |
3 |
5 |
2 |
HW Landfilled (metric tons) |
2 |
3 |
5 |
2 |
HW Fuel recovery (metric tons) |
n/a |
<1 |
10 |
4 |
Total Waste Generated (NHW+HW) (metric tons) |
809 |
1,348 |
1,564 |
1,545 |
Total Waste Generated per Revenue (metric tons/$USD millions) |
0.03 |
0.06 |
0.06 |
0.04 |
Contract Wafer Manufacturing HW Generated (metric tons) |
47,018 |
44,292 |
60,025 |
85,83920 |
Contract Wafer Manufacturing HW Generated per Revenue (metric tons / $USD millions)1 |
1.99 |
1.95 |
1.76¹¹ |
1.77 |
n/a |
97% |
100% |
100% |
Wastewater
| 2022 | 2023 | 2024 | 2025 | |
| Total Wastewater Discharge21 (million liters) | 16 | 18 | 37 | 45 |
| Austin | 3 | 7 | 25 | 29 |
| Longmont | 12 | 11 | 12 | 16 |
| Wastewater generated per Revenue (million liters / $USD Millions)1 | 0.0007 | 0.0008 | 0.0015 | 0.0013 |
Compliance in AMD Operations
| 2022 | 2023 | 2024 | 2025 | |
| Number of Environmental Non-Compliances | 0 | 122 | 223 | 224 |
| Fines (USD) | 0 | $8,072 | $3,000 | $6,000 |
Products
| 2024 | 2025 | |
| Total Product Materials Used by Weight (metric tons)25 | 5,998 | 6,046 |
| Total Returned Product Materials by weight repaired or refurbished (metric tons)26 | 107 | 91 |
Global Workforce Data
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| Total Workforce27 | 34,648 | 35,294 | 39,691 | 45,842 |
| Total Employees | 24,525 | 25,768 | 28,343 | 31,359 |
| Temporary Workers and Contractors | 10,123 | 9,526 | 11,348 | 14,483 |
| Employees By Region | ||||
| Americas | 11,989 | 12,590 | 13,774 | 15,062 |
| % in Americas | 49% | 49% | 49% | 48% |
| EMEA | 1,478 | 1,546 | 2,003 | 2,288 |
| % in EMEA | 6% | 6% | 7% | 7% |
| Asia-Pacific/China/Japan | 11,058 | 11,632 | 12,566 | 14,009 |
| % in Asia-Pacific/China/Japan | 45% | 45% | 44% | 45% |
| Temporary Workers and Contractors by Region | ||||
| Americas | 2,271 | 1,977 | 2,378 | 2,978 |
| % in Americas | 22% | 20% | 21% | 21% |
| EMEA | 613 | 721 | 739 | 1,008 |
| % in EMEA | 6% | 8% | 6.5% | 7% |
| Asia-Pacific/China/Japan | 7,239 | 6,828 | 8,231 | 10,497 |
| % in Asia-Pacific/China/Japan | 72% | 72% | 72.5% | 72% |
| Employees by Gender | ||||
| Male | 18,744 | 19,703 | 21,737 | 24,105 |
| % of male | 76% | 76% | 77% | 77% |
| Female | 5,780 | 6,064 | 6,605 | 7,253 |
| % of female | 24% | 24% | 23% | 23% |
| Genderqueer28 | 1 | 1 | 1 | 1 |
| % of genderqueer | <1% | <1% | <1% | <1% |
| Employees by Age Group | ||||
| iGen (born 1997 or later) | 1,648 | 2,346 | 3,305 | 4,498 |
| % of iGen | 7% | 9% | 12% | 14% |
| Millennials (born 1981-1996) | 13,336 | 13,973 | 15,402 | 17,152 |
| % of Millennials | 54% | 54% | 54% | 55% |
| Generation X (born 1965-1980) | 7,930 | 7,967 | 8,255 | 8,503 |
| % of GenX | 32% | 31% | 29% | 27% |
| Baby Boomers (born 1946-1964) | 1,610 | 1,481 | 1,381 | 1,206 |
| % of Baby Boomers | 7% | 6% | 5% | 4% |
| Traditionalists (born 1927-1945) | 1 | 1 | 0 | 0 |
| % of Traditionalist | <1% | <1% | 0% | 0% |
| Employees by Work Status | ||||
| Full-Time Employees | 24,475 | 25,726 | 28,295 | 31,317 |
| % of FTE | 99.8% | 99.8% | 99.8% | 99.9% |
| Part-Time Employees | 50 | 42 | 48 | 42 |
| % of PTE | <1% | <1% | <1% | <1% |
Employee Category by Gender
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| AMD Executive Team29 | ||||
| Male | 13 | 19 | 17 | 19 |
| % of male | 81% | 83% | 77% | 79% |
| Female | 3 | 4 | 5 | 5 |
| % of female | 19% | 17% | 23% | 21% |
| Genderqueer | 0 | 0 | 0 | 0 |
| % of genderqueer | 0% | 0% | 0% | 0% |
| Senior Management30 | ||||
| Male | 1,217 | 1,336 | 1,416 | 1,511 |
| % of male | 86% | 86% | 86% | 85% |
| Female | 191 | 219 | 227 | 259 |
| % of female | 14% | 14% | 14% | 15% |
| Genderqueer | 0 | 0 | 0 | 0 |
| % of genderqueer | 0% | 0% | 0% | 0% |
| Engineering | ||||
| Male | 15,854 | 16,827 | 18,803 | 20,969 |
| % of male | 81% | 81% | 81% | 81% |
| Female | 3,701 | 3,980 | 4,452 | 4,972 |
| % of female | 19% | 19% | 19% | 19% |
| Genderqueer | 1 | 1 | 1 | 1 |
| % of genderqueer | <1% | <1% | <1% | <1% |
Diversity of Board of Directors
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| Total number of Board of Directors | 9 | 9 | 9 | 8 |
| Generation X (born 1965-1979) | ||||
| Total Male | 1 | 1 | 1 | 1 |
| % of male | 11% | 11% | 11% | 12.5% |
| Total Female | 1 | 1 | 1 | 1 |
| % of female | 11% | 11% | 11% | 12.5% |
| Baby Boomers (born 1946-1964) | ||||
| Total Male | 5 | 5 | 5 | 4 |
| % of male | 56% | 56% | 56% | 50% |
| Total Female | 2 | 2 | 2 | 2 |
| % of female | 22% | 22% | 22% | 25% |
Women in Management
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| % of Women in Management | n/a | 19% | 18% | 19% |
| % of Women in Middle Management | n/a | 20% | 19% | 20% |
| % of Women in Non-Managerial Positions | n/a | 25% | 24% | 24% |
New Employee Hires and Employee Turnover
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| Total New Employee Hires & Acquisitions | 11,278 | 3,058 | 4,474 | 5,638 |
| New Hires as % of Prior Year Employee Count | 73% | 12% | 17% | 20% |
| New Hires Distribution by Region | ||||
| Americas | 4,926 | 1,488 | 2,004 | 2,772 |
| % of Americas | 44% | 49% | 45% | 49% |
| EMEA | 1,199 | 268 | 608 | 448 |
| % of EMEA | 11% | 9% | 13% | 8% |
| Asia-Pacific/China/Japan | 5,153 | 1,302 | 1,862 | 2,418 |
| % of Asia-Pacific/China/Japan | 45% | 42% | 42% | 43% |
| New Hires by Gender | ||||
| Male | 8,757 | 2,362 | 3,453 | 4,414 |
| % of male | 78% | 77% | 77% | 78% |
| Female | 2,520 | 696 | 1,021 | 1,224 |
| % of female | 22% | 23% | 23% | 22% |
| Genderqueer | 1 | 0 | 0 | 0 |
| % of genderqueer | <1% | 0% | 0% | 0% |
| New Hires by Age Group | ||||
| iGen (born 1997 or later) | 1,225 | 857 | 1,182 | 1,544 |
| % of iGen | 11% | 28% | 26% | 27% |
| Millennials (born 1981-1996) | 6,269 | 1,581 | 2,446 | 3,006 |
| % of Millennials | 56% | 52% | 55% | 53% |
| Generation X (born 1965-1980) | 3,159 | 542 | 769 | 982 |
| % of GenX | 28% | 18% | 17% | 18% |
| Baby Boomers (born 1946-1964) | 625 | 77 | 77 | 106 |
| % of Baby Boomers | 5% | 2% | 2% | 2% |
| Traditionalists (born 1927-1945) | 0 | 1 | 0 | 0 |
| % of Traditionalist | 0% | <1% | 0% | 0% |
| Terminations | ||||
| Total Terminations | 2,251 | 1,805 | 1,892 | 2,619 |
| Total Turnover Rate (%)31 | 10.6% | 7.2% | 7.0 % | 8.9% |
| Voluntary Turnover32 Rate (%) | 9.5% | 4.1% | 4.3 % | 4.9% |
| Other Turnover Rate33 | 1.1% | 3.1% | 2.7 % | 4.0% |
| Male Voluntary Turnover Rate (%) | 9.6% | 4% | 4.2 % | 4.9% |
| Female Voluntary Turnover Rate (%) | 9.1% | 4.3% | 4.5 % | 4.8% |
| Turnover Rate by Gender | ||||
| Male | 1,740 | 1,395 | 1,430 | 2,050 |
| % of male | 10.7% | 7.3% | 6.9 % | 9.0% |
| Female | 511 | 410 | 462 | 569 |
| % of female | 10.1% | 6.9% | 7.3 % | 8.3% |
| Genderqueer | 0 | 0 | 0 | 0 |
| % of genderqueer | 0% | 0% | 0% | 0% |
| Turnover Rate by Region | ||||
| Total in Americas | 1,146 | 907 | 822 | 1,474 |
| Turnover rate (%) in Americas | 10.8% | 7.4% | 6.2% | 10.3% |
| Total in Asia-Pacific/China/Japan | 995 | 712 | 927 | 979 |
| Turnover rate (%) in Asia-Pacific/China/Japan | 10.4% | 6.3% | 7.6% | 7.4% |
| Total in Europe/Africa | 110 | 186 | 143 | 166 |
| Turnover rate (%) in Europe/Africa | 10% | 12.7% | 8.3 % | 7.8% |
| Total Turnover Rate by Age Group | ||||
| Total number of iGen (born 1997 or later) | 141 | 158 | 222 | 346 |
| Turnover rate (%) of iGen (born 1997 or later) | 12.3% | 7.9% | 7.7% | 9.0% |
| Total number of Millennials (born 1981-1996) | 1,426 | 945 | 1,019 | 1,269 |
| Turnover rate (%) of Millennials (born 1981-1996) | 12.3% | 7% | 6.9% | 7.8% |
| Total number of Generation X (born 1965-1980) | 533 | 498 | 475 | 726 |
| Turnover rate (%) of Generation X (born 1965-1980) | 7.6% | 6.3% | 5.8% | 8.8% |
| Total number of Baby Boomers (born 1946-1964) | 151 | 203 | 175 | 278 |
| Turnover rate (%) of Baby Boomers (born 1946-1964) | 10.2% | 13.1% | 12.1% | 22.4% |
| Total number of Traditionalists (born 1927-1945) | 0 | 1 | 1 | 0 |
Parental Leave34
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 | |
| Employees that were entitled to parental leave35 | |||||
| Male | 18,770 | 19,704 | 21,740 | 24,105 | |
| Female | 5,759 | 6,063 | 6,604 | 7,253 | |
| Genderqueer | 0 | 1 | 1 | 1 | |
| Total | 24,529 | 25,768 | 28,345 | 31,359 | |
| Employees that took parental leave | |||||
| Male | 549 | 950 | 1,025 | 1,068 | |
| Female | 304 | 485 | 519 | 581 | |
| Genderqueer | 0 | 0 | 0 | 0 | |
| Total | 853 | 1,435 | 1,544 | 1,649 | |
| Employees that returned to work in the reporting period after parental leave | |||||
| Male | 531 | 896 | 977 | 1,035 | |
| Female | 239 | 397 | 504 | 555 | |
| Genderqueer | 0 | 0 | 0 | 0 | |
| Total | 770 | 1,293 | 1,481 | 1,590 | |
| Employees that returned to work after parental leave ended that were still employed 12 months after their return to work | |||||
| Male | 507 | 880 | 947 | 1,010 | |
| Female | 229 | 391 | 486 | 544 | |
| Genderqueer | 0 | 0 | 0 | 0 | |
| Total | 736 | 1,271 | 1,433 | 1,554 | |
| Return to work rates of employees that took parental leave (%) | |||||
| Male return to work rate (%) | 97% | 94% | 95% | 97% | |
| Female return to work rate (%) | 79% | 82% | 97% | 95.5% | |
| Genderqueer return to work rate (%) | 0% | 0% | 0% | 0% | |
| Total return to work rate (%) | 90% | 90% | 96% | 96% | |
| Retention rates of employees that took parental leave (%) | |||||
| Male retention rate (%) | 92% | 93% | 92% | 95% | |
| Female retention rate (%) | 75% | 81% | 94% | 94% | |
| Genderqueer retention rate (%) | 0% | 0% | 0% | 0% | |
| Total retention rate (%) | 86% | 89% | 93% | 94% | |
Performance and Career Development36
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| Employees by gender who received a regular performance review during the reporting period. | ||||
| % of male | 100% | 100% | 100% | 100% |
| % of female | 100% | 100% | 100% | 100% |
| % of genderqueer | 100% | 100% | 100% | 100% |
| Employees by employee category37 who received a regular performance review during the reporting period. | ||||
| % of Executive | 100% | 100% | 100% | 100% |
| % of Managers | 100% | 100% | 100% | 100% |
| % of Technical Staff | 100% | 100% | 100% | 100% |
| % of Staff | 100% | 100% | 100% | 100% |
Minimum Wage38
| Annual Salary (USD) | 2022 | 2023 | 2024 | 2025 |
| Lowest U.S. entry-level wage | $35,292 | $38,468 | $40,103 | $36,474 |
Training and Education
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| Total hours of employee training | 178,944 | 156,561 | 166,994 | 244,297 |
| Average hours of training per year per employee | 7.3 | 6 | 5.9 | 7.8 |
| % of the total employees across all locations who received career or skills-related training39 | n/a | 26% | 23% | 23% |
| % of the total workforce across all locations who received training (internally or externally) on environmental issues40 | n/a | 67% | 75% | 89% |
Business Ethics
| 2025 | |
| Number of reported cases of business ethics related violations through internal or external whistleblower procedure | 355 |
| % of AMD employees that are current on their Global Code of Conduct Training | 99%41 |
AMDer Survey
| (Headcount, year-end) | 2022 | 2023 | 2024 | 2025 |
| % of employees that responded | 96% | 95% | 93% | 92% |
| I am proud to work for AMD (% favorable) | 92% | 91% | 91% | 94% |
| I would recommend AMD as a great place to work (% favorable) | 86% | 83% | 88% | 91% |
| AMD creates an environment where people of diverse backgrounds can succeed (% favorable) | 92% | 92% | 92% | 93% |
| Engagement Index (% favorable) | 86% | 86% | 86% | 89% |
| Belonging and Inclusion Index (% favorable) | 90% | 89% | 91% | 92% |
| Manager Quality Index (% favorable) | 89% | 88% | 87% | 88% |
Workforce Health and Safety42
| 2022 | 2023 | 2024 | 2025 | |
| Number of Recordable Employee Injury Cases | 2 | 11 | 10 | 18 |
| Rate of Employee Recordable Injury Cases (per 200,000 work hours) | n/a | n/a | 0.03 | 0.06 |
| Number of Recordable Contractor Injury Cases | 1 | 3 | 0 | 4 |
| Rate of Contractor Recordable Injury Cases (per 200,000 work hours) | n/a | n/a | 0.00 | 0.10 |
| Number of Recordable Injury Cases43 | 3 | 14 | 10 | 22 |
| Rate of Recordable Injury Cases (per 200,000 work hours) | 0.01 | 0.04 | 0.03 | 0.06 |
| Number of Recordable Illness Cases | 9 | 0 | 0 | 0 |
| Rate of Recordable Illness Cases (per 200,000 work hours) | 0.03 | 0 | 0 | 0 |
| Number of Recordable Combined Injury and Illness Cases | 12 | 14 | 10 | 22 |
| Rate of Recordable Combined Injury and Illness Cases (per 200,000 work hours) | 0.04 | 0.04 | 0.03 | 0.06 |
| Number of Employee Work-Related Fatalities | 0 | 0 | 0 | 0 |
| Number of Contractor Work-Related Fatalities | 0 | 0 | 0 | 0 |
| Number of Combined Employee and Contractor Work-Related Fatalities | 0 | 0 | 0 | 0 |
| Rate of Combined Employee and Contractor Work-related Fatalities (per 200,000 work hours) | 0 | 0 | 0 | 0 |
| Number of Employee High-Consequence Work-Related Injuries (excluding fatalities) | 0 | 0 | 0 | 0 |
| Number of Contractor High-Consequence Work-Related Injuries (excluding fatalities) | 0 | 0 | 0 | 0 |
| Rate of Combined Employee and Contractor High-Consequence Work-Related Injuries (excluding fatalities) (per 200,000 work hours) | 0 | 0 | 0 | 0 |
| Rate of Employee High-Consequence Work-Related Injuries (excluding fatalities) (per 200,000 work hours) | 0 | 0 | 0 | 0 |
| Rate of Contractor High-Consequence Work-Related Injuries (excluding fatalities) (per 200,000 work hours) | 0 | 0 | 0 | 0 |
| Number of Recordable Employee Illness Cases | 9 | 0 | 0 | 0 |
| Number of Recordable Contractor Illness Cases | 0 | 0 | 0 | 0 |
| Number of Days Lost Due to Employee Injuries | n/a | n/a | 33 | 44 |
| Lost Time Injury (LTI) Severity Rate for Contractors (per 1,000 work hours) | n/a | n/a | 0 | 0 |
| Number of Days Lost Due to Contractor Injuries | n/a | n/a | 0 | 6 |
| Number of Employee Hours Worked | n/a | n/a | 60,007,176 | 64,228,488 |
| Number of Contractor Hours Worked | n/a | n/a | 6,306,528 | 8,360,856 |
| Number of Employee Lost Time Events | n/a | n/a | 1 | 8 |
| Number of Contractor Lost Time Events | n/a | n/a | 0 | 1 |
| Total Lost Time Events | n/a | n/a | 1 | 9 |
| Employee Lost Time Injury (LTI) Frequency Rate (per 1,000,000 work hours) | n/a | n/a | 0.02 | 0.12 |
| Contractors Lost Time Injury (LTI) Frequency Rate (per 1,000,000 work hours) | n/a | n/a | 0 | 0.12 |
| Lost Time Injury (LTI) Severity Rate for Employees (per 1,000 work hours) | n/a | n/a | 0 | 0 |
| Total number of Hours Worked | 59,955,000 | 74,238,696 | 66,313,704 | 72,589,344 |
| Number of Health or Safety Non-Compliances | 144 | 345 | 146 | 147 |
| % of the total workforce across all locations represented in joint management-worker health & safety committees | n/a | 28% | 29.5% | 26% |
| Total amount of monetary losses as a result of legal proceedings associated with employee health and safety violations (USD) | n/a | $0 | $0 | $0 |
| % of AMD Managed Sites for which an EHS risk assessment has been conducted | n/a | 30% | 44% | 41% |
Volunteerism48
| 2022 | 2023 | 2024 | 2025 | |
| AMD Volunteers | 4,977 | 5,720 | 8,178 | 8,050 |
| AMD Volunteer Hours | 17,244 | 25,368 | 33,388 | 34,721 |
| Number of Volunteer Events | 162 | 288 | 358 | 326 |
| Number of Employee Donors | 1,888 | 4,167 | 6,424 | 12,391 |
Additional Workforce Disclosures
Gender Pay Gap Report - Xilinx Ireland Unlimited Company
AMD EEO-1 Reports:
WGEA Gender Equality Group Report - AMD Australia:
Economic Performance Indicators
The following tables provide information about our company’s annual financial data, social investments and U.S. political activities.
n/a = not available
Values shown in italics represent adjusted data and are different from values shown in previous Corporate Responsibility Reports.
Financial Data49
| 2022 | 2023 | 2024 | 2025 | |
| Total Revenue (millions USD) | $23,601 | $22,680 | $25,785 | $34,639 |
| Operating Expense (millions USD) | $9,339 | $10,059 | $10,825 | $13,458 |
Community Investments
Advanced Micro Devices, Inc. Political Action Committee (PAC)51
| 2022 | 2023 | 2024 | 2025 | |
| Total Contributions (USD) | $36,500 | $33,000 | $82,120 | $5,140 |
| Total Receipts (USD) | $62,465 | $6,000 | $37,500 | $12,500 |
Footnotes
- For financial intensity metrics, AMD reports the environmental metric and divides by the corresponding year annual revenue (in millions of USD). Energy use intensity is MWh of energy per $1M USD revenue. Emissions intensity is metric tCO2e per $1M USD in revenue, shown both for location-based and market-based scope 2 emissions. Water use and wastewater discharge intensity are reported as millions of liters of water used or discharged per $1M USD in revenue. Waste is reported as metric tons generated per $1M USD in revenue.
- Indirect energy consumption is comprised of purchased grid electricity, electricity generated from chilled water, on-site solar generation and green tariff electricity.
- Electricity consumption does not include chilled water; it includes purchased grid electricity, on-site solar generation and green tariff electricity
- AMD sources renewable energy through utility providers (green tariffs), and sellers of third-party certified renewable energy credits (RECs), as well as generates renewable energy onsite.
- San Jose has fuel cells that convert diesel into electricity. The emissions related to the fuel cells are reported under Scope 1. The site also generates renewable energy onsite, which is included in the site electricity use but assigned a zero emission factor.
- 2020 data include energy use and Scope 1 and 2 GHG emission data from the Xilinx and Pensando acquisitions in 2022 in order to reset the 2020 GHG goal baseline per the GHG Protocol. All 2022 and later year environmental metrics include Xilinx and Pensando data.
- AMD follows the GHG Protocol for Scope 1 emission estimates, the internationally recognized standard for the corporate accounting and reporting of GHG emissions. Scope 1 emissions factors estimated based on quantity of refrigerants and fuel consumed in each geography, including natural gas and refrigerants such as hydrofluorocarbons (HFCs). The scope is based on operational control (i.e., ability to implement operational policies) and the method includes Site Metrics Coordinators entering the monthly amount of fuel and chemicals use, by type, into AMD’s central database, and then applying the emission factors. Emission factors were obtained from three main sources: DESNZ 2025 (previously referred to as DEFRA or UK BEIS in previous documents), IPCC AR6 (6th Assessment Report) or US EPA Fluorinated GHG Report.) Values have undergone third-party limited assurance. Beginning 2025 AMD will account for Fluorinated ethers (FEs) outside of Scope 1 as they are not one of the eight gases or categories of gases required for inclusion in GHG inventories according to the GHG Protocol.
- AMD follows the GHG Protocol for Scope 2 emission estimates by multiplying the quantity of electricity consumed at each site by relevant emission factors. If electricity use data is not available, as for small offices, then an average value for U.S. office buildings is used for all AMD locations (16.9 kWh/sq ft) based on EIA CBECS results for the average administrative office, and the emission factor for the location is applied. AMD applies both the market-based and location-based methods for estimating scope 2 emissions. Market-based emission calculations are based on grid electricity use excluding renewable energy sourced through green tariffs and renewable energy credits (RECs) allocated to each site. Location-based emission calculations are based on the quantity of grid electricity used plus renewable energy sourced through utility “green tariff” programs. Emission factors for locations in the U.S. are based on eGRID and Green-e 2025 databases. Location-based electricity emission factors for Canada are provided in the 2025 Canada NIR report, specific to each Canadian province. Market-based emission factors for Europe were taken from the IEA 2025 emission factor database. Location-based emission factors for all countries other than the US and Canada were taken from the IEA 2025 emission factor database. 2020-25 values have undergone third-party limited assurance.
- Our goal is a 50 percent reduction in absolute Scope 1 and 2 GHG emissions from 2020-2030.
- Our value chain emissions are estimated following the guidance from the GHG Protocol. Annually we update emission factors as available by the referenced sources. Category 1: Emissions are calculated by directly surveying suppliers representing ~85% of our total supply chain spend (includes Foundry, final test/assembly, substrates and others). It includes their estimated allocations to AMD (typically using revenue-based accounting), at a factory-level where available, for their Scope 1 and 2 emissions, as well as upstream Scope 3 if available and third-party verified. For Foundries, we use a third-party lifecycle assessment to estimate upstream emissions. For other suppliers we apply a spend-based method by mapping spend categories to the associated Scope 3 CEDA emission factor. Emissions from all other vendors (including marketing, professional services, real estate, software providers, telecom and networking providers and other manufacturing services) are calculated using a spend-based method. Category 2: Calculations follow a spend-based method and are included in the disclosed figure in Category 1. Category 3: Emissions are calculated using fuel and electricity consumption data collected from our sites globally, and Well-to-tank (WTT) emission factors for natural gas and diesel were obtained from the DEFRA 2025 Conversion Factors database. Emission factors for transmission & distribution-related electricity losses, and electricity-related WTT generation and transmission and distribution, were obtained from the IEA Emission Factors 2025 database and the IPCC AR6 report. Category 4: Emissions are estimated using a hybrid methodology combining supplier-specific emissions reported by two of our major shipping providers and a mode-specific, spend-based calculation on all other logistics spend. Category 5: Waste data is collected from our sites globally. End-of-life, disposal method-specific emissions factors were obtained from the 2025 DESNZ Conversion Factors database. A recycling emission factor was obtained from the 2021 DESNZ Conversion Factors database. Category 6: Emissions are calculated in accordance with the GHG Protocol based on commercial and private jet air travel data and spend data provided by our travel agencies. Flights were categorized as either short-, medium-, or long-haul and the appropriate DESNZ 2024 factors are applied. For car rental and train travel, we used spend-based estimates from the CEDA 5.0 EEIO database. Category 7: Emission estimates are based on employee survey data conducted every 2-3 years. Offices are split into geographic regions, and commuting benchmarks were calculated per these regions (EMEA, North America, South Asia, East Asia/Southeast Asia.) The commuting modes are cycling/biking, van transport, public transport, and driving alone. Public transport is assumed to be a mix of train and bus travel, as determined by regional benchmarks. All employees are assumed to work 250 days with an assumption applied for full time onsite versus part time (hybrid) versus remote. Data for 2022-2023 included the ~10 largest campuses whereas 2024 and 2025 included all employees globally. Category 8: For serviced sites, where AMD does not have operational control and metered electricity data is not available, electricity consumption is estimated using square footage, capacity or occupancy, applied in that order. Location-based grid emission factors from IEA 2025 were applied to calculate emissions. Category 11: Emissions are calculated based on total sales volume, average product electricity consumption, and average product lifetime split by product category for products sold in the reporting year. Emissions were calculated by multiplying total energy consumption by the corresponding country-level emission factor from IEA 2025. A percentage of data center-related products are assumed to be powered with renewable electricity based on public reporting from our customers. Category 12: Emissions are calculated based on the average product weight by product category and the total sales volume within the reporting year. A weight-based calculation is used, with the disposal method estimated using region-specific e-waste disposal benchmarks obtained from the Global e-Waste Statistics Partnership (2025). Region-specific waste disposal benchmarks were obtained from the World Bank. Disposal type-specific emission factors were obtained from the EPA GHG Emission Factor Hub (2025). Region-specific blended average waste disposal emission factors were calculated using waste disposal benchmarks. Category 15: FY25 is the first-year AMD estimates investment-related emissions. Estimates are based on revenue, ownership, and holding value data, with investees mapped to CEDA EEIO emission factors using NAICS classifications. For investees with reported revenue, emissions are calculated by multiplying revenue by AMD’s ownership share and the applicable Scope 1 & 2 emissions factor. For investees without reported revenue, emissions are estimated using average holding-value based intensity factors derived from investees with available revenue data.
- The 2024 Scope 3 Category 1, Contract Manufacturing Waste and Water totals were restated to reflect a shift from corporate-level (company-average) to facility-level allocated data to AMD from our primary foundry supplier. The site-specific allocation more accurately reflects the emissions, water and waste attributable to our production. The result of the 2024 restatement reduces the previously reported figures. 2025 emissions reflects the same facility-level allocation.
- Starting in FY25 onward, initially reported Scope 3 Category 1 estimated emissions will incorporate a one-year data lag, meaning the AMD current reporting year value corresponds to the year prior for suppliers, thereby allowing suppliers to have sufficient time to complete their full inventories and provide accurate emission values. An exception is for our largest supplier whereby emissions data from the reporting year is used. For estimating all other suppliers’ emissions in the reporting year, we factor each suppliers’ carbon intensity from the prior year (facility level Scope 1 and 2 and corporate-level Scope 3 divided by AMD spend with the supplier) with the reporting-year spend with that supplier, thereby using spending changes as a proxy for changes in emissions when previous year supplier data is used. Once reporting year data is received and reviewed, AMD will recalculate Scope 3 Category 1 emissions for the prior reporting period and clarify restatements attributed to the difference between spend-based estimates and verified supplier values. The restatement may also require restatement on our progress to goal for related to manufacturing emissions intensity.
- Beginning in 2025, emissions from AMD serviced locations, meaning AMD leases space and does not have operational control, are reported under Scope 3 Category 8 (Upstream Leased Assets).
- In 2023, AMD restated estimated 2022 product use and end of use estimates (Scope 3 categories 11 and 12) to include Xilinx and Pensando data following 2022 acquisitions.
- Adjustments to data center lifetime assumptions and clarifications within the embedded product category led to recalculation of 2024 and applied to 2025 reported values.
- Adjustments to weight assumptions of products between the data center and client applications led to a recalculation of both 2024 and applied to 2025 total product weight as well as the resulting end of life emissions.
- Increase in emissions was from a higher share of incineration globally per the World Bank’s disposal method benchmarks. This was notable in North America and East Asia
- HFE-related F-gas emissions (zero ozone depletion potential) were reported under Scope 1 in prior years. Beginning 2025, they are disclosed separately under Other Emissions as F-Gas Emissions
- AMD moved its data center operations from Atlanta in 2021 to a collocated facility but decommissioned server and other equipment since 2023.
- 2025 hazardous waste generated by contract wafer foundries was estimated because actual data was not available at the time of reporting. The prior year intensity value (hazardous waste generated / $ USD spend) for each supplier was applied to the reporting year spend to estimate hazardous waste generated by each supplier for the reporting year. Once reporting year data is received and reviewed, AMD will recalculate hazardous waste generated for the prior reporting period and clarify restatements.
- AMD generates a limited amount of wastewater that requires treatment by the municipal wastewater treatment plant, in accordance with water quality permitting.
- On 12/14/2023, the City of Markham performed an unannounced McNabb construction site visit and notified AMD of a failure to meet the “Tree Preservation By-Law.” An approved permit and protection barriers were needed prior to beginning outdoor work laying conduit and the construction contractors failed to comply. The City required an arborist to inspect the trees and provide a tree impact report. The arborist, hired by the construction company (DPI), came to the site on 12/16/2023 and provided a report to the City which stated that three trees (each 14 inch+ in diameter) were severely damaged and will need to be removed. The City required AMD to remove the three trees by Jan 31, 2024, and gave an option of planting 18 trees or pay a fine of USD$10,800 (USD$600 for each tree not planted) by May 31, 2024. An arborist (hired by DPI) removed the trees on January 24, 2024. After discussion with the property owner, AMD opted to replace three trees and pay cash in lieu for fifteen. Action of tree removal was confirmed by the City on January 26, 2024, and cash in lieu on April 17, 2024. The City confirmed compliance via email on May 9, 2024.
- A stormwater inspection at San Jose found two environmental issues related to outdoor cleaning which were quickly fixed and a follow-up inspection confirmed all concerns were resolved. A notice of violation and a USD$3,000 settlement offer were received from the San Francisco Bay Regional Water Quality Control Board regarding a xylene exceedance in groundwater associated with the AMD Superfund site. Despite evidence that there was no bypass, violation or unauthorized discharge, AMD decided not to contest the violation and paid the fine. An inspection by the San Jose Fire Department noted missing signage. Temporary signs were placed during the inspection and permanent signs were installed shortly thereafter.
- The City of San Jose issued a notice of violation for the lack of building specific organic waste diversion service registration for two buildings at the San Jose site. The administrative error was corrected. The San Francisco Bay Regional Water Quality Control Board issued a Notice of Violation, and a $6,000 settlement offer regarding a zinc exceedance in groundwater discharge associated with AMD’s Superfund site. Despite evidence that there was no bypass, violation or unauthorized discharge, AMD decided not to contest the violation and paid the fine.
- Reported data include product materials only and excludes packaging materials.
- In 2025, we repaired, refurbished or recycled over 768,000 units, weighing 91 metric tons, including categories such as quality returns for defects, damage in transit or end-of-life evaluation units across all geographies. Stock rotation, fulfillment error and other return to stock categories are not included. In 2024, we did not include in our reporting, but we repaired, refurbished or recycled over 1,086,783 units, weighing 106.9 metric tons. Cumulative data since 2022 followed these same assumptions, totaling 3.3 million units, weighing over 250 metric tons.
- Reported data include AMD employees, temporary workers and contractors.
- Genderqueer was added as a new category for employees to self-identify their gender beginning in 2022.
- AMD Executive Leadership Team (c-suite).
- Includes people managers in director- or fellow-level positions and above. Reported employee data exclude employees with unknown gender.
- Number of total terminations divided by the average monthly regular headcount throughout the year.
- Voluntary terminations are defined as the number of regular employees leaving AMD voluntarily, excluding retirements.
- Other turnover rate (%) is defined as the number of other terminations divided by the average monthly regular headcount throughout the year. Other termination reasons include involuntary termination, reduction in force (RIF), death and retirement.
- Based on employees who were on parental or maternity leave with a start date in that year and who returned to work after the leave end date (with voluntary or family reasons). Reported data are for AMD employees.
- AMD offers paid parental leave globally to its FTE employees. “Parental leave” can encompass various types of leave in different countries; it provides time off for parents to nurture their families and make necessary life adjustments.
- Includes AMD employees hired on or before September 30 of the reporting year.
- Employee categories are defined as follows: Executive: director-level positions and above; Manager: manager and senior manager Job Titles; Technical Staff: member of technical staff through corporate fellow; Professional: supervisor positions; and Staff: employees from junior-level positions to professional.
- AMD employee base pay is benchmarked against the competitive high-tech market of companies of our size, with midpoint targets set at or above the market median.
- Reported value is based on the total number of employees across all locations who received career- or skills-related instructor led training - excluding compliance, function-specific or eLearning trainings.
- Reported value is calculated based on the number of employees who completed the AMD Global HSSE Workplace Orientation training during the reporting period.
- As of December 31, 2025, 99% of eligible employees had completed their 2025 training.
- The reported data include AMD employees and contract workers who report directly to an AMD employee. Our reporting guidelines are based on OSHA reporting criteria. Minor (first-aid-level) injuries are not included. Lost days are calculated based on scheduled workdays. The scope of our well-being (health and safety) reporting was expanded in 2024 to include additional metrics and categories. Consequently, some data for prior years are not available for comparison.
- Reported data includes workers, employees and contractors, including Service Provider Workers.
- A missing placard was identified during a San Jose Fire Department hazardous materials inspection at the San Jose site. A placard was posted and the agency confirmed the violation was resolved.
- The California Department of Public Health issued an NOV on July 6, 2023, for missing signage on X-ray equipment at the San Jose site. Signs were added and the NOV was closed on Aug. 8, 2023. The San Jose Fire Department issued two NOVs in November 2023 for delayed submittal of fire alarm inspections. The information was submitted and the NOVs were closed on Dec. 1, 2023.
- An inspection by the San Jose Fire Department noted missing signage. Temporary signs were posted during the inspection, and permanent signs were installed shortly thereafter.
- The San Jose Fire Department issued a notice of violation for the omission of liquid nitrogen and carbon dioxide from a CERS submittal. The omitted materials were added to the submittal during the inspection.
- Reported data include AMD employees and contractors.
- Economic data for current and past years are updated annually to reflect the company’s most recent financial reports.
- Reported data by region and focus area exclude Matching Gift and Volunteer Rewards.
- The AMD PAC is 100% funded by personal contributions from eligible donors, including certain employees and their spouses, and is used to support candidates running for public office. AMD cannot and does not make any direct or indirect financial contributions to federal political candidates or committees. Totals of U.S. contributions and receipts associated with the AMD PAC are available on the U.S. Federal Election Commission website. Additional information on AMD Public Policy Engagement and U.S. Political Activities can be found here: https://www.amd.com/en/corporate/corporate-responsibility/governance/public-policy-engagement.html.
Footnotes
- For financial intensity metrics, AMD reports the environmental metric and divides by the corresponding year annual revenue (in millions of USD). Energy use intensity is MWh of energy per $1M USD revenue. Emissions intensity is metric tCO2e per $1M USD in revenue, shown both for location-based and market-based scope 2 emissions. Water use and wastewater discharge intensity are reported as millions of liters of water used or discharged per $1M USD in revenue. Waste is reported as metric tons generated per $1M USD in revenue.
- Indirect energy consumption is comprised of purchased grid electricity, electricity generated from chilled water, on-site solar generation and green tariff electricity.
- Electricity consumption does not include chilled water; it includes purchased grid electricity, on-site solar generation and green tariff electricity
- AMD sources renewable energy through utility providers (green tariffs), and sellers of third-party certified renewable energy credits (RECs), as well as generates renewable energy onsite.
- San Jose has fuel cells that convert diesel into electricity. The emissions related to the fuel cells are reported under Scope 1. The site also generates renewable energy onsite, which is included in the site electricity use but assigned a zero emission factor.
- 2020 data include energy use and Scope 1 and 2 GHG emission data from the Xilinx and Pensando acquisitions in 2022 in order to reset the 2020 GHG goal baseline per the GHG Protocol. All 2022 and later year environmental metrics include Xilinx and Pensando data.
- AMD follows the GHG Protocol for Scope 1 emission estimates, the internationally recognized standard for the corporate accounting and reporting of GHG emissions. Scope 1 emissions factors estimated based on quantity of refrigerants and fuel consumed in each geography, including natural gas and refrigerants such as hydrofluorocarbons (HFCs). The scope is based on operational control (i.e., ability to implement operational policies) and the method includes Site Metrics Coordinators entering the monthly amount of fuel and chemicals use, by type, into AMD’s central database, and then applying the emission factors. Emission factors were obtained from three main sources: DESNZ 2025 (previously referred to as DEFRA or UK BEIS in previous documents), IPCC AR6 (6th Assessment Report) or US EPA Fluorinated GHG Report.) Values have undergone third-party limited assurance. Beginning 2025 AMD will account for Fluorinated ethers (FEs) outside of Scope 1 as they are not one of the eight gases or categories of gases required for inclusion in GHG inventories according to the GHG Protocol.
- AMD follows the GHG Protocol for Scope 2 emission estimates by multiplying the quantity of electricity consumed at each site by relevant emission factors. If electricity use data is not available, as for small offices, then an average value for U.S. office buildings is used for all AMD locations (16.9 kWh/sq ft) based on EIA CBECS results for the average administrative office, and the emission factor for the location is applied. AMD applies both the market-based and location-based methods for estimating scope 2 emissions. Market-based emission calculations are based on grid electricity use excluding renewable energy sourced through green tariffs and renewable energy credits (RECs) allocated to each site. Location-based emission calculations are based on the quantity of grid electricity used plus renewable energy sourced through utility “green tariff” programs. Emission factors for locations in the U.S. are based on eGRID and Green-e 2025 databases. Location-based electricity emission factors for Canada are provided in the 2025 Canada NIR report, specific to each Canadian province. Market-based emission factors for Europe were taken from the IEA 2025 emission factor database. Location-based emission factors for all countries other than the US and Canada were taken from the IEA 2025 emission factor database. 2020-25 values have undergone third-party limited assurance.
- Our goal is a 50 percent reduction in absolute Scope 1 and 2 GHG emissions from 2020-2030.
- Our value chain emissions are estimated following the guidance from the GHG Protocol. Annually we update emission factors as available by the referenced sources. Category 1: Emissions are calculated by directly surveying suppliers representing ~85% of our total supply chain spend (includes Foundry, final test/assembly, substrates and others). It includes their estimated allocations to AMD (typically using revenue-based accounting), at a factory-level where available, for their Scope 1 and 2 emissions, as well as upstream Scope 3 if available and third-party verified. For Foundries, we use a third-party lifecycle assessment to estimate upstream emissions. For other suppliers we apply a spend-based method by mapping spend categories to the associated Scope 3 CEDA emission factor. Emissions from all other vendors (including marketing, professional services, real estate, software providers, telecom and networking providers and other manufacturing services) are calculated using a spend-based method. Category 2: Calculations follow a spend-based method and are included in the disclosed figure in Category 1. Category 3: Emissions are calculated using fuel and electricity consumption data collected from our sites globally, and Well-to-tank (WTT) emission factors for natural gas and diesel were obtained from the DEFRA 2025 Conversion Factors database. Emission factors for transmission & distribution-related electricity losses, and electricity-related WTT generation and transmission and distribution, were obtained from the IEA Emission Factors 2025 database and the IPCC AR6 report. Category 4: Emissions are estimated using a hybrid methodology combining supplier-specific emissions reported by two of our major shipping providers and a mode-specific, spend-based calculation on all other logistics spend. Category 5: Waste data is collected from our sites globally. End-of-life, disposal method-specific emissions factors were obtained from the 2025 DESNZ Conversion Factors database. A recycling emission factor was obtained from the 2021 DESNZ Conversion Factors database. Category 6: Emissions are calculated in accordance with the GHG Protocol based on commercial and private jet air travel data and spend data provided by our travel agencies. Flights were categorized as either short-, medium-, or long-haul and the appropriate DESNZ 2024 factors are applied. For car rental and train travel, we used spend-based estimates from the CEDA 5.0 EEIO database. Category 7: Emission estimates are based on employee survey data conducted every 2-3 years. Offices are split into geographic regions, and commuting benchmarks were calculated per these regions (EMEA, North America, South Asia, East Asia/Southeast Asia.) The commuting modes are cycling/biking, van transport, public transport, and driving alone. Public transport is assumed to be a mix of train and bus travel, as determined by regional benchmarks. All employees are assumed to work 250 days with an assumption applied for full time onsite versus part time (hybrid) versus remote. Data for 2022-2023 included the ~10 largest campuses whereas 2024 and 2025 included all employees globally. Category 8: For serviced sites, where AMD does not have operational control and metered electricity data is not available, electricity consumption is estimated using square footage, capacity or occupancy, applied in that order. Location-based grid emission factors from IEA 2025 were applied to calculate emissions. Category 11: Emissions are calculated based on total sales volume, average product electricity consumption, and average product lifetime split by product category for products sold in the reporting year. Emissions were calculated by multiplying total energy consumption by the corresponding country-level emission factor from IEA 2025. A percentage of data center-related products are assumed to be powered with renewable electricity based on public reporting from our customers. Category 12: Emissions are calculated based on the average product weight by product category and the total sales volume within the reporting year. A weight-based calculation is used, with the disposal method estimated using region-specific e-waste disposal benchmarks obtained from the Global e-Waste Statistics Partnership (2025). Region-specific waste disposal benchmarks were obtained from the World Bank. Disposal type-specific emission factors were obtained from the EPA GHG Emission Factor Hub (2025). Region-specific blended average waste disposal emission factors were calculated using waste disposal benchmarks. Category 15: FY25 is the first-year AMD estimates investment-related emissions. Estimates are based on revenue, ownership, and holding value data, with investees mapped to CEDA EEIO emission factors using NAICS classifications. For investees with reported revenue, emissions are calculated by multiplying revenue by AMD’s ownership share and the applicable Scope 1 & 2 emissions factor. For investees without reported revenue, emissions are estimated using average holding-value based intensity factors derived from investees with available revenue data.
- The 2024 Scope 3 Category 1, Contract Manufacturing Waste and Water totals were restated to reflect a shift from corporate-level (company-average) to facility-level allocated data to AMD from our primary foundry supplier. The site-specific allocation more accurately reflects the emissions, water and waste attributable to our production. The result of the 2024 restatement reduces the previously reported figures. 2025 emissions reflects the same facility-level allocation.
- Starting in FY25 onward, initially reported Scope 3 Category 1 estimated emissions will incorporate a one-year data lag, meaning the AMD current reporting year value corresponds to the year prior for suppliers, thereby allowing suppliers to have sufficient time to complete their full inventories and provide accurate emission values. An exception is for our largest supplier whereby emissions data from the reporting year is used. For estimating all other suppliers’ emissions in the reporting year, we factor each suppliers’ carbon intensity from the prior year (facility level Scope 1 and 2 and corporate-level Scope 3 divided by AMD spend with the supplier) with the reporting-year spend with that supplier, thereby using spending changes as a proxy for changes in emissions when previous year supplier data is used. Once reporting year data is received and reviewed, AMD will recalculate Scope 3 Category 1 emissions for the prior reporting period and clarify restatements attributed to the difference between spend-based estimates and verified supplier values. The restatement may also require restatement on our progress to goal for related to manufacturing emissions intensity.
- Beginning in 2025, emissions from AMD serviced locations, meaning AMD leases space and does not have operational control, are reported under Scope 3 Category 8 (Upstream Leased Assets).
- In 2023, AMD restated estimated 2022 product use and end of use estimates (Scope 3 categories 11 and 12) to include Xilinx and Pensando data following 2022 acquisitions.
- Adjustments to data center lifetime assumptions and clarifications within the embedded product category led to recalculation of 2024 and applied to 2025 reported values.
- Adjustments to weight assumptions of products between the data center and client applications led to a recalculation of both 2024 and applied to 2025 total product weight as well as the resulting end of life emissions.
- Increase in emissions was from a higher share of incineration globally per the World Bank’s disposal method benchmarks. This was notable in North America and East Asia
- HFE-related F-gas emissions (zero ozone depletion potential) were reported under Scope 1 in prior years. Beginning 2025, they are disclosed separately under Other Emissions as F-Gas Emissions
- AMD moved its data center operations from Atlanta in 2021 to a collocated facility but decommissioned server and other equipment since 2023.
- 2025 hazardous waste generated by contract wafer foundries was estimated because actual data was not available at the time of reporting. The prior year intensity value (hazardous waste generated / $ USD spend) for each supplier was applied to the reporting year spend to estimate hazardous waste generated by each supplier for the reporting year. Once reporting year data is received and reviewed, AMD will recalculate hazardous waste generated for the prior reporting period and clarify restatements.
- AMD generates a limited amount of wastewater that requires treatment by the municipal wastewater treatment plant, in accordance with water quality permitting.
- On 12/14/2023, the City of Markham performed an unannounced McNabb construction site visit and notified AMD of a failure to meet the “Tree Preservation By-Law.” An approved permit and protection barriers were needed prior to beginning outdoor work laying conduit and the construction contractors failed to comply. The City required an arborist to inspect the trees and provide a tree impact report. The arborist, hired by the construction company (DPI), came to the site on 12/16/2023 and provided a report to the City which stated that three trees (each 14 inch+ in diameter) were severely damaged and will need to be removed. The City required AMD to remove the three trees by Jan 31, 2024, and gave an option of planting 18 trees or pay a fine of USD$10,800 (USD$600 for each tree not planted) by May 31, 2024. An arborist (hired by DPI) removed the trees on January 24, 2024. After discussion with the property owner, AMD opted to replace three trees and pay cash in lieu for fifteen. Action of tree removal was confirmed by the City on January 26, 2024, and cash in lieu on April 17, 2024. The City confirmed compliance via email on May 9, 2024.
- A stormwater inspection at San Jose found two environmental issues related to outdoor cleaning which were quickly fixed and a follow-up inspection confirmed all concerns were resolved. A notice of violation and a USD$3,000 settlement offer were received from the San Francisco Bay Regional Water Quality Control Board regarding a xylene exceedance in groundwater associated with the AMD Superfund site. Despite evidence that there was no bypass, violation or unauthorized discharge, AMD decided not to contest the violation and paid the fine. An inspection by the San Jose Fire Department noted missing signage. Temporary signs were placed during the inspection and permanent signs were installed shortly thereafter.
- The City of San Jose issued a notice of violation for the lack of building specific organic waste diversion service registration for two buildings at the San Jose site. The administrative error was corrected. The San Francisco Bay Regional Water Quality Control Board issued a Notice of Violation, and a $6,000 settlement offer regarding a zinc exceedance in groundwater discharge associated with AMD’s Superfund site. Despite evidence that there was no bypass, violation or unauthorized discharge, AMD decided not to contest the violation and paid the fine.
- Reported data include product materials only and excludes packaging materials.
- In 2025, we repaired, refurbished or recycled over 768,000 units, weighing 91 metric tons, including categories such as quality returns for defects, damage in transit or end-of-life evaluation units across all geographies. Stock rotation, fulfillment error and other return to stock categories are not included. In 2024, we did not include in our reporting, but we repaired, refurbished or recycled over 1,086,783 units, weighing 106.9 metric tons. Cumulative data since 2022 followed these same assumptions, totaling 3.3 million units, weighing over 250 metric tons.
- Reported data include AMD employees, temporary workers and contractors.
- Genderqueer was added as a new category for employees to self-identify their gender beginning in 2022.
- AMD Executive Leadership Team (c-suite).
- Includes people managers in director- or fellow-level positions and above. Reported employee data exclude employees with unknown gender.
- Number of total terminations divided by the average monthly regular headcount throughout the year.
- Voluntary terminations are defined as the number of regular employees leaving AMD voluntarily, excluding retirements.
- Other turnover rate (%) is defined as the number of other terminations divided by the average monthly regular headcount throughout the year. Other termination reasons include involuntary termination, reduction in force (RIF), death and retirement.
- Based on employees who were on parental or maternity leave with a start date in that year and who returned to work after the leave end date (with voluntary or family reasons). Reported data are for AMD employees.
- AMD offers paid parental leave globally to its FTE employees. “Parental leave” can encompass various types of leave in different countries; it provides time off for parents to nurture their families and make necessary life adjustments.
- Includes AMD employees hired on or before September 30 of the reporting year.
- Employee categories are defined as follows: Executive: director-level positions and above; Manager: manager and senior manager Job Titles; Technical Staff: member of technical staff through corporate fellow; Professional: supervisor positions; and Staff: employees from junior-level positions to professional.
- AMD employee base pay is benchmarked against the competitive high-tech market of companies of our size, with midpoint targets set at or above the market median.
- Reported value is based on the total number of employees across all locations who received career- or skills-related instructor led training - excluding compliance, function-specific or eLearning trainings.
- Reported value is calculated based on the number of employees who completed the AMD Global HSSE Workplace Orientation training during the reporting period.
- As of December 31, 2025, 99% of eligible employees had completed their 2025 training.
- The reported data include AMD employees and contract workers who report directly to an AMD employee. Our reporting guidelines are based on OSHA reporting criteria. Minor (first-aid-level) injuries are not included. Lost days are calculated based on scheduled workdays. The scope of our well-being (health and safety) reporting was expanded in 2024 to include additional metrics and categories. Consequently, some data for prior years are not available for comparison.
- Reported data includes workers, employees and contractors, including Service Provider Workers.
- A missing placard was identified during a San Jose Fire Department hazardous materials inspection at the San Jose site. A placard was posted and the agency confirmed the violation was resolved.
- The California Department of Public Health issued an NOV on July 6, 2023, for missing signage on X-ray equipment at the San Jose site. Signs were added and the NOV was closed on Aug. 8, 2023. The San Jose Fire Department issued two NOVs in November 2023 for delayed submittal of fire alarm inspections. The information was submitted and the NOVs were closed on Dec. 1, 2023.
- An inspection by the San Jose Fire Department noted missing signage. Temporary signs were posted during the inspection, and permanent signs were installed shortly thereafter.
- The San Jose Fire Department issued a notice of violation for the omission of liquid nitrogen and carbon dioxide from a CERS submittal. The omitted materials were added to the submittal during the inspection.
- Reported data include AMD employees and contractors.
- Economic data for current and past years are updated annually to reflect the company’s most recent financial reports.
- Reported data by region and focus area exclude Matching Gift and Volunteer Rewards.
- The AMD PAC is 100% funded by personal contributions from eligible donors, including certain employees and their spouses, and is used to support candidates running for public office. AMD cannot and does not make any direct or indirect financial contributions to federal political candidates or committees. Totals of U.S. contributions and receipts associated with the AMD PAC are available on the U.S. Federal Election Commission website. Additional information on AMD Public Policy Engagement and U.S. Political Activities can be found here: https://www.amd.com/en/corporate/corporate-responsibility/governance/public-policy-engagement.html.
Social Performance Indicators
The following tables provide annual data on our global workforce including employee diversity, new hires, employee turnover, parental leave, career development, training, health and safety, and volunteerism.
n/a = not available
Values shown in italics represent adjusted data and are different from values shown in previous Corporate Responsibility Reports.