In FY 2024, 77.0% of E-Rate allocation went to communities where 14.2 million households lack home internet and 9.0 million lack a large-screen device
With $1.8 billion of the $2.76 billion committed in FY 2024 covering recurring internet bills, this study examines whether the program's core objectives are being met and what the communities it serves look like today
Every year, the federal government helps schools and libraries pay their internet bills through a program called E-Rate, part of the Universal Service Fund. In 2024, it approved $2.76 billion in discounts for 21,102 institutions across the country: 18,507 schools and school districts, and 2,595 libraries. This study matches every funded institution to Census data for the communities around it, to examine who the program serves and what those neighborhoods look like.
The discount each institution receives is set by the federal government based on how many students qualify for the National School Lunch Program (NSLP) -- free or reduced-price meals, a standard federal measure of household income. Schools where 75 percent or more of students qualify receive the deepest discounts: 80 to 90 percent off their bill. Schools where fewer than 35 percent qualify receive 20 to 49 percent off.
In 2024, 77.0 percent of E-Rate dollars went to the highest-need category. In the ZIP codes where those institutions operate, Census data shows 14.2 million households have no home internet subscription and 9.0 million have no large-screen device such as a laptop, desktop, or tablet.
With $1.8 billion (64%) of 2024 E-Rate spending going to recurring internet bills rather than one-time infrastructure, and with 14.2 million households in those communities still lacking home internet access, can E-Rate's core objectives be considered met? And if the program were stopped, what would be the consequence for those communities?
E-Rate spending falls into four categories defined by the FCC. The largest is internet access: the recurring monthly bill a school or library pays to an internet provider. In this funding year, that category alone accounted for $1.72 billion, or 62.1% of total commitments. This is not a one-time purchase. It is a bill that returns every month. If the discount were removed, the bill would not disappear -- each institution would need to cover the full cost on its own budget.
$2.76 billion across four FCC-defined categories
The FCC defines each category. Descriptions below use plain terms.
The Census Bureau tracks not just whether households have internet, but what devices they use. A household with internet access only on a smartphone but no laptop, desktop, or tablet is counted as smartphone-only. A household with no computing device of any kind is counted separately. The figures below cover ZIP codes served by the highest-discount institutions, weighted by household count.
For comparison, in ZIP codes served by the lowest-discount institutions (20-49% band): 2.0 million households (16.1%) have no home broadband, and 1.0 million (8.0%) have no large-screen device.
The table below shows all four discount bands, the institutions in each, the committed dollars, how that money splits between recurring internet service and equipment, and the Census figures for surrounding communities. All household figures are estimates derived from Census data.
Where the $2.76 billion goes
Discount band set by NSLP enrollment share per FCC matrix. Broadband, device, and poverty figures are household-weighted averages from ACS 2020-2024. Household counts are Census estimates.
| Discount | NSLP threshold | Institutions | Committed | Recurring vs Equipment | Population | Total households | Households without broadband | Households without large-screen device | ZIP poverty rate |
|---|---|---|---|---|---|---|---|---|---|
| 80-90%+ | 75% or more of students | 11,869 10,001 schools, 1,868 libraries |
$2.13B 77.0% of total |
64.0% recurring, 36.0% equipment
|
129.5M | 49.7M | 14.2 million (28.5%) | 9.0 million (18.2%) | 16.7% |
| 70-79% | 50-74% rural / 75% or more urban | 2,276 1,898 schools, 378 libraries |
$199M 7.2% of total |
79.0% recurring, 21.0% equipment
|
11.3M | 4.5M | 1.4 million (30.4%) | 700K (15.6%) | 10.8% |
| 50-69% | 35-74% of students | 4,446 4,141 schools, 305 libraries |
$366M 13.3% of total |
55.0% recurring, 45.0% equipment
|
49.6M | 19.3M | 3.8 million (19.9%) | 2.0 million (10.3%) | 9.2% |
| 20-49% | Fewer than 35% of students | 2,655 2,610 schools, 45 libraries |
$70M 2.5% of total |
47.0% recurring, 53.0% equipment
|
32.1M | 12.6M | 2.0 million (16.1%) | 1.0 million (8.0%) | 8.1% |
The two charts below show where institutions and dollars are concentrated across the four discount bands.
Number of schools and libraries at each discount level
Committed dollars at each discount level
The 20 cities below received the most E-Rate funding in 2024. Each row shows the number of schools and libraries receiving funding, the average discount level, and Census estimates for the surrounding communities.
Average discount across all institutions in that city. Broadband and device figures are household-weighted averages from ACS 2020-2024. Household figures are Census estimates. Geographic limitation applies: figures reflect the administrative billing ZIP, not all communities served.
| # | City | Committed | Schools and Districts | Libraries | Average Discount | Total Households in Area | Households Without Broadband | Households Without Large-Screen Device | Smartphone-Only Households | ZIP Poverty Rate |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | New York, NY | $137.0M | 97 | 3 | 77% | 732K1,551K people | 143K(20.0%) | 88K(12.0%) | 55K(8.0%) | 16.0% |
| 2 | Bethel, AK † | $96.6M | 1 | 0 | 90% | < 1K< 1K people | < 1K(52.0%) | < 1K(24.0%) | —(22.0%) | 14.1% |
| 3 | Los Angeles, CA | $36.9M | 132 | 3 | 75% | 835K2,257K people | 203K(24.0%) | 122K(15.0%) | 90K(11.0%) | 18.3% |
| 4 | Chicago, IL | $29.9M | 76 | 2 | 64% | 832K1,963K people | 223K(27.0%) | 134K(16.0%) | 90K(11.0%) | 17.4% |
| 5 | Amarillo, TX | $29.1M | 5 | 0 | 77% | 41K98K people | < 1K(23.0%) | < 1K(19.0%) | < 1K(14.0%) | 12.5% |
| 6 | Houston, TX | $29.0M | 64 | 2 | 79% | 653K1,722K people | 182K(28.0%) | 123K(19.0%) | 97K(15.0%) | 19.1% |
| 7 | Unalakleet, AK | $25.5M | 1 | 0 | 89% | —< 1K people | —(75.0%) | —(27.0%) | —(26.0%) | 8.1% |
| 8 | Montgomery, AL | $23.4M | 5 | 1 | 83% | 28K71K people | < 1K(31.0%) | < 1K(25.0%) | < 1K(17.0%) | 30.1% |
| 9 | San Antonio, TX | $23.0M | 34 | 1 | 77% | 307K834K people | 81K(26.0%) | 64K(21.0%) | 49K(16.0%) | 18.3% |
| 10 | Columbia, SC | $21.0M | 6 | 2 | 80% | 64K157K people | 14K(23.0%) | 10K(16.0%) | < 1K(13.0%) | 23.9% |
| 11 | Kotzebue, AK | $19.6M | 1 | 0 | 89% | < 1K< 1K people | —(29.0%) | —(15.0%) | —(13.0%) | 14.0% |
| 12 | Salt Lake City, UT | $19.5M | 7 | 1 | 66% | 63K153K people | 18K(28.0%) | < 1K(14.0%) | < 1K(11.0%) | 14.6% |
| 13 | Albuquerque, NM | $18.7M | 58 | 2 | 79% | 248K590K people | 63K(25.0%) | 37K(15.0%) | 26K(11.0%) | 16.2% |
| 14 | Memphis, TN | $18.6M | 35 | 1 | 82% | 213K533K people | 70K(33.0%) | 52K(25.0%) | 39K(18.0%) | 23.7% |
| 15 | Dallas, TX | $18.3M | 41 | 0 | 74% | 355K899K people | 85K(24.0%) | 55K(15.0%) | 39K(11.0%) | 14.7% |
| 16 | San Juan, PR | $17.9M | 23 | 18 | 85% | 131K289K people | 67K(51.0%) | 50K(38.0%) | 27K(21.0%) | 36.3% |
| 17 | Springfield, IL | $17.5M | 2 | 0 | 68% | 27K60K people | 11K(42.0%) | < 1K(27.0%) | < 1K(19.0%) | 25.9% |
| 18 | La Mirada, CA | $16.7M | 3 | 0 | 77% | 15K46K people | < 1K(18.0%) | < 1K(8.0%) | < 1K(5.0%) | 6.2% |
| 19 | Flint, MI | $15.6M | 15 | 2 | 82% | 50K118K people | 14K(28.0%) | 13K(26.0%) | < 1K(19.0%) | 29.6% |
| 20 | Orlando, FL | $14.6M | 42 | 1 | 68% | 288K781K people | 64K(22.0%) | 32K(11.0%) | 25K(9.0%) | 14.4% |
† Bethel, AK is the billing address for Lower Kuskokwim School District, which serves approximately 56 remote villages across 22,000 square miles of western Alaska. Census figures shown reflect Bethel city only.
Large-screen device: a laptop, desktop computer, or tablet (ACS B28010). Smartphone-only household: has internet access but no laptop, desktop, or tablet. No computing device: no laptop, desktop, or tablet of any kind, regardless of internet access. Home broadband: a wired or fixed-wireless subscription at the residence; mobile data plans are not counted (ACS B28011).
Methodology
All rates are household-weighted averages. ZIP codes are assigned to one discount band using the dominant committed dollar amount to prevent double-counting of population across bands. Committed figures represent approved discounts; disbursements lag by one to two years as institutions invoice after service is delivered and are not used in this study. Entity counts reflect unique Billed Entity Numbers after deduplication. Consortia are grouped with schools as they represent joint school-district applications. Correlation between E-Rate discount rate and reported NSLP enrollment: r = 0.904.
Geographic limitation
93% of committed dollars flow through school districts and consortia matched to their administrative billing ZIP. That ZIP may not represent all communities those institutions serve. Community figures should be read as characteristics of the billing ZIP area, not the full service geography.