Safety-Net Reliance Has Tripled at Amazon and Tops the List for Gig Drivers, Federal Audit Finds

Between February 2020 and September 2025, the working-age safety net's burden concentrated into a small set of employer and platform categories. A Government Accountability Office audit commissioned by Sen. Bernie Sanders finds Amazon employees' reliance on food stamps and Medicaid nearly tripled, and gig drivers for Uber, Lyft, DoorDash, GrubHub, and Instacart moved from outside the Medicaid top five to inside the top three while holding the No. 1 spot for food stamp use among occupational categories.

The data tracks where the shift is concentrated, not why.

The audit drew on U.S. Census data and state-level employment records from 11 states. It counted working-age adults who enrolled in the Supplemental Nutrition Assistance Program (SNAP, commonly known as food stamps) or Medicaid, then matched those records to employers and platform categories. Among working people using these programs, about two-thirds worked full time, mostly for private-sector employers concentrated in transportation, restaurants and food preparation, and retail sales. The Washington Post first reported the audit's findings.

Amazon employee enrollment in both programs rose from a low base to a roughly threefold increase over a five-and-a-half-year window. Walmart employees also ranked high in raw counts, as did workers at other large private-sector employers. Gig drivers as a category ranked first for SNAP use and rose from outside the top five for Medicaid in 2020 to the top three by the report's window. The report does not break out Amazon's share of national SNAP or Medicaid enrollment, and the 11-state sample limits how broadly the findings generalize.

Amazon pushed back on the comparison. Spokesperson Rachael Lighty called the report "misleading" and argued that it relies on raw counts rather than percentages, a methodological point that matters because Amazon's overall workforce is much larger than the workforces of any single gig platform. Lighty also noted that Amazon's part-time options can push some workers above Medicaid and SNAP income thresholds in ways full-time work at other employers might not. DoorDash cited a survey it commissioned in which roughly one-third of its workers said they take gig work specifically to avoid needing public benefits, casting the work as a buffer against program dependence rather than a path into it. Flex, an industry group that represents Uber, made a similar point, saying flexible platform work lets workers bridge gaps that might otherwise require government assistance.

These responses are not refutations. They are a methodological debate the data itself does not settle. Raw counts capture the absolute size of the public subsidy flowing through a given employer or platform. Percentages capture the rate of program reliance among a given workforce. The audit's authors chose raw counts. Amazon's spokesperson argues percentages are fairer. Both numbers are defensible, and neither answers the question of why a working full-time job in transportation, retail, or warehouse fulfillment now leaves a large share of workers still needing federal nutrition or health benefits.

The question sits outside the report itself. Sanders cast the findings as evidence that taxpayers are subsidizing 'starvation wages' at Walmart and Amazon while those companies pay CEOs 'exorbitant' compensation and spend billions on stock buybacks. The political reading is Sanders', not the audit's. The audit documents the concentration of program reliance. The question of who should pay for low-wage work, and how to count it, is the one the data raises but does not answer.