Amazon Workers on Federal Aid Tripled Since 2020 and Delivery Apps Rank Worse

A Government Accountability Office report estimated that Amazon employee enrollment in Medicaid and SNAP nearly tripled from February 2020 to September 2025 across 11 reporting states, as reported by The Washington Times. The finding that matters most for job seekers is not about any single company. It is that most people enrolled in these programs are employed. At Metaintro, we track roughly 50 million job postings, and this dataset is a rare, unusually direct measure of whether a job actually covers the cost of living. Treat it as a due diligence tool when you are comparing offers, not as a headline about one employer.

What Did the GAO Report Find?

The central estimate is that Amazon employee enrollment in Medicaid and SNAP nearly tripled from February 2020 to September 2025 across 11 reporting states.

The ranking shift is equally notable. The collective workforce of Uber, Lyft, DoorDash, Grubhub and Instacart now surpasses Walmart as the employer with the most SNAP recipients. Amazon moved into second place on the Medicaid list, behind Walmart.

The scope of the data deserves attention before anyone over-reads it. The report examined Medicaid data from six states and SNAP data from nine states, with four states reporting both programs. That is a partial national picture assembled from the states that report, not a complete census of every employer in the country.

The national totals give the trend its weight. 13.8 million Americans were enrolled in Medicaid in 2024, up from 12 million in 2020. For SNAP, 10.6 million adults relied on the program in 2024, up from 9 million in 2020.

Why Does This Matter If You Have a Job?

Because the population using these programs is overwhelmingly employed. Among those Medicaid enrollees, 66.1% were working full-time. Among SNAP recipients, 71.3% were employed full-time.

That reframes the entire subject. Benefit enrollment among full-time workers is not primarily a story about unemployment or work ethic. It is a measure of whether full-time earnings at a given employer clear the cost of living in a given state, and increasingly the answer is that for a significant share of workers they do not.

For a job seeker, this converts a political talking point into a practical screening question. When you evaluate an offer, the relevant test is not whether the headline wage sounds acceptable. It is whether that wage, after the specific schedule and benefits on offer, leaves you above the threshold where public assistance becomes necessary in the state where you will actually live.

That threshold varies enormously by state, which is exactly why the report's state-level construction matters. The same hourly rate can be comfortable in one state and insufficient in another, and national averages hide that completely.

What Do the Companies Say in Response?

The responses are worth reading carefully, because they are substantive rather than dismissive and they change how you should interpret the numbers.

The Flex Association countered that 80% of their drivers work part-time to supplement other jobs, and emphasised that app-based workers choose when and how they work.

DoorDash noted that its average delivery driver worked four hours weekly, with 63% supplementing other employment. An Amazon spokesperson highlighted that 74% of the company's employees are enrolled in a company health insurance plan.

These points are genuinely relevant. If the average app-based driver works four hours weekly, counting them as that platform's "workers" alongside a full-time warehouse employee compares two very different things. A platform with many occasional workers will naturally show a large count of benefit recipients without that saying much about the pay rate for an hour worked.

The counterpoint is equally real. Full-time employment dominates the enrollment figures overall, at 66.1% for Medicaid and 71.3% for SNAP, so part-time supplemental work cannot explain the whole pattern. Both things are true at once, and a careful reader holds both.

How Should You Use This When Comparing Offers?

Convert the wage into a monthly net figure before you compare anything. Gross hourly rates are close to meaningless across employers with different schedules, benefit costs and guaranteed hours. What matters is what reaches your account each month after payroll deductions and premiums.

Then interrogate the hours guarantee specifically. A high hourly rate attached to variable or unguaranteed hours can produce lower monthly income than a modest rate with reliable full-time scheduling. The gig platform responses make this point implicitly, since a four-hour weekly average describes flexibility and also describes income that cannot sustain a household on its own.

Price the benefits properly rather than treating them as a checkbox. Amazon's response cited 74% enrollment in a company health plan, which is a meaningful figure, but the questions that determine your actual position are the premium, the deductible, the waiting period before coverage starts, and whether dependents are affordable. Two jobs with nominally identical benefits can differ by thousands of dollars a year on those four variables alone.

For app-based work, calculate net earnings rather than gross. Vehicle costs, fuel, insurance, maintenance and self-employment tax come out of the headline figure, and the responsibility for all of them sits with the worker. Flexibility has genuine value, particularly alongside caring responsibilities or study, but it should be chosen with the true hourly net in view.

Finally, use state context deliberately. Because eligibility thresholds and living costs vary by state, the same offer can be viable in one location and not in another. If you have any geographic flexibility, that is often a larger lever on your real income than negotiating the rate itself.

What Questions Should You Ask Before Accepting a Role?

Ask about guaranteed hours in writing, not scheduled hours. Many roles advertise full-time and deliver variable weeks, and the distinction determines whether your income is predictable enough to plan around. If a recruiter cannot state a guaranteed weekly minimum, treat the advertised annual figure as an optimistic ceiling rather than an expectation.

Ask when benefits actually begin and what they cost you per pay period. A waiting period of several months changes the real value of an offer substantially, particularly if you are leaving a role where you are already covered. The relevant number is not whether health coverage exists but what your premium, deductible and dependent costs total across a year.

Ask how overtime, shift differentials and peak-season hours work. In warehousing, logistics and retail, a meaningful share of annual earnings can come from premium hours, and whether those are guaranteed, offered or rationed makes a large difference to the yearly total. It also affects how much of your income disappears when demand softens.

Ask what the progression looks like in concrete terms. The most useful question is what someone hired into this role two years ago is doing now and what they earn. Vague answers about opportunity are common. Specific answers about typical progression are far more informative, and an employer that cannot give one is telling you something.

For app-based work, ask what the platform counts as an active worker. The reported figures are shaped by that definition, since a workforce including anyone who completed a delivery in a given period looks very different from one counting regular full-time drivers. DoorDash's own point about a four-hour weekly average illustrates precisely why the denominator matters when interpreting any platform statistic.

Why Are These Figures Rising Across Multiple Employers?

Several forces are operating at once, and separating them is more useful than attributing the trend to any single cause.

Cost of living has risen faster than entry-level wages in many states, which mechanically pushes more full-time workers across eligibility thresholds even when their nominal pay increases. A raise that trails housing and food costs leaves a worker further from self-sufficiency despite earning more, and that dynamic is invisible if you look only at wage growth.

The composition of available work has shifted toward flexible and app-mediated arrangements. That shift explains why delivery and ride-hailing platforms now collectively exceed a traditional retailer on SNAP counts. Flexible work suits many people genuinely well, and it also tends to shift costs such as vehicles, insurance and equipment onto the worker, which reduces net income without reducing headline pay.

Health coverage economics play a role too. Where premiums and deductibles absorb a large share of take-home pay, a worker can hold employer coverage and still fall within assistance thresholds. Amazon's citation of 74% enrollment in a company plan and the presence of employees on Medicaid are not contradictory statements, which is exactly why the cost of coverage matters more than its existence.

Finally, scale itself distorts the ranking. The largest employers appear at the top of these lists partly because they employ enormous numbers of people. A ranking by absolute count will always favour the biggest workforces, which is why rate-based comparisons would be more informative than the raw ordering and why the ordering alone should not drive your decision about where to work.

What Does This Say About the Wider Labour Market?

It says the traditional signal that a job is sufficient has weakened. Full-time work has historically been treated as the dividing line between needing assistance and not needing it, and these figures show that line has moved for a substantial share of workers.

The trend direction is consistent across both programs. Medicaid enrollment rose from 12 million to 13.8 million, and SNAP from 9 million to 10.6 million over the same period. Two independent programs moving the same way over four years is a pattern rather than a measurement artefact. It also predates and outlasts any single hiring cycle, which means it is unlikely to resolve simply because the labour market tightens for a quarter or two. Structural gaps between pay and living costs close slowly, and usually only where wages rise faster than housing, which has not been the recent pattern in most states.

The compositional change matters too. Delivery and ride-hailing platforms collectively overtaking a traditional retailer on SNAP counts reflects how much employment has shifted toward app-mediated and flexible arrangements. That shift brings genuine autonomy and also transfers cost and risk from company to worker, and the enrollment data is one of the few places where that transfer becomes measurable.

None of this means avoiding these employers. Large retailers and platforms remain among the most accessible entry points into work, particularly for people without formal credentials, and accessibility has real value. The right conclusion is narrower and more useful. Treat any single job as a starting position rather than a destination, verify that its actual net income covers your actual costs in your actual state, and keep building the skills that let you move up from it. That progression is the part you control, and over a few years it matters far more to your position than which large employer you started with.

There is one more habit worth adopting from this data. Because the report is built state by state, it is a reminder that almost every meaningful pay comparison is local rather than national. Two identical job titles at the same company can leave workers in very different financial positions depending on where they are based, and that gap is usually larger than any raise you could negotiate on the same offer. When you assess where to work, treat location as a compensation variable in its own right, not merely as a lifestyle preference, and run the net-income calculation for the specific state before you weigh anything else. The workers appearing in these figures are, in large part, people for whom that calculation was never made explicit, and making it explicit is the single most useful thing this report can prompt you to do.

Related Articles

• Amazon Layoffs Are Crushing Workers in a Saturated Job Market
• The K Shaped Job Market and the Closing Wage Gap
• Surge Pricing Is Squeezing Gig Workers and How to Fix It
• Underemployed and Overlooked, When a Job Offers Too Little Work
• Transit Worker Overtime and Out-Earning Most Professionals
• Airport Workers May Not See the New Benefits Fee
• Can Your Boss Dock Your Pay When You Are Late
• Walmart AI Reshapes Associate Jobs Without Headcount Cuts
• Best States to Get Trained and Hired in 2026

People Also Asked

Q: Do most people on SNAP and Medicaid work?

A: The report indicates most do. Among Medicaid enrollees counted, 66.1% were working full-time, and among SNAP recipients 71.3% were employed full-time.

Q: Which employers have the most workers on assistance?

A: On SNAP, the combined workforce of Uber, Lyft, DoorDash, Grubhub and Instacart surpassed Walmart. On Medicaid, Amazon moved into second place behind Walmart. The data covers Medicaid in six states and SNAP in nine, so it is a partial rather than national ranking.

Q: Does this mean I should avoid these employers?

A: Not necessarily. These are among the most accessible entry points into work, and the companies note context, including that 80% of app drivers work part-time to supplement other income. The practical step is to calculate net monthly income after benefits costs for your own state before accepting.

Wondering what you should earn? Metaintro shows you what employers are actually paying for your skills right now so you can judge an offer against the market instead of guessing. Create your free profile and compare your options.