EPA's Final Nationwide PFAS Tests Are In: About 1 in 13 Water Systems Over the Limit, Plus a Lithium Surprise

For three years, EPA had nearly every mid-size and large water system in the country test its treated water for 29 PFAS "forever chemicals" and lithium. The final results are now public. They answer a question many households have been asking since the 2024 PFAS limits were announced: how many water systems are actually over the line?

The Headline Numbers

EPA released the final dataset from the Fifth Unregulated Contaminant Monitoring Rule (UCMR 5) in August 2026, with a summary report dated July 2026. Sampling ran from January 2023 through December 2025 and covered all community and non-transient water systems serving more than 3,300 people, plus a nationally representative sample of 800 smaller systems. Water was sampled where it enters the distribution system, after treatment.

Because single samples bounce around, EPA averaged each sampling location's results (two samples for groundwater sources, four for surface water) and compared those averages to the April 2024 limits. About 10,130 water systems had complete results to compare. Here is what that comparison found:

The weighted figure matters because the survey sampled only some of the country's many very small systems. EPA's 7.8% adjusts for that, so it is the best single estimate for the whole country.

Why bigger systems were more likely to be over

EPA counted a system as over if any one of its sampling locations averaged above a limit. Large utilities often draw from many wells, rivers, or plants, so they have more locations and more chances for one to land above the line. That means a large-system result does not tell you that every household on that system gets water at that level. It tells you at least one entry point did. Your utility's Consumer Confidence Report or its PFAS page should say which sources were affected.

Which Chemicals Drove the Results

Almost all of the exceedances came from the two oldest and best-studied PFAS, PFOS and PFOA. Both have a limit of 4.0 parts per trillion (ppt). The other three individually regulated PFAS, and the combined "hazard index" mixture limit, were rarely exceeded. Number and share of systems whose average was over each limit:

A system that was over for more than one chemical appears in more than one row, so the rows do not add up to the totals above.

This table bears directly on the policy fight. In May 2026, EPA proposed rescinding the limits for PFHxS, PFNA, GenX, and the hazard index while keeping the PFOA and PFOS limits. By EPA's own data, those four standards would have flagged roughly 1% of systems or fewer. PFOA and PFOS, which EPA proposes to keep, account for most of the systems that were over. For the people served by the systems that were over on PFHxS or GenX, though, the proposal matters a great deal.

PFAS rarely show up alone

EPA found that 66% of sampling locations with any PFAS detection had more than one. 1,874 water systems had at least one location with three or more different PFAS, and one location had 13. That supports the case for treating PFAS as a group rather than one chemical at a time, and it is why the treatment you choose should reduce a broad range of PFAS, not just one compound.

There was some reassuring news too. Two common short-chain PFAS that have screening levels, PFBA and PFHxA, were detected at about 1,900 and 1,800 systems respectively, but no result at any system exceeded the USGS health-based screening level for either one.

The Lithium Finding Nobody Is Talking About

UCMR 5 also tested for lithium, which occurs naturally in some groundwater and is also used in batteries and psychiatric medication. Lithium has no federal drinking water standard and no health advisory. EPA uses a non-enforceable screening value, called a Health Reference Level, of 10 micrograms per liter (µg/L).

• 2,747 of 10,313 water systems (about 27%) had at least one lithium result EPA counts as above that screening level.
• That includes 24.1% of large systems and 28.5% of smaller ones.
• Across all 66,675 lithium samples, 19.7% were above the screening level.

What this means for health is genuinely unsettled. EPA's lithium fact sheet says the agency cannot yet confidently estimate risk for drinking water levels between its 10 µg/L screening value and the far higher doses used in medicine. The screening value exists to help EPA decide whether lithium deserves a closer look, not to mark a known danger line. A result above it is a reason to pay attention to future EPA action, not a reason for alarm.

If you want to reduce lithium anyway, the treatment differs from PFAS. EPA notes that boiling and disinfection do not remove lithium. State health and well-water guidance identifies reverse osmosis and ion exchange as the effective household options, and says ordinary activated carbon filters do not reliably reduce it.

What Changes Next, and When

UCMR 5 results are not compliance results. EPA says so directly in the report. Enforcement will be based on a running annual average of quarterly samples at each entry point, reported to state regulators. The timeline as of November 1, 2026:

• April 2027: Water systems must complete initial PFAS monitoring under the 2024 rule. Many can use their UCMR 5 data toward this.
• April 26, 2029: The legal compliance date for the PFAS limits. It has not changed.
• Proposed, not final: In May 2026, EPA proposed letting water systems apply for two extra years, to 2031, to meet the PFOA and PFOS limits, and proposed rescinding the PFHxS, PFNA, GenX, and hazard index limits. As of this writing we found no final rule in the Federal Register, so the 2024 limits remain the law.

Some states are not waiting. Several, including Illinois, have adopted their own PFAS standards that stay in place regardless of what happens federally.

How to Check Your Own Water System

1. Look up your system in EPA's data. EPA's UCMR 5 Data Finder lets you search by state and water system. You will need your system's name or its public water system ID, which is printed on your water bill or in your utility's annual Consumer Confidence Report.
2. Read the results the right way. A single high result is not the same as an average above the limit. A result marked "below MRL" means the lab could not reliably measure it at that level. For PFOA and PFOS that floor is 4 ppt, the same as the limit, so "below MRL" does not mean zero.
3. Ask your utility what it is doing. Systems that were over have until 2029 to comply, and many have already started treatment projects. Your utility's website or Consumer Confidence Report should describe its plan.
4. On a private well? UCMR 5 did not test private wells at all. To know your levels, have your well tested by a certified lab. See our guide to PFAS testing for well owners.

You can also look up your city on CheckMyTap. Our city data was refreshed against this final dataset in November 2026: city pages show PFOA and PFOS as each system's highest single entry-point average, and our worst water quality list names every city we track where either averaged above 4 ppt (172 of 924 tested).

If Your System Was Over: Filter Options

You do not have to wait until 2029. Point-of-use filters are a practical way to lower PFAS in the water you drink and cook with.

• Look for certification to NSF/ANSI 53 or 58 for PFOA and PFOS reduction. NSF moved PFAS claims into these standards in 2021, replacing the older P473 protocol. Check the specific model on NSF's or WQA's online listings, because certification is model-specific.
• Reverse osmosis systems certified under NSF/ANSI 58 reduce a broad range of PFAS and also address lithium and other dissolved minerals. See our reverse osmosis guide.
• Activated carbon filters certified under NSF/ANSI 53 for PFOA and PFOS are a lower-cost option for PFAS, but they are not a lithium solution.
• Do not boil to remove PFAS. Boiling does not remove PFAS and can concentrate them slightly as water evaporates. Here is why.