What changed
Two things converge: (1) EPA finalized enforceable Maximum Contaminant Levels for six PFAS in April 2024, setting PFOA/PFOS at 4.0 ppt — a binding nationwide obligation forcing thousands of utilities to test and, where they exceed, install treatment (FACT, federalregister.gov 2024-07773). (2) Schema-defined structured extraction from arbitrary websites is now a single API call (Context.dev), which removes the scraper/parser build that previously made harvesting thousands of non-standardized utility water-quality disclosures uneconomic for a solo dev (FACT that the API exists; that it cleanly parses every CCR PDF/page is a HYPOTHESIS to validate).
Why now
The compliance-testing wave is front-loaded: UCMR 5 already compels systems serving ≥3,300 people plus a sample of smaller systems to collect PFAS occurrence data (FACT, federalregister.gov 2021-27858), and the 2024 MCL rule converts those exceedances into mandatory treatment decisions. Vendors are hunting for the exact list of newly-out-of-compliance systems right now, before the buying decisions are made and before the data aggregates into commodity.
Converging signals
A regulation that creates a forced-filer/forced-remediator class (utilities) + a new cheap capability to structure their public disclosures + a discretionary buyer (treatment vendors) who profits from reaching those utilities first.
Customer pain
Treatment-equipment makers and consultancies must find which of ~50,000 US public water systems actually exceed 4.0 ppt to prioritize sales outreach. Today this is manual: reps read scattered Consumer Confidence Reports, state dashboards, and eventual EPA UCMR releases, or buy generic construction/permit lead lists that are not PFAS-qualified. HYPOTHESIS (needs validation): the pain is acute enough that they will pay for a clean, monitored exceedance list.
Who pays
PFAS treatment-equipment makers (GAC, ion-exchange, RO systems), water-engineering consultancies, and testing labs buying qualified leads. NOTE: the buyer is NOT the forced filer — the utility is compelled, the vendor is a DISCRETIONARY buyer of leads. That is the central risk: the mandate proves the market exists, not that vendors will pay YOU for the list.
Solved today
Vendor reps manually monitor CCRs/state portals, attend industry conferences, buy generic infrastructure lead lists, or wait for EPA to publish UCMR 5 occurrence data (which EPA does publicly and for free).
Why current solutions are bad
Manual monitoring is slow and misses freshness; generic lead lists aren't PFAS-qualified; the free EPA UCMR release is periodic, not continuous, doesn't map cleanly to a rep's territory, and lags smaller-system CCR disclosures. But this free public path is also the single biggest threat to the product (see kill arguments).
Proposed product
A subscription lead/monitoring feed: normalize utility CCRs and state drinking-water dashboards into a structured record per water system (system ID, population served, latest measured PFOA/PFOS/four others, exceedance flag vs 4.0 ppt, sample date, contact, county/territory), with change-detection alerts when a system newly exceeds. Delivered as a filterable dashboard + CSV/API export, sliceable by state and sales territory.
MVP version
Pick ONE state with a decent public water-quality dashboard (e.g. a state with a searchable PFAS results portal). Use the extraction API to harvest and normalize its systems to the six regulated PFAS + 4.0 ppt threshold, validate against the official state dataset and EPA UCMR 5 for accuracy, and produce a ranked exceedance list with contacts. Wrap in a simple gated web view.
30-day build
Build the extractor+normalizer for one state; validate accuracy against official data; assemble the target list of treatment vendors/consultancies operating in that state; run the founder's own kill test — cold-pitch three vendors with a free sample of the exceedance list and ask for a paid pilot.
60-day build
If ≥1 vendor commits to pay, harden the pipeline, add change-detection/alerts, expand to 3-5 states, and add territory filtering. Stand up billing (card, monthly). Add a data-accuracy/QA step since a bad lead destroys trust fast.
90-day revenue plan
Convert pilots to paid subscriptions ($300-1,500/mo per vendor depending on geography/seats); expand state coverage as sales demand pulls it. Target 5-15 paying vendor/consultant accounts. Replicate the same feed for adjacent regulated contaminants (lead, nitrate) once PFAS proves the model.
Distribution path
Direct, demonstrated-value outreach: a defined, findable set of PFAS treatment vendors, reps, and consultancies (LinkedIn, trade associations like AWWA/WQA, industry directories). Lead with a free single-state sample list — the product sells itself if the leads are real. No procurement cycle.
Pricing hypothesis
Per-seat/per-territory monthly SaaS: ~$300-800/mo entry (single state), scaling to $1,500-3,000/mo for multi-state or national + API access. Possibly a per-lead or annual-territory-exclusive tier.
Technical difficulty
Low-to-moderate. Core build is small thanks to the extraction API; the real work is source discovery (where each state/utility publishes results), normalization to a common schema, threshold logic, and — critically — accuracy validation and change-detection.
Legal / regulatory risk
Low. Aggregating public water-quality disclosures is lawful; CCRs and state portals are public records. No licensure required. Avoid implying official/EPA endorsement and keep a clear data-provenance/disclaimer trail.
Platform dependency
Moderate. Heavy reliance on one third-party extraction API (Context.dev, a new YC startup) is a single point of failure for cost/uptime — mitigate by keeping the normalizer portable so the extraction backend is swappable. No government platform can deplatform a read-only public-data aggregator.
Founder fit
Strong. This is a data/report + public-records + compliance-monitoring product — squarely in the founder's stated preferences and industrial/regulatory comfort zone. It leans on his public-records and automation strengths and sells through demonstrated value, not relationship sales. It is NOT the per-filing government-portal model he has proven, but it's an adjacent, well-matched data play.
Breakout potential
Good. If PFAS works, the same harvest-normalize-alert engine replicates across every regulated drinking-water contaminant and all 50 states, and the underlying capability generalizes to any 'monitor thousands of public disclosures for a threshold breach' lead product (environmental, permits, violations).
Final recommendation
BUILD THE ONE-STATE MVP AND RUN THE KILL TEST FIRST. The convergence is real and the build is cheap and fast, but the entire thesis rests on an unproven assumption — that treatment vendors will PAY for a PFAS-qualified exceedance list rather than use free EPA/state data. Do not scale before a paying pilot. The differentiator must be freshness + territory-mapped contacts + alerts, not raw data access.
Next action
Pick one state with a public PFAS results portal, build the extract+normalize pipeline, validate it against the official dataset, generate a ranked exceedance-with-contacts sample, and cold-pitch three treatment vendors/consultancies for a paid pilot within 30 days.