The Rule Keeps Changing. The Hazard Doesn't.
The RMP rule's Inherently Safer Technology provision, known as IST/STAA, has been finalized, rescinded, and reproposed four times in roughly a decade.
Obama's EPA required it in 2017. Trump's EPA killed it in 2019. Biden's EPA brought it back with teeth in March 2024. Trump's EPA published a proposed rollback on February 24, 2026, with a comment period that closed April 10.
You probably think you missed the memo on a debate that never really involved you. That's partly true. The 2024 STAA requirement, the one currently being rolled back, applied to Program 3 facilities in NAICS (North American Industry Classification System) codes 324 and 325: petroleum refining and chemical manufacturing. Cold storage, food processing, ice manufacturing weren't the primary target. IIAR said it plainly in their 2025 technical paper: the IST requirement is "not applicable to most facilities with ammonia refrigeration."
The NAICS carve-out is real. The concept behind the mandate has been embedded in your PHA since the first time your team asked whether you could reduce the system charge.
The regulatory scope is correct: the 2024 STAA mandate wasn't written for this industry. But IIAR's statement tends to get read as something broader: that IST thinking doesn't apply to ammonia refrigeration. That reading is wrong, and the evidence against it is the AIM Act.
The same federal government currently rolling back STAA is actively mandating more ammonia installations through the AIM Act, the 2020 law phasing down HFCs that is pushing new industrial refrigeration installations toward ammonia. Every new ammonia system being specified right now is the outcome of an IST analysis. Someone asked which refrigerant is safer over a relevant time horizon, accounting for relevant second-order consequences, and the answer came back: ammonia.
The CCPS (Center for Chemical Process Safety) foundational text on IST uses refrigeration history as its opening example — and it's not hard to see why. Ammonia was the original refrigerant — and for good reason. It's the most thermodynamically efficient refrigerant available. It announces itself when it leaks. It has zero global warming potential and zero ozone depletion. CFCs replaced it mid-century, and from a pure acute-toxicity standpoint, that was inherently safer. Then CFCs destroyed the ozone layer. The industry came back to ammonia — not because regulations forced it, not because it was cheaper, but because after a century of alternatives, nothing else does the job as well with a cleaner environmental ledger. The AIM Act is just the federal government finally catching up to what the industry already knew.
That's what IST actually is. "Inherently safer" has always been relative to the hazard you're measuring, the time horizon you're using, and the second-order effects you're willing to count. For facilities where that analysis is settled by physics and a century of operational history, requiring documentation of it on a federal compliance timeline every five years is where the mandate falls apart — not the concept.
The engineers who live with these decisions have been doing IST analysis for decades. The AIM Act just made it federal policy.
What happens to the engineering rationale behind your system when nobody wrote it down?
Every time a PHA team asks whether you could reduce the system charge, that's an IST conversation. Whether a secondary refrigerant loop is appropriate for a particular load: same thing. Every charge minimization decision, every cascade discussion, every inventory review against the design load maps directly to the framework's four principles: minimize, substitute, moderate, simplify. A plant that runs charge calculations against design load per IIAR 2 and selects a secondary CO₂ or glycol loop for a high-occupancy area is running the minimize and substitute analysis, whether or not anyone called it that.
Most of those conversations aren't getting captured.
Facilities that treat STAA as a compliance checkbox: document it when required, archive the report, move on. They have nothing when the rule disappears. When it returns, they're starting from zero. No record of what was considered, what was rejected, why the engineering judgment landed where it did.
Facilities that capture IST-informed PHA decisions as engineering records, not because a regulation required it but because engineering rationale is worth keeping, build something different. When an incident happens, when OSHA shows up, when the rule changes again: they have evidence. Not a checkbox. A record.
The practical move isn't complicated: make IST thinking an explicit part of your PHA, regardless of whether STAA is currently required. Map your charge minimization decisions to the four principles. Document why you're at the charge you're at and what alternatives were evaluated. That record has value independent of any regulation.
The 2026 proposed rule strips STAA requirements for existing processes but retains them for all new Program 3 processes regardless of NAICS code. A new ammonia system, the kind the AIM Act is pushing this industry toward, could face STAA documentation requirements for the first time, even at a facility that was never in NAICS 324 or 325.
The rule isn't finalized. EPA is targeting completion before May 2027 compliance deadlines.
The ammonia charge in your system exists for reasons. The engineering decisions that determined that charge were made by someone, with reasoning that either got written down or didn't.
The rule will change again. The hazard won't. Whether minimizing your charge exempts you from that question entirely is a different argument — and a harder one.