GRAS Reform Does Not Mention Nanoencapsulation, Which Is Why It Reshapes the Market

The rule never says nano, and that is the point

The FDA’s proposed rule on substances generally recognized as safe, published on 11 August 2026 as 91 FR 51834, runs 48 pages. It amends 21 CFR parts 170 and 570. It converts a voluntary notification programme into a mandatory one, creates seven categories of exception, and rewrites the threshold-of-regulation maths. It does not mention nanoencapsulation once.

That absence is not an oversight, and it is not reassurance. The rule is drafted at the level of substance identity and conditions of intended use, and nanoencapsulation is precisely the technology that changes those two things while leaving the marketing story — “it’s vitamin D”, “it’s omega-3”, “it’s curcumin” — completely intact. The proposed rule does not need a nanotechnology chapter to reach nanoencapsulated ingredients. It reaches them through the definition of what a GRAS conclusion covers.

For an industry that has spent two decades commercialising delivery systems on the back of a parent molecule’s regulatory file, that is the whole ballgame. The reform does not add a nano-specific hurdle. It removes the thing that made a nano-specific hurdle unnecessary in the notifier’s own reading.

What the rule actually does — and what it does not

Start with the thing the rule is most often mis-described as doing. It does not introduce premarket approval for GRAS substances. The proposed text is explicit: the programme

“do[es] not establish a premarket review program for purportedly GRAS substances… a company may continue marketing a purported GRAS substance before submitting a GRAS notice or after submitting a GRAS notice before it is filed by FDA.”

So the commercial risk is not a licence to be refused. It is administrative non-compliance with a disclosure duty — and the rule says what that non-compliance costs. Failure to meet the notification requirement “would be a factor in FDA’s prioritization of food substances for post-market review”. In a regime where the agency has finite post-market capacity, being placed at the front of that queue is the enforcement mechanism. There is no fine schedule and no approval letter; there is a list, and your name can move up it.

What the rule would require is a notice for any substance introduced into interstate commerce under the GRAS provision, for the use in question, with seven exceptions:

The seven exceptions

1. A no-questions letter already covers the substance under those conditions of use;
2. the substance is listed or affirmed as GRAS in parts 182, 184 or 186;
3. the substance qualifies under § 170.30(d) (pre-1958 conventional food ingredient) or the revised § 170.30(i)(1);
4. FDA has already considered the use through an established process — a voluntary premarket consultation, an animal cell culture consultation — and did not point to a GRAS notice;
5. the use has a granted threshold-of-regulation exemption;
6. an effective food contact notification covers it and you are the listed manufacturer or supplier; or
7. you file a streamlined submission during a one-year window after the effective date, for substances already in commerce before that date.

Exceptions one and seven are where nanoencapsulation lives, and both are narrower than they look. Exception seven is the one everybody in the sector will reach for. Exception one is the one that has quietly been doing the work already.

The parent-file defence, and why the rule ends it

The standard commercial position on a nanoencapsulated ingredient is that the active is GRAS, the carrier is GRAS, and therefore the product is GRAS. It is a reasonable reading of the statute and it is exactly the reading the proposed rule declines to accept.

Exception one — the no-questions letter — comes with a condition stated in the rule itself. A GRAS conclusion within a notice that received a no-questions letter “may not apply to a use of a substance if the identity of, manufacturing process for, or the conditions of use (e.g., food categories, use levels, technical effect, specifications) of that substance are significantly different from those discussed in the GRAS notice”. The rule then spells out the consequence: “a change in manufacturing process may alter the composition, and perhaps the toxicity, of the substance”, and “a substance may not be food grade following a manufacturing process change that introduces impurities”.

Nanoencapsulation changes all three variables at once. It changes the identity — a 200 nm capsule is not the same article of commerce as the free active. It changes the manufacturing process in the most material way available to a formulator. And it changes the technical effect, which is the entire reason anyone pays for it: the same dose delivered at higher bioavailability.

This is not a novel reading invented for this article. FDA laid the ground in 2014 guidance on significant manufacturing process changes, which addresses what happens to the regulatory status of a food substance when its process changes, and which names the case directly: consultation with the agency is prudent “particularly when the manufacturing process change involves emerging technologies, such as nanotechnology”. That guidance is nonbinding and twelve years old. It has never been the operative test, because nothing forced anyone to apply it. The proposed rule gives it teeth by making the notification duty — and therefore the agency’s visibility — unconditional.

The same 2014 guidance is why the “the carrier is GRAS too” half of the defence is weaker than it sounds. It asks whether a change “affects the identity of the food substance” and whether it “affects the regulatory status of the use”. A new wall material, a new encapsulation route, a new particle-size specification: each is a candidate for a new determination, not an inherited one.

The clock, and the gate before it

The rule’s most under-appreciated provision is that a submission is not a notification. The requirement is met only when FDA files the submission as a GRAS notice. A submission that fails the filing gate leaves the notifier non-compliant, and the rule adds a 45-day cap to the pre-filing evaluation precisely so that this gate is a known quantity.

Just over thirteen months, at the outside, from submission to a response letter. That is the planning number for anyone whose launch calendar assumes a GRAS file can be assembled in a quarter. And note what the response can be. A “cease to evaluate” letter now explicitly means the notification requirement is not met — the submission is treated as though it never arrived. An insufficient-basis letter does not breach the notification duty, but it does inform post-market action.

The one-year window is cheap, and it is public

The streamlined submission in new subpart F is the provision most nanoencapsulation developers will use, and it is genuinely light. It requires the submitter’s name and address, the substance’s name, the conditions of intended use (foods, levels, purposes), and evidence that the substance was in interstate commerce before the effective date of the final rule. It does not require underlying safety data. It permits, but does not require, a statement of the statutory basis. It must be filed electronically within one year of the effective date.

Two features deserve a hard look before anyone treats it as a free pass.

One: it is a public list. FDA will post what it receives, and inclusion on the list “would not represent a determination by FDA that the use of the substance is GRAS”. So the trade is a public disclosure of your ingredient, its food categories, its use levels and its purpose, in exchange for skipping the notice. For a category where the technical effect is the competitive advantage, publishing the use levels and the intended foods is not a neutral act.

Two: FDA can escalate you out of it. The rule provides that the agency “may issue a determination that a GRAS notice or food additive petition must be submitted for the intended use of a substance”. Inclusion on the list is a starting position, not a settlement.

And the eligibility condition is where nanoencapsulation runs into trouble again. The evidence required is presence in interstate commerce before the effective date. If the nanoencapsulated version is not yet commercial — if it is in pilot, or in a customer trial, or was reformulated last year onto a different carrier — it is not eligible for the cheap route, and it faces the full notice.

The scale of the gap the agency is trying to close

The numbers in the preamble explain why this is happening, and they are stark in a way that has not been widely reported.

FDA’s Human Foods Program had filed more than 1,200 GRAS notices in total by 25 March 2025, since the programme began in 1998; the Center for Veterinary Medicine had filed 75 by 28 March 2025. Combined, the two centres had filed more than 670 since January 2016.

Against that, the rule cites estimates that as of January 2011 there were more than 10,000 additives in use in food, “including an estimated 1,000 human food substances for which firms had claimed independent conclusions of GRAS status” — that is, substances already on the market that the voluntary programme had never seen.

So the agency’s entire multi-decade dataset is comparable in order of magnitude to the population it knows it is missing. The mandatory regime is not being built on top of a complete register; it is being built to create one for the first time.

The register will also be international. Of the more than 1,200 filed notices, 617 were submitted by foreign firms — 51.4% — and 503 of those came from firms in countries where English is not the primary language, 81.5% of the foreign total. That is why the proposal adds a requirement that any material in a foreign language be accompanied by an accurate and complete English translation. For a nano-ingredient supply chain that runs through contract manufacturers in Switzerland, Germany, Japan and Korea, that translation burden lands on the party least equipped to carry it.

The agency’s own cost estimate is modest against the compliance reality: a present value of USD 89.6 million at a 3% discount over ten years, range USD 34.9 million to USD 210.0 million, annualised at USD 10.5 million. The upper bound is 2.34 times the point estimate — a range that wide is a signal that the agency does not know how many substances will actually file.

The threshold-of-regulation, and the bioavailability problem

One provision deserves more attention from this sector than it has had: the rewrite of the threshold-of-regulation exemption, and its expansion from food contact substances to substances used in food generally. A granted TOR exemption becomes an exception from the notification duty.

FDA presents the numerical change as safety-equivalent, and the arithmetic behind that claim is worth setting out because it is exact and it is not obviously neutral.

Under the current rule, a substance is at the threshold if dietary concentration is at or below 0.5 parts per billion, corresponding to 1.5 micrograms per person per day, on an assumed diet of 1,500 g solid plus 1,500 g liquid food. Check: 0.5 µg/kg across 3 kg of diet = 1.5 µg.

Under the proposal, the test becomes an estimated daily intake of 0.025 micrograms per kilogram bodyweight per day. Check: 0.025 µg/kg × 60 kg = 1.5 µg.

Both land on exactly 1.5 micrograms per person per day. The change is not a loosening or a tightening — it is a switch of denominator from a fixed 3 kg of food to a 60 kg body. That is defensible, and it is also the whole of the “equivalent level of safety” claim: it is equivalence at one body weight.

Now the part that matters for encapsulation. The TOR is a threshold on systemic exposure. Nanoencapsulation exists to raise the fraction of a dose that reaches systemic circulation — that is what a delivery system is for. A formulation that keeps the use level identical while doubling bioavailability has not changed its dietary concentration at all, and has changed the quantity the threshold is actually protecting against. Any TOR reasoning that runs off the label’s grams-per-kilogram rather than off absorbed dose is going to be wrong in exactly the direction that favours the applicant. The proposed rule does not resolve this; it does not have to, because the exemption is granted case by case.

Europe is already there, and the asymmetry is the real commercial fact

None of this is novel in Brussels. In the EU, an engineered nanomaterial sitting on the wrong side of the line falls within the novel food definition in Regulation (EU) 2015/2283, which means authorisation rather than notification. And the data package is specified. EFSA’s 2021 guidance on nanomaterials in the food and feed chain sets out physico-chemical characterisation requirements, the parameters that must be measured, the methods and techniques for characterising nanomaterials in complex matrices, and a tiered toxicological framework. Depending on the initial tier, additional studies may be needed for reproductive and developmental toxicity, chronic toxicity and carcinogenicity, immunotoxicity and allergenicity, neurotoxicity, effects on the gut microbiome and endocrine activity.

Put the two regimes beside each other. In Europe, a nanoencapsulated ingredient can require particle characterisation in food matrices and a full tiered toxicology programme. In the United States, the same ingredient under this proposal requires a notice whose content is defined by the existing GRAS notice format — which is keyed to the identity, manufacturing process and specifications of the notified substance, with no nano-specific data requirement in the rule at all.

Two consequences follow, and both are commercial rather than legal.

The US file will not carry the EU dossier. A no-questions letter is now a meaningful asset in the United States — it is exception number one, the only exception that lets a substance skip notification entirely on the strength of prior review. But its value is US-only and it is thin in a way that will be discovered at the point of a European launch.

The EU dossier is the better evidence. A company that has already produced EFSA-grade characterisation and toxicology for a nano-encapsulate holds almost everything a US GRAS notice needs, and rather more. The reverse is not true.

What this means if you are holding one of these files

Three things, in order of urgency.

Establish the presence evidence now. If the encapsulated version was in interstate commerce before the effective date of the final rule, the one-year streamlined window is available and it is cheap. That determination has to be evidenced — invoices, bills of lading, purchase orders, dated specifications — and it has to be assembled before the window opens, not during it. Note the asymmetry: the window opens on the effective date, which is set by a final rule that does not exist yet. The evidence is retrospective; the decision is not.

Answer the identity question honestly. If the particle size, the wall material, the process or the specification is materially different from whatever the parent file describes, the parent file does not cover it, and the proposed rule makes that consequential in a way it currently is not. The strongest position is a notice for the encapsulated substance in its own right, filed before anyone asks. The weakest is an assertion of inherited GRAS status with a specification sheet that contradicts it.

Price the disclosure, not just the fee. The run-rate cost of a notice is a rounding error against a commercial launch. What is not a rounding error is that either route — the full notice or the streamlined list — puts your ingredient, its use levels and its intended food categories into a public database. For an industry whose differentiation is a delivery system, that is the real cost of the reform, and it is the one nobody has costed.

Where this stands

The proposal was published on 11 August 2026. Comments close on 9 December 2026, 120 days later, under docket FDA-2025-N-3262. The docket had 766 public comments by late October. Industry response so far has been supportive in principle and focused on capacity — the American Feed Industry Association’s president, Constance Cullman, put the operative question plainly: “it is critical that the agency has the resources, capacity and tools needed to administer ingredient reviews in a timely manner.”

That is the right question, and it is not rhetorical. The rule asks FDA to absorb a population of substances it has never counted, at a moment when it is simultaneously promising 45-day filing decisions and a doubled extension window. The 405-day clock is a commitment, not a guarantee, and the agency’s own upper-bound cost estimate is more than twice its point estimate.

But the timing risk is not symmetric across the industry. A company making a commodity ingredient has a filing. A company making a nanoencapsulated ingredient has a filing and a question to answer about whether its existing regulatory position ever existed. The first is a cost. The second is a re-underwriting of the asset. And it is the second that the reform, without ever using the word, has put on the table.

Where this shows up in demand

The exposure is concentrated in categories where the delivery system is the product claim rather than an incidental processing aid. That includes the functional-nutrition formats where encapsulated actives are the differentiator — the dietary supplements channel, functional gummies and jellies and functional confectionery where heat-sensitive actives require protection through processing; fruit and vegetable powders, where encapsulation governs whether a colour or a vitamin survives the dryer; microalgae, where the value is in carotenoids and long-chain omega-3s that oxidise unless protected; and essential oils, the category where the carrier and the active are hardest to separate and where the regulatory question of what the “substance” actually is has never been settled.

Those markets do not get smaller because of a notification duty. What changes is who can enter them cheaply. The incumbent advantage in encapsulation has always been technical; the reform converts part of it into a regulatory file advantage, and file advantages compound in a way that process know-how does not.

Pheonix has made a related argument before about this industry’s relationship with the word nano — that the sector optimises its language to avoid it. Read that piece here: The Industry Optimises to Avoid the Word Nano. The proposed rule is, in effect, a mechanism that stops the language optimisation from working, because the disclosure is keyed to identity and use rather than to vocabulary.