Ultra-processed foods: getting ahead of a definition that will hit your recipes
Long confined to scientific debate through the NOVA classification, the ultra-processed food (UPF) concept is entering the regulatory arena. The FDA opened a consultation in mid-2025 to define a common standard, and a body of work proposes moving beyond NOVA toward more operational "markers of ultra-processing".
For food & beverage, the focus shifts: a product is no longer judged on its nutrients alone, but on how it is processed and why certain ingredients are in it.
From a fuzzy concept to an operational definition
NOVA sorts foods into four groups, the fourth — ultra-processed — covering industrial formulations that combine extracted substances (sugars, starches, protein isolates) with cosmetic additives (colours, flavours, emulsifiers). The criticism of this definition is its imprecision.
Regulators are therefore seeking measurable criteria based on the presence of certain ingredient families and process markers. An operational definition will make a product classifiable — and therefore targetable by policy.
The focus shifts to additives and processes
An emulsifier, flavour or colour added "for texture or appearance" becomes a signal of ultra-processing, regardless of the nutritional profile. The question is no longer only "how much sugar" but "which ingredients, with what function, and why".
This puts qualitative composition — ingredient families, additives, technological functions — back at the centre of a product's assessment.
School canteens and public procurement, the first testing ground
Without waiting for a final consensus, policies are already acting: school catering, public procurement and private certifications are starting to exclude ultra-processed products. A listing can therefore depend on a product's classification well before any general law.
What this means for your product data
Getting ahead means knowing, for every recipe, which ingredients count as a marker of ultra-processing and what function they serve — then simulating clean-label alternatives with their impact on cost, texture and compliance.
A PLM that structures each ingredient by family and technological function lets you instantly flag products exposed to a future UPF definition and drive the reformulation. crumble-ai turns this watch into data work, not a manual portfolio review.
Is your portfolio ready?
See how crumble-ai continuously audits your recipes and blocks non-conformities at the source.
Book a demo