Abstract
In 2025–2026, a clear shift is observed from a descriptive approach to the design of nutritional and nutraceutical products based on mechanistic hypotheses, translational research, and systemic analyses, including multi-omics, and considering the role of the food matrix and processing history in shaping bioavailability and physiological effects [1]. Simultaneously, "smart" nanocarriers are identified as a breakthrough in formulation, addressing low solubility, poor stability, and lack of controlled release of active substances, while also enabling stimulus-responsive release at target sites (e.g., pH/enzyme/redox-dependent) [2]. "Green" pathways for obtaining and modifying ingredients are increasingly emphasized, including supercritical CO2 extraction for vitamin D3-containing fractions and biotransformations using microbial fermentation and enzymes, which can increase bioactivity (e.g., conversion of hesperidin to hesperetin) and generate new molecules under mild conditions [2, 3]. In clinical and nutritional medicine, the importance of precision nutrition supported by AI and patient biological data is growing, while regulatory and quality aspects are strengthened: the need for an "evidence–dose–claim" framework and rigorous grading of evidence certainty (e.g., GRADE) for credible health claims and regulatory acceptance [1, 2]. Concurrently, a convergence of food–digital technologies is emerging: from models predicting glycemic response based on CGM and microbiota, to agent-based AI platforms accelerating ingredient and product development [4, 5].
Introduction
The collected works from 2025–2026 describe the "methodological maturation" of research on bioactive compounds and functional foods through the implementation of multi-omics strategies (microbiome profiling, metabolomics, lipidomics) and a shift towards systemic and translational paradigms, more strongly rooted in biological mechanisms [1]. In these same materials, it is consistently emphasized that bioactivity cannot be separated from the food matrix, processing history, and physicochemical stability, as these shape the bioaccessibility and physiological impact of ingredients [1]. In practice, this means that innovation in 2025–2026 is understood not only as a "new ingredient" but also as a new way of manufacturing, stabilizing, delivering, measuring effect, and proving efficacy in a properly defined target population [1, 2].
During this period, technology and regulations become more closely linked: a scientific evaluation framework centered on the "evidence–dose–claim" axis is directly postulated, intended to rigorously validate efficacy, safe dose-response relationships, and suitability for target populations for each ingredient [2]. Concurrently, in the digital realm, the necessity of standardizing data formats and analytical pipelines is emphasized to ensure predictive models are reliable and reproducible, and the importance of responsible implementation for translationality into clinically meaningful and equitable applications [4, 6].
Dietary Supplements
In the provided citations, technological solutions primarily concern ingredient sourcing and "design" (extraction, biotransformations) and their formulation (carriers, stimulus-responsive release), as well as the quality of evidence and the logic of health claims. The conclusions below therefore apply to technologies typical for both supplements and ingredients for functional foods and medical products, but in this chapter, they are interpreted from the perspective of supplements as formulations focused on delivering specific bioactive compounds [2].
Sourcing and Biotransformations of Ingredients
As an example of advanced ingredient sourcing, the obtention of a vitamin D3-containing fraction using supercritical CO2 extraction (SFE-CO2) carried out in a pilot plant was indicated [3]. In the same vein of "eco-solutions," the importance of biotransformation technologies, including microbial fermentation, was emphasized as a strategy for enriching and diversifying the value of raw materials [2].
At the mechanism level of biotransformation, the use of microbial enzyme systems, such as β-glucosidases and esterases, for hydrolysis and modification of bound compounds in the starting material was described [2]. A detailed example demonstrated that such processing can significantly increase bioavailability and bioactivity by converting hesperidin into more active hesperetin, and simultaneously lead to the formation of new molecules not present in the raw material, while maintaining mild and "green manufacturing" conditions [2].
Delivery and Stabilization
It was explicitly described that a significant breakthrough in formulation technology has been the development of smart nanocarriers aimed at overcoming in vivo application barriers of bioactive ingredients resulting from their complex physicochemical properties and suboptimal pharmacokinetic profiles [2]. These systems are designed to systematically address key practical problems: low solubility, poor stability, non-specific distribution, and lack of controlled release of active ingredients [2]. Stimulus-responsive release was identified as a particularly promising variant, enabling precise release at target sites through materials that react to pathological microenvironments, e.g., specific pH, enzymes, or redox levels [2].
In the area of functional foods (often transferred to supplements), the role of systems maintaining the quality and stability of oxygen-sensitive ingredients is additionally emphasized, as this strategy is presented as crucial for preserving the quality, potency, and shelf-life of "oxygen-sensitive nutrients" in functional foods and supplements [7].
Clinical Evidence and Claims Logic
In the collected materials, the "evidence–dose–claim" approach strongly resonates as a condition for credibility: it was indicated that it is necessary to establish a scientific evaluation system based on "evidence–dose–claim," which rigorously validates efficacy, safe dose-response relationships, and target populations for each ingredient [2]. From the perspective of supplementation practice, an important example is the synthesis of evidence for a specific substance: the inclusion of a systematic review and meta-analysis assessed by the GRADE methodology for β-hydroxy-β-methylbutyrate (HMB) supplementation was described as an important contribution, and more broadly as a pattern for transparent synthesis and critical assessment of evidence certainty [1]. It was directly emphasized that the path to credible health claims and regulatory acceptance is "paved" with methodological discipline and rigorous grading of evidence quality [1].
Functional Food
The landscape in 2025–2026 shows a shift towards mechanism-based and systemic design, where diet–microbiota–host interactions are analyzed translationally, not just descriptively [1]. Methodologically, maturation through the use of multi-omics, including microbiome profiling, metabolomics, and lipidomics, was emphasized [1]. At the same time, it was consistently indicated that bioactivity is inextricably linked to the food matrix, processing history, and physicochemical stability, which determine bioaccessibility and physiological impact [1].
New Ingredient Classes and Product Concepts
Within the innovation stream of ingredients, postbiotics hold a special place, for which the ISAPP definition has been adopted as standard: "preparation of inanimate microorganisms and/or their components that confers a health benefit on the host" [8]. At the same time, it was emphasized that the adoption of this definition does not invalidate other approaches, and nomenclature remains an "open and ongoing debate" that should not be a major obstacle to research progress [8].
On a mechanistic level, it was described that postbiotics can strengthen epithelial barrier function, regulate innate and acquired immune responses, and modulate host gene expression through pattern recognition receptors and epigenetic modifications [9]. Concurrently, key barriers to clinical translation were identified, including formulation variability, limited understanding of gut-brain interactions, degradation during gastrointestinal transit, and inter-individual variability of the microbiome [10].
Evidence and Health Effect Models
In the area of metabolic personalization, a study was cited where data from continuous glucose monitoring were combined with clinical, behavioral data, and gut microbiota variables to train a gradient-boosted regression model on a cohort of over 800 individuals and 46,898 meals [4]. This model was able to accurately predict individual glycemic response to specific meals, enabling the formulation of personalized dietary recommendations aimed at optimizing metabolic effects [4].
In the field of food tolerance immunology, a mechanism was identified in which Treg cell responses to seed storage proteins may constitute a common pathway leading to oral tolerance [11]. Furthermore, in the area of delivery technology and complex (synbiotic/multi-ingredient) interventions, results showed that a "double emulsion gel" could be used to deliver probiotics and CBD to the gastrointestinal tract [12]. In the SHIME® model, an increase in beneficial bacterial families (Lachnospiraceae and Clostridiaceae), effective delivery, release, and persistence of _L. plantarum_, and increased butyrate and lactate production were reported [12], and quantitative analysis showed effective release of probiotic bacteria from the gel (significantly higher counts after intervention) [12].
Processing and Manufacturing Technologies
In the area of processing, it was emphasized that non-thermal methods have distinct advantages: they allow for the preservation of heat-sensitive ingredients, increase bioavailability by modifying the matrix, and support innovative encapsulation systems that overcome the limitations of classical thermal methods [13]. High-pressure processing (HPP) was indicated as an example, inactivating microorganisms at pressures of 400–600 MPa at room temperature, and in fortified beverages, ensuring better retention of vitamins and polyphenols and sensory attributes than thermal pasteurization [13]. PEF technology was also mentioned, which, through short pulses of high voltage, leads to reversible permeabilization of cell membranes, increases phytochemical extraction, and inactivates microorganisms with an insignificant thermal effect [13].
In biomanufacturing, the development of fermentation control automation was highlighted: embedded edge computing devices (e.g., NVIDIA Jetson AGX Orin) run reinforcement learning algorithms that dynamically optimize bioreactor parameters in real-time (temperature, pH, stirring speed) [14]. On a systemic scale, a review of "precision fermentation" was presented, which integrates strain design, bioprocess engineering, techno-economic feasibility, environmental outcomes, and regulatory readiness in a single framework, addressing gaps in previous reviews focused on products or organisms [15]. From a sustainability perspective, it was indicated that precision fermentation, in typical comparison to conventional animal husbandry and crop systems, requires less land and water, generates lower greenhouse gas emissions, and provides products of consistent quality, free from contaminants [15], although significant barriers were simultaneously emphasized, such as high capital and energy costs, scaling problems, complexity of downstream processing, and regulatory uncertainties and consumer acceptance [15].
Medical Food
In the provided materials, the "medical food" axis is linked to precision nutrition and personalized medicine, where interventions are tailored to the patient's biological profile (genomics, microbiota, metabolic markers), and AI is indicated as the most probable path for implementing precision nutrition in chronic disease management [16, 17]. In clinical practice, data were also cited that clinical trials "progressively demonstrated" better outcomes of therapies based on the patient's genomic profile, gut microbiota, or metabolic markers compared to traditional, generic dietary recommendations [16].
Clinical Evidence and Intervention Examples
In the area of microbiotic interventions, results were cited that supplementation with the _B. BBr60_ strain in a clinical trial was associated with a significant improvement in the lipid profile through increased HDL and decreased total cholesterol [18]. In the same results, an improvement in gastrointestinal health, measured by a decrease in NDI (p = 0.002), was also noted, interpreted as a reduced impact of gastrointestinal discomfort on daily functioning [18], as well as an improvement in gastrointestinal symptoms and emotional states along with a significant decrease in ADS scores (p = 0.000), including symptoms related to alcohol consumption [18].
Concurrently, in regulatory-clinical materials, it was shown that the FDA approved an sBLA for PALYNZIQ (pegvaliase-pqpz), extending the indication to pediatric patients aged 12+ with phenylketonuria [19], and the communication indicated a significantly greater reduction in blood phenylalanine levels at week 72 in the PALYNZIQ arm compared to the "diet only" arm [19]. For ultra-rare diseases, LOARGYS was cited as a therapy targeting the "primary driver" of the disease (persistently elevated arginine in ARG1-D), with accelerated FDA approval based on results from the Phase 3 PEACE study, in which LOARGYS significantly reduced plasma arginine compared to placebo after 24 weeks [20].
In the field of rare diseases and neurology, it was also cited that leucovorin is the first treatment for the rare genetic condition "cerebral folate deficiency" [21], and that Avlayah (weekly intravenous infusion) has been approved for the treatment of neurological manifestations of Hunter syndrome under specific clinical and population conditions [22].
Personalization Technologies and Patient Data
Within the scope of hyper-personalization of nutrition, machine learning applications on data collected "in the field" were indicated: federated learning models can predict glycemic responses based on biometrics from wearable devices, and neural networks decode sensory preferences from social media discussions [14]. In a review context, it was emphasized that AI can transform both precision fermentation towards sustainable biosynthesis of proteins, enzymes, and functional compounds, and hyper-personalization diet systems, integrating genomics, metabolomics, and consumer psychology for "real-time" recommendation matching [14].
Safety and Research Gaps
From the perspective of safety and the development of medical and nutritional products, it was emphasized that before plant proteins can be commercialized in critical applications (e.g., infant nutrition), extensive assessment of allergenic potential is necessary, including in vitro tests, animal studies, and ultimately clinical trials in infants [23]. Additionally, it was indicated that currently no other plant proteins (e.g., pea, lentil, fava bean) are approved for use in infant formulas for children < 1 year of age, and there is a data gap regarding their allergenicity in this age group [23].
In the area of nutrition-pharmacology interactions, it was highlighted that the impact of food processing on drug absorption, metabolism, and subsequent pharmacological activity is "pressing yet insufficiently explored," which implies a significant gap for designing dietary and medical interventions in real nutritional conditions [24].
Breakthrough Technologies
In 2025–2026, "cross-cutting" technologies combine three layers: (1) ingredient manufacturing and modification (extraction and biotransformations), (2) advanced formulation and delivery systems, and (3) data platforms, standardization, and AI supporting design and effect validation. In the collected citations, these elements appear as components of a single, coherent direction of development, in which products are "designed" simultaneously at the level of chemistry, carrier, and clinical evidence [2].
Delivery Engineering
Smart nanocarriers were presented as a critical breakthrough in formulation technology, allowing the overcoming of in vivo barriers resulting from the physicochemical and pharmacokinetic properties of bioactive ingredients [2]. A set of problems that these systems are intended to solve was also defined, including solubility, stability, non-specific distribution, and release control [2]. Stimulus-responsive release, through materials reacting to pH, enzymes, or redox in pathological microenvironments, is particularly promising, as it is intended to enable precise release at the target site [2].
Biotransformations and Fermentation
In the area of biotransformations, microbial fermentation was indicated as an "eco-friendly" solution for enriching and diversifying the value of raw materials [2]. Mechanistically, the role of microbial enzymes, such as β-glucosidases and esterases, in the hydrolysis and modification of bound components in the starting material was described [2]. Consequently, the possibility of increasing bioavailability and bioactivity was reported, including the transformation of hesperidin into more active hesperetin and the generation of new molecules under mild conditions consistent with "green manufacturing" principles [2].
Personalization and AI Platforms
In the cited sources, the concept of a development and personalization platform appears, whose "cornerstones" include multi-dimensional individual assessment, adaptive interventions and feedback systems, and "AI-powered smart formulation and design" [2]. In a similar vein, it was indicated that personalized health management is to be achieved through an integrated data and product platform, analyzing individual differences and providing tailored solutions [2].
As an example of digital R&D acceleration, AMBROSIA was presented, an agent-based AI platform integrating biological data with "intelligent research operations," intended to accelerate product development, optimize extract characterization, and identify new target markets for existing ingredients [5]. In the manufacturing area, edge-RL was also indicated for dynamic real-time optimization of bioreactor parameters, forming the technological basis for more stable and efficient fermentation processes [14].
Standardization of Evidence
The materials indicated two complementary axes of standardization: data standardization and evidence standardization. From the data perspective, it was emphasized that standardization of formats, preprocessing, and analytical frameworks is essential for creating reliable, reproducible, and transferable models [4]. From the clinical evidence perspective, the importance of the GRADE approach (using HMB as an example) was highlighted as a model for transparent synthesis and assessment of evidence certainty [1], and it was also emphasized that regulatory acceptance and credible health claims require methodological discipline and rigorous grading of evidence quality [1]. Additionally, the "evidence–dose–claim" framework was directly postulated as a scientific system for evaluating efficacy, safety, and suitability for target populations [2].
Trends
Across 2025–2026, several trends can be identified that consistently recur in the cited sources and connect supplements, functional foods, and medical products into a single innovation ecosystem.
The first trend is the transition from "descriptive cataloging" to hypothesis-driven, mechanistic, and translational research, which is explicitly described as a field "in active transition" [1]. The second is the institutionalization of multi-omics approaches as a tool for methodological maturation of research on bioactivity and diet–microbiota–host interactions [1].
The third trend is "matrix-aware" product design, treating the food matrix, processing, and stability as determinants of bioaccessibility and physiological effects, rather than as second-order technological details[1]. The fourth is a shift from probiotics as the sole axis of microbiotic innovation towards postbiotics, with an attempt to unify the definition through the ISAPP standard, while recognizing that the nomenclature remains an open debate[8].
The fifth trend is the increasing role of AI in personalization and R&D: from predictive models of glycemic response built on CGM and microbiota variables, to agent platforms integrating biological data with research operations to accelerate product development[4, 5]. The sixth trend is automation and "process intelligence" in biomanufacturing, including reinforcement learning optimizing bioreactor parameters in real time, which supports the stability and efficiency of fermentation processes producing functional ingredients[14].
Challenges
The identified challenges for 2025–2026 are translational, regulatory, and engineering in nature, and many of them concern what happens "between" the laboratory, production, market, and clinical practice.
In the area of microbiota and postbiotics, translational barriers include formulation variability, degradation in the gastrointestinal tract, limited understanding of gut-brain interactions, and inter-individual variability of the microbiome, which complicates both study design and the predictability of effects in the population[10]. Simultaneously, in the area of new ingredient classes, it was indicated that nomenclature (e.g., postbiotics) remains an open debate, although it should not block progress, which in practice means the need for parallel work on definition, quality standards, and evidence criteria[8].
In the area of data and AI, it was indicated that standardization of data formats, preprocessing, and analytical frameworks is a condition for reproducible and transferable models[4]. At the same time, it was emphasized that further improvement of methodology and responsible implementation are crucial to translate innovations into clinically significant and equitable applications[6].
In the area of biomanufacturing and "precision fermentation," barriers related to capital and energy costs, scaling problems, complexity of downstream processing, consumer acceptance, and regulatory uncertainty were described[15]. From the perspective of feasibility and sustainability, it was also emphasized that strain selection, process design, and downstream processing strongly influence sustainability and commercial viability, setting the tone for technology development priorities in the coming years[15].
In the area of safety and clinical practice, a research gap was identified regarding the impact of food processing on drug absorption, metabolism, and pharmacological activity, deemed urgent but insufficiently studied[24]. In critical nutritional applications, it was also indicated that a full assessment of allergenicity (in vitro, animals, finally infant studies) is necessary before commercializing new plant proteins, and at the same time, the lack of approvals for alternative plant proteins in infant formulas < 1 year and the lack of allergenicity data in this group were noted[23].
Implications
The cited sources reveal practical implications that can be organized around three questions: how to produce ingredients, how to deliver them in the body, and how to prove and scale their effects.
For producers, two parallel technological directions are key: on the one hand, advanced sourcing and biotransformations (SFE-CO2 for fractions with vitamin D3 and microbial fermentation/enzymes that can increase bioactivity and generate new molecules)[2, 3], and on the other hand, the development of formulation systems that solve problems of solubility, stability, and controlled release, including stimulus-responsive release in target microenvironments[2]. For R&D departments, an additional "accelerator" are AI platforms integrating biological data with research operations to accelerate product development and characterize extracts, an example of which is AMBROSIA[5].
For clinicians and dietary teams, the growing support of data and models is important: it was shown that models based on CGM combined with clinical, behavioral, and microbiota data can accurately predict glycemic response and enable personalization of recommendations[4]. The position that therapies and recommendations based on genomic profile, microbiota, or metabolic markers yield better results than generic recommendations, and that AI is the most likely path to implementing precision nutrition in the management of chronic diseases, is also supportive[16, 17].
For regulators and quality compliance teams, the necessity of "evidence–dose–claim" frameworks was directly formulated, and the role of rigorous assessment of evidence certainty (GRADE) was emphasized as the foundation for reliable health claims and regulatory acceptance[1, 2]. For data systems and digital surveillance, it was emphasized that standardization of data formats and pipelines is essential for reproducible and transferable models, and responsible implementation is a condition for translating into clinically significant and equitable applications[4, 6].
To synthetically show how the implications form a "decision map" at the product development stage, the table below compiles the four most frequently cited innovation axes along with typical benefits and barriers, exactly as they emerge from the cited sources.
Perspectives
The cited materials indicate that the most probable development in the coming years will involve further integration of mechanistic design, evidence standardization, and digital personalization platforms. On the one hand, the field of research on bioactive ingredients has already been described as moving towards hypothesis-driven, mechanistic, and translational paradigms, supported by multi-omics[1]. On the other hand, it was emphasized that bioactivity is not independent of the matrix and processing, which suggests further intensification of "matrix-aware" approaches, where non-thermal processing and delivery systems will be co-designed with the biological objective[1, 13].
In the area of digitalization, the expected trajectory is twofold: (1) the development and adoption of platforms integrating data and product to deliver tailored health solutions, and (2) the use of AI to shorten R&D cycles and automate manufacturing processes. This trajectory is directly supported by the thesis that personalized health management is to materialize through an integrated data and product platform[2], as well as the example of the AMBROSIA platform, which combines biological data with research operations to accelerate product development[5]. At the same time, the cited works indicate the necessary conditions for this transformation: data standardization and responsible implementation, so that innovations translate into clinically significant, reproducible, and equitable applications[4, 6].
Finally, in the area of regulation and evidence, further strengthening of assessment rigor is most probable, as both the necessity of "evidence–dose–claim" frameworks[2] and the central role of grading the quality of evidence (GRADE) on the path to reliable health claims and regulatory acceptance[1] were emphasized.
