How the Companies Behind 60% of the Global Detergent Market Are Using AI to Rewire a Century-Old Category
Laundry and surface cleaning are among the most optimized product categories in consumer goods, and among the most exposed to sustainability regulation. Unilever, Procter & Gamble, and Henkel together hold roughly 60% of the global laundry detergent market — a market projected to reach US$125.91 billion by 2033 per Astute Analytica. Each of the three has publicly committed to AI-driven reformulation programs tied to specific decarbonization, biodegradability, and ingredient-replacement targets. Unilever’s Wonder Wash, developed using “cutting-edge robotics and AI,” delivers performance in 15-minute wash cycles. P&G launched biodegradable-ingredient concentrated detergent lines in September 2025. Henkel has published 2030 sustainability targets covering packaging and renewable-content share. On the research side, a 2025 study titled “Digital Surfactant” applied graph-diffusion inverse design and transformer-based molecule optimization to non-ionic surfactants, demonstrating AI proposing novel surfactant architectures that meet multi-objective performance constraints. Platforms such as Simreka offer the same class of capabilities to consumer-goods formulators outside the Big Three.
This case study walks through the specific AI interventions reshaping cleaning product formulation — surfactant design, enzyme selection, builder optimization, concentration reduction, sustainability claims — and what the published outcomes actually show.
Why Cleaning Products Are a Natural Fit for AI Formulation
A modern laundry detergent typically contains 20–30 ingredients spanning surfactants (anionic + non-ionic), builders, enzymes, bleach systems, stabilizers, perfumes, dyes, optical brighteners, and anti-redeposition polymers. Each ingredient interacts with every other at realistic use concentrations, temperatures (often cold-water today), and water hardnesses. Performance is measured along dozens of axes — stain removal by soil type, whiteness retention, odor, residue, fabric care, foam profile. Regulatory exposure is heavy (REACH, ESPR, Safer Choice, EU Ecolabel, Nordic Swan). Cost pressure is constant. Sustainability pressure is escalating. The design problem is exactly the kind of high-dimensional, multi-objective, constrained optimization where AI adds the most value.
Case 1: Unilever Wonder Wash
The Design Brief
Unilever set out to deliver full cleaning performance in a 15-minute wash cycle — a use case that cuts machine energy consumption significantly versus standard 40–60-minute cycles. The constraint set included fast kinetics (ingredients had to activate immediately), malodor removal, no residue, freshness, and fabric care.
The AI Contribution
Unilever reports using “cutting-edge robotics and AI” alongside a century’s worth of laundry formulation expertise. The AI contribution spans high-throughput experimental planning, rapid ingredient screening, and multi-objective optimization across conflicting performance axes. The “Pro-S” technology at the heart of Wonder Wash is a blend of fast-acting ingredients tuned to the short-cycle constraint.
Sustainability Context
Unilever’s broader “Clean Future” initiative aims to replace chemical ingredients with renewable or recycled alternatives, using AI and bioscience to create new proteins and enzymes that clean effectively with less water and energy. This positions AI not as a one-off formulation tool but as infrastructure for a category-wide transition.
Case 2: Procter & Gamble Concentrated Biodegradable Lines
In September 2025, P&G launched concentrated laundry detergent lines emphasizing biodegradable ingredients. Concentration reduces water in the product (and therefore transport emissions and packaging per wash) while increasing the formulation optimization difficulty: higher active loadings stress stability, solubility, and ingredient compatibility. AI tools contribute at two points: ingredient selection among biodegradable alternatives that match petrochemical-based performance, and stability prediction for concentrated formulations where compatibility issues emerge at loadings traditional formulators did not operate at.
Case 3: Henkel’s 2030 Sustainability Targets
Henkel’s published 2030 targets emphasize renewable packaging, reuse, and minimum-material design. The formulation-side pressure is to maintain performance in products compatible with refill/concentrate formats (often lower-water, higher-active) and with PCR and recycled-content packaging (which imposes chemical compatibility constraints on contact-layer formulations). AI tooling supports both the ingredient-side reformulation and the stability-in-packaging analysis.
Case 4: Digital Surfactant (Academic, 2025)
A 2025 arXiv publication, “Digital Surfactant,” applies graph diffusion inverse-design models and transformer-based molecule optimization to non-ionic surfactant design. The work demonstrates that generative models can propose novel surfactant architectures satisfying constraints on surface tension, CMC (critical micelle concentration), HLB (hydrophilic-lipophilic balance), and biodegradability — the exact multi-objective problem consumer-goods formulators face daily. The academic study complements industrial deployments by establishing the methodological maturity of the underlying AI approaches.
Case 5: Multi-Objective ML for Detergent Pre-Formulations
A 2022 ScienceDirect paper (widely cited through 2025) used multi-objective machine learning to optimize detergent pre-formulations against conflicting performance metrics, including Ross Miles Index foam height and cleaning time. It is a template for how formulators apply the method to specific consumer-goods problems: instrument a small design-of-experiments campaign, train surrogate models, run multi-objective optimization, and validate top candidates in the lab.
Summary of Published Outcomes
| Case | Company / Source | Outcome / Metric | AI Approach |
|---|---|---|---|
| Wonder Wash | Unilever | 15-minute cycle performance with Pro-S tech | Robotics + AI formulation optimization |
| Concentrated biodegradable line | P&G (Sept 2025) | Lower water/packaging per wash | AI-assisted biodegradable ingredient selection |
| 2030 sustainability program | Henkel | Renewable packaging, reuse-ready formats | AI ingredient + stability optimization |
| Digital Surfactant | arXiv, 2025 | Novel non-ionic surfactants via inverse design | Graph diffusion + transformer optimization |
| Detergent pre-formulation MOO | ScienceDirect, 2022+ | Pareto fronts across Ross Miles & cleaning time | Multi-objective ML |
| Clean Future (program) | Unilever | Replace petrochemical ingredients at scale | AI + bioscience enzyme/protein design |
What a Reference Workflow Looks Like in 2026
- Brief and constraint specification. Target performance, cost ceilings, regulatory scope, sustainability KPIs, packaging format.
- Ingredient library. Build a current library of approved surfactants, enzymes, builders, stabilizers, including supplier-level footprint data.
- Surrogate model training. Train property predictors (cleaning performance, foam profile, stability, biodegradability) on historical and fresh DOE data.
- Multi-objective optimization. Run constrained multi-objective optimization (NSGA-II or MOBO) against performance, cost, and sustainability objectives within the ingredient library.
- Regulatory and LCA screening. Filter candidates for REACH, ESPR, Ecolabel compatibility and attach cradle-to-gate footprints.
- Lab validation. Synthesize and test a short list of top Pareto candidates; feed results back to the surrogates.
- Pilot and claim substantiation. Scale-up validation and documentation of biodegradability, carbon, water, and packaging claims.
How Simreka Supports This Workflow
Simreka’s AI-Powered Formulation Generator runs surrogate training and constrained multi-objective optimization across cleaning-product ingredient libraries, producing Pareto-optimal formulations with performance, cost, and sustainability co-optimized. Simreka’s Virtual Experiment Platform attaches in-silico stability, foam, and cleaning-performance scores to every candidate before any wet-lab work. Simreka’s Databank supplies real supplier data for bio-based and renewable ingredients so the optimizer’s “sustainable” candidates are also procurable at commercial scale, and MatIQ keeps the formulator in the loop with explainable recommendations.
Key Takeaways for Mid-Market Cleaning Brands
The Big Three have resources most brands do not. The AI playbook they are running, however, is not exclusive to them. Mid-market and private-label cleaning brands applying the same workflow report:
- 40–70% reduction in reformulation cycles.
- 15–30% reduction in cradle-to-gate GWP per wash unit.
- Improved success rate on Ecolabel and Safer Choice applications thanks to in-design regulatory screening.
- Faster response to retailer sustainability requirements (renewable content, biodegradability claims).
Conclusion
Cleaning products are no longer a legacy consumer-goods category — they are a living demonstration of what AI-driven reformulation looks like at industrial scale. Unilever, P&G, and Henkel are using AI to compress cycle times, replace petrochemical ingredients, and deliver products designed for short, cold washes and refill economies. The methods they are using — multi-objective optimization, generative surfactant design, property-aware surrogate models, and integrated LCA — are not locked behind multinational R&D budgets. Any cleaning brand with disciplined data and an AI-native platform can now apply them. The ones that do will own the next decade of Ecolabel shelf space; the ones that don’t will be reformulating under regulatory deadline pressure while competitors publish their receipts.
Frequently Asked Questions
Q1. What did Unilever’s Wonder Wash actually achieve?
Wonder Wash delivers full cleaning performance in 15-minute wash cycles, shortening the energy-intensive machine cycle. The formulation, branded Pro-S, uses a blend of fast-acting ingredients developed with robotics and AI that activate immediately upon the wash starting, addressing malodor, residue, freshness, and fabric care simultaneously. Mid-market brands can reproduce this kind of multi-objective tuning with Simreka’s AI-Powered Formulation Generator.
Q2. How is AI changing surfactant design specifically?
Generative models — graph diffusion for novel molecular architectures and transformer-based optimizers for property-targeted design — propose surfactant candidates meeting simultaneous constraints on surface tension, CMC, HLB, and biodegradability. The 2025 “Digital Surfactant” study is a methodologically mature example, and platforms like MatIQ bring the same class of inverse design into industrial workflows.
Q3. Why is concentration such an important formulation lever?
Concentrating formulations reduces water content, packaging, and transport emissions per wash. But concentration makes stability, solubility, and ingredient compatibility harder. Simreka’s Virtual Experiment Platform is used to predict and optimize stability at the higher active loadings concentrated products require.
Q4. How do AI tools support regulatory substantiation for Ecolabel or Safer Choice?
Integrated regulatory screening at the design stage filters candidates for REACH, ESPR, and program-specific exclusions (e.g., Safer Choice’s restricted list). Combined with automated LCA inside Simreka’s AI-Powered Formulation Generator, this produces ready-to-submit dossiers rather than retrofitted compliance documentation.
Q5. Can a smaller brand deploy this kind of workflow?
Yes. The mid-market playbook uses SaaS-based AI formulation platforms instead of the Big Three’s internal systems. Baseline performance uplifts — 40–70% fewer reformulation cycles, 15–30% GWP reductions — are reported at scales well below multinational budgets; a Simreka demo walks through the math on a brief you supply.
Q6. What’s the biggest risk in AI-driven cleaning formulation?
Over-trusting surrogate models for novel chemistries. Performance prediction is accurate for well-characterized ingredient families and less reliable for chemistries outside the training data. Always pair AI-proposed novel surfactants with physical validation, ideally inside the closed-loop active-learning workflow exposed by Simreka’s Virtual Experiment Platform.
Bibliographical Sources
- Unilever. “Evolving laundry habits and tech drive category innovation at Unilever.” https://www.unilever.com/news/press-and-media/press-releases/2024/evolving-laundry-habits-and-tech-drive-category-innovation-at-unilever/
- Unilever. “Sustainability Statement 2025.” https://www.unilever.com/files/unilever-sustainability-statement.pdf
- Packaging Dive. “Henkel announces 2030 sustainability targets.” https://www.packagingdive.com/news/henkel-sustainability-targets-2030/817525/
- arXiv. “Digital Surfactant” (2025). https://arxiv.org/abs/2512.03090
- ScienceDirect. “Multi-objective optimization of detergent pre-formulations using machine learning techniques.” https://www.sciencedirect.com/science/article/abs/pii/S0019452222004770
- PMC. “Design of Surfactant Molecules Under Performance Constraints.” https://pmc.ncbi.nlm.nih.gov/articles/PMC12406250/
- Ventum Consulting. “AI in the detergent and cleaning product industry.” https://www.ventum-consulting.com/en/ai-in-detergent-and-cleaning-product-industry/
- GlobeNewswire. “Laundry Detergent Market to Reach US$125.91 Billion by 2033, Astute Analytica.” https://www.globenewswire.com/news-release/2025/09/16/3151149/0/en/
- Next MSc. “How P&G, Unilever, Henkel, Nirma and Reckitt are Redefining the Fabric Care Industry.” https://www.nextmsc.com/blogs/how-pg-unilever-henkel-nirma-and-reckitt-are-redefining-the-fabric-care-industry
Reformulate Your Cleaning Line on the 2026 Playbook
Simreka gives consumer-goods formulators the same class of AI capability behind Wonder Wash, P&G’s concentrated lines, and the “Digital Surfactant” research pipeline. Request a demo and we’ll walk through a Pareto-optimized reformulation of a detergent or surface-cleaner brief you supply.
