A significant new package of EU cosmetic ingredient amendments has moved into the formal regulatory process.

On 8 July 2026, the European Union notified the World Trade Organization of draft Commission Regulation G/TBT/N/EU/1219, proposing amendments to Annexes II, III and V of Regulation (EC) No 1223/2009 on cosmetic products. The proposal combines newly harmonised CMR classifications with the outcome of several recent Scientific Committee on Consumer Safety assessments. 

The package has been described within the cosmetics industry as a combination of Omnibus CMR IX and Omnibus Ingredients III. It would introduce new ingredient prohibitions, establish concentration limits for previously unregulated substances, revise the conditions for certain preservatives and remove the remaining cosmetic preservative exemptions for mercury compounds.

However, one point must remain clear from the beginning:

The notified text is still a draft. It has not yet been adopted or published in the Official Journal of the European Union.

The draft still contains placeholders for dates and annex entry numbers. Cosmetic companies should therefore begin preparing for the proposed changes, but final legal decisions should be based on the adopted Regulation when it is published. 

What the draft would change

The proposed Regulation covers four main areas:

1. New Annex II prohibitions resulting from recent CMR classifications.
2. Prohibitions based on SCCS opinions for benzophenones, hair dyes and prostaglandins.
3. New Annex III restrictions for CBD, BHA and nano-hydroxyapatite.
4. Revised Annex V conditions for Butylparaben and Propylparaben, together with removal of mercury-containing preservatives.

The commercial impact will vary considerably by product category. Eyelash serums, CBD skincare, children’s cosmetics, oral-care products, non-oxidative hair dyes and formulas containing BHA are likely to require the most immediate attention.

New CMR-related prohibitions

Article 15 of the EU Cosmetics Regulation generally prohibits substances classified as carcinogenic, mutagenic or toxic for reproduction under the CLP Regulation, unless the strict conditions for an exemption are fulfilled.

Commission Delegated Regulation (EU) 2025/1222 introduced new harmonised classifications. According to the cosmetic draft, no requests were submitted to continue cosmetic use of the affected CMR substances under the exemption provisions. The substances that are not already covered would therefore be added to Annex II of the Cosmetics Regulation. 

The proposed Annex II additions include substances such as:

• Ozone
• Dinitrogen oxide, commonly known as nitrous oxide
• Trimethyl phosphate
• Tetrahydrofurfuryl methacrylate
• Fluoroethylene
• 2-phenylpropene, also known as α-methylstyrene
• 2,3-epoxypropyl isopropyl ether
• Several additional industrial chemical substances covered by the new classifications

These may not appear frequently as intentionally added ingredients in mainstream finished cosmetics. Nevertheless, companies should consider their possible presence in raw materials, processing aids, propellants, monomers, impurities or supplier blends.

Benzophenone-1 would be prohibited

Benzophenone-1 is used as a light stabiliser to protect cosmetic formulations from deterioration caused by ultraviolet radiation.

The SCCS concluded in March 2025 that Benzophenone-1 could not be considered safe for use as a cosmetic light stabiliser. The Committee identified genotoxicity concerns and evidence of endocrine activity, including estrogenic and weak anti-androgenic activity. The draft therefore proposes adding Benzophenone-1 to Annex II as a prohibited cosmetic substance. 

This could affect more than products where Benzophenone-1 is presented as an active or prominent ingredient. It may also appear in:

• Fragrance compositions
• Colour cosmetics
• Formulation stabiliser systems
• UV-sensitive product bases
• Supplier blends intended to protect colour or fragrance stability

Finished-product manufacturers will need to check both their own formulas and the detailed composition of purchased blends.

Benzophenone-2 would also be prohibited

Benzophenone-2 has been used as a UV filter, light stabiliser and fragrance-related ingredient.

The SCCS was unable to conclude that Benzophenone-2 was safe because its genotoxic potential could not be excluded. The Committee also noted limited or unavailable repeated-dose and reproductive-toxicity data, together with clear evidence of estrogenic activity.

The draft responds by proposing a complete Annex II prohibition rather than a concentration restriction. 

This distinction matters. The proposal does not create a lower acceptable use level for Benzophenone-2. Under the notified text, reformulation or removal would be required.

Basic Brown 16 and Basic Blue 99 would be banned

The draft proposes prohibiting two colourants used in non-oxidative hair dye products:

• Basic Brown 16
• Basic Blue 99

The SCCS concluded that the weight of evidence for Basic Brown 16 indicated mutagenic potential. For Basic Blue 99, the available evidence indicated potential genotoxicity. Both substances would consequently be added to Annex II. 

Hair-colour manufacturers should review not only finished formulas but also premixed colour systems and supplier shade blends. A colour mixture may contain one of the affected dyes even where it is not obvious from the commercial name of the blend.

Prostaglandins and their analogues face a broad class prohibition

One of the most commercially important parts of the draft concerns prostaglandins and prostaglandin analogues used in eyelash and eyebrow enhancement products.

The proposal is not limited to one named substance. It would add “prostaglandins and their analogues” as a broad Annex II entry.

The SCCS reviewed ingredients including ethyl tafluprostamide, methylamido-dihydro-noralfaprostal and isopropyl cloprostenate. It concluded that these substances raised safety concerns because of their pharmacological activity, even at very low concentrations, and their potential to cause serious undesirable effects, particularly involving ocular health.

The Committee also highlighted insufficient evidence to exclude reproductive and developmental toxicity concerns. This was considered especially relevant because many users of lash and brow products are women of childbearing age. No conditions of cosmetic use could be established under which the assessed substances were considered safe. 

This means lash and brow brands should not limit their review to ingredients explicitly labelled as “prostaglandin.”

A proper screening exercise should include:

• INCI names
• CAS numbers
• Supplier trade names
• Eyelash-conditioning active blends
• Ingredients making growth or density claims
• Compounds with prostaglandin-like pharmacological activity

The breadth of the proposed wording could make this one of the most disruptive changes in the entire package.

Mercury-containing preservatives would lose their remaining exemptions

Mercury and its compounds are already generally prohibited under Annex II. However, Annex V currently contains limited preservative allowances for Thiomersal and certain phenylmercuric salts.

The draft would remove Annex V entries 16 and 17 and revise the general Annex II mercury entry so that it no longer refers to exceptions for the special cases listed in Annex V. 

The SCCS concluded that the currently permitted preservative uses could not be considered safe. Its assessment identified an inadequate margin of safety based on renal toxicity, while the genotoxicity evidence remained unclear.

These preservatives are unlikely to be widely used in modern cosmetic portfolios, but companies should still check:

• Legacy eye-area products
• Old formulas still marketed in limited volumes
• Specialist preservative systems
• Long-standing supplier specifications
• Products acquired through mergers or brand purchases

CBD would become expressly restricted at 0.19%

The proposal does not impose a general ban on Cannabidiol.

Instead, CBD would be added to Annex III and permitted in:

• Leave-on products
• Rinse-off products
• Oral-care products

The proposed maximum concentration would be 0.19% in the ready-for-use cosmetic product.

The presence of delta-9-tetrahydrocannabinol as an impurity would be limited to 0.00025%, equivalent to 2.5 ppm. 

This would give CBD a clearer ingredient-specific regulatory framework, but it would also introduce demanding impurity controls.

A simple supplier statement declaring a raw material “THC-free” may not provide sufficient evidence. Brands and Responsible Persons may need:

• A quantitative THC specification
• A validated analytical method
• Appropriate limits of detection and quantification
• Batch-specific or risk-based certificates of analysis
• Confirmation of the botanical source and extraction route
• Calculation of CBD concentration in the finished product
• Assessment of every cannabis-derived ingredient contributing THC

The 0.19% limit applies to the ready-for-use finished cosmetic, not merely to the CBD concentration in the purchased raw material.

BHA would be limited to dermal products at 0.07%

Butylated Hydroxyanisole, commonly known as BHA, is used as an antioxidant and may also be present in fragrance compositions or stabilised raw-material blends.

The draft would permit BHA at a maximum concentration of 0.07% in leave-on and rinse-off cosmetic products.

It would not be permitted in:

• Oral-care products
• Products that may expose the end user’s lungs through inhalation

The SCCS assessment addressed dermal use and did not support oral or inhalation-related applications. 

The practical challenge is that BHA may enter a formula indirectly through:

• Fragrances
• Essential-oil blends
• Oil-soluble active preparations
• Colourant dispersions
• Stabilised oils
• Supplier antioxidant packages

Companies may therefore need full compositional information rather than relying only on the finished product’s intentionally added ingredient list.

Sprays, aerosols, powders and other products capable of generating inhalable exposure deserve particular attention.

Nano-hydroxyapatite receives updated permitted conditions

The nano-hydroxyapatite amendment should not be presented simply as a new restriction or prohibition.

The proposal would permit Hydroxyapatite in nano form at:

• Up to 29.5% in toothpaste
• Up to 10% in mouthwash

These concentrations would be subject to strict particle specifications. The permitted material must be composed of rod-shaped particles, with at least 87% by particle number having an aspect ratio of three or less. The remaining particles must have an aspect ratio not exceeding nine.

The particles must also be uncoated, not surface modified and have a specified maximum length of approximately 122 ± 43 nanometres. Applications that may expose the lungs through inhalation would not be permitted. 

For oral-care manufacturers, this may create useful formulation opportunities. At the same time, compliance cannot be demonstrated through a document that simply states “nano-hydroxyapatite.”

Supplier evidence should cover:

• Particle shape
• Particle-number distribution
• Aspect-ratio distribution
• Maximum particle length
• Coating status
• Surface-modification status
• Test methods and representative batch data

These characteristics should also be reflected in the Product Information File and Cosmetic Product Safety Report.

Butylparaben would receive separate and tighter conditions

The draft would separate Butylparaben and Propylparaben into individual Annex V entries.

For Propylparaben, the familiar maximum of 0.14%, expressed as acid, would broadly continue, subject to the existing combined paraben limits and restrictions involving leave-on products for the nappy area of children under three.

Butylparaben would receive its own entry. The general maximum of 0.14% would remain, but products intended for children under ten would be subject to tighter product-specific limits:

• 0.14% in rinse-off products
• 0.002% in leave-on products
• 0.092% in oral-care products

The SCCS conclusion would not apply to sprayable products, including mouth sprays, capable of exposing the lungs. The draft therefore proposes prohibiting Butylparaben in those applications. 

The 0.002% limit for children’s leave-on products is particularly significant. A formula that complies with the general Butylparaben limit could still fail the more specific condition when intended or marketed for children under ten.

Companies should review more than products explicitly labelled “children’s cosmetics.” Relevant products may include:

• Family skincare
• Sensitive-skin lotions
• Multi-age personal-care products
• Products visually marketed toward children
• Products included in children’s gift sets
• Leave-on products routinely promoted for use by the whole family

The intended user group should be assessed through the complete product presentation, not only a single statement on the label.

Two different transition systems are proposed

The draft creates an important distinction between the CMR amendments and the remaining ingredient measures.

CMR-related measures

The amendments linked to classifications under Delegated Regulation (EU) 2025/1222 are intended to apply from 1 February 2027.

Companies should not assume that the longer general transition periods will apply to these CMR prohibitions. 

Other ingredient measures

For the non-CMR ingredient changes, the notified annex proposes:

• A 12-month transition after entry into force for placing non-compliant products on the Union market.
• A 24-month transition after entry into force for continuing to make those products available on the Union market.

The first deadline concerns the initial placement of a product on the EU market. The later deadline concerns continued distribution and sale within the market.

The Regulation would enter into force on the twentieth day following publication in the Official Journal. Because publication has not yet occurred, the final calendar deadlines for these measures cannot currently be calculated. 

The WTO notification process indicates a proposed adoption timetable during late 2026. That timetable remains provisional until the final Regulation is adopted and published. 

What cosmetic companies should do now

Waiting for the final publication before beginning any review could leave insufficient time for reformulation, testing and supply-chain changes.

A practical preparation programme should begin with six activities.

1. Screen the complete portfolio

Search formulas using INCI names, chemical names and CAS numbers.

Do not limit the review to ingredients deliberately added by the finished-product manufacturer. Include fragrance mixtures, colour blends, active preparations, preservatives, impurities and processing-related substances.

2. Prioritise high-impact product groups

The first review should cover:

• Eyelash and eyebrow serums
• CBD skincare and oral-care products
• Children’s leave-on products containing Butylparaben
• Non-oxidative hair dyes
• Products containing Benzophenone-1 or Benzophenone-2
• Formulas containing BHA
• Nano-hydroxyapatite oral-care products
• Legacy products containing mercury preservatives

3. Request more precise supplier declarations

Broad declarations such as “EU compliant,” “THC-free” or “cosmetic grade” may not address the new requirements.

Supplier questionnaires should request substance-specific concentration, impurity and particle-characterisation data.

4. Assess reformulation consequences

Removing an ingredient can affect more than regulatory compliance.

Reformulation may require:

• Stability testing
• Preservative efficacy testing
• Packaging compatibility studies
• Colour-performance testing
• Updated exposure calculations
• Claim substantiation review
• New supplier qualification
• Revised manufacturing instructions

5. Update compliance documentation

Affected products may require changes to:

• Cosmetic Product Safety Reports
• Product Information Files
• Raw-material specifications
• Safety assessment calculations
• CPNP information
• Labels and warnings
• Internal regulatory databases
• Distributor and Responsible Person documentation

6. Separate regulatory dates by substance

Do not create one general deadline for the entire package.

CMR substances, prohibited SCCS-assessed ingredients and restricted ingredients may follow different compliance pathways. Each substance should have its own internal regulatory record, deadline and product-impact assessment.

Final perspective

This draft is more than another routine cosmetics annex update.

It could remove entire ingredient classes from cosmetic use, particularly prostaglandins used in lash and brow products. It would also create precise new limits for CBD and BHA, materially tighten the use of Butylparaben in children’s products and demand much stronger nanomaterial evidence for Hydroxyapatite.

For regulatory and formulation teams, the main lesson is simple: compliance will depend increasingly on what is hidden inside raw-material blends, supplier preparations and impurity profiles, not only on the ingredients deliberately added at the finished-product stage.

The final Regulation may still differ from the WTO-notified draft. Companies should therefore prepare against the proposed requirements while keeping final reformulation, withdrawal and market-transition decisions aligned with the adopted Official Journal text.

Regulatory Updates Matter Only When Teams Can Act on Them

Changes such as these do not affect regulatory teams alone. They influence formulation decisions, raw-material selection, supplier documentation, product safety assessments, testing programmes, claims and market-transition planning.

That is where OnlyTRAININGS helps.

OnlyTRAININGS provides advanced, expert-led training for cosmetic formulators, R&D professionals, regulatory specialists, product developers and technical decision-makers who need more than a summary of changing regulations.

Our cosmetics training programmes focus on the practical questions professionals face every day:

• How should an affected formulation be screened?

• Which supplier documents are no longer sufficient?

• When is reformulation actually required?

• What evidence should be added to the CPSR and PIF?

• How can regulatory, safety and performance requirements be managed together?

Whether you are developing a new cosmetic product, reviewing an existing portfolio or preparing for upcoming EU requirements, OnlyTRAININGS helps your team move from regulatory awareness to informed technical action.

Build Stronger Cosmetic Formulation and Compliance Decisions

Explore advanced cosmetics training programmes covering formulation, ingredient safety, regulatory compliance, product performance, scale-up and troubleshooting.

Explore Cosmetics Trainings at OnlyTRAININGS

OnlyTRAININGS | Where Expertise Matters Most.

Regulatory status checked on 17 July 2026. This article discusses a draft regulatory measure and should not be treated as legal advice.

EU cosmetics regulation 2026, EU cosmetic ingredient restrictions, Omnibus CMR IX, Omnibus Ingredients III, cosmetic ingredient bans EU, prostaglandins cosmetics ban, CBD cosmetics EU limit, BHA cosmetics restriction, Butylparaben children products, Benzophenone-1 ban, Benzophenone-2 ban, nano hydroxyapatite cosmetics

Responsible sourcing is no longer limited to choosing a supplier with the right certificate.

For a global chemical company, it now involves tracing raw materials back through complex supply networks, examining environmental and human-rights risks, supporting suppliers that need to improve and determining whether renewable or circular alternatives can be introduced without disrupting production.

BASF’s latest responsible-sourcing activities show how this transition is taking shape across palm oil, coconut oil, castor oil and circular chemical feedstocks.

The company’s approach also demonstrates something procurement teams across the chemical industry are increasingly discovering: sustainability cannot be managed as a separate reporting exercise. It must be built into supplier selection, purchasing conditions, risk assessments, technical decisions and long-term supply relationships.

Traceability Comes Before Improvement

A company cannot properly manage a raw-material risk when it cannot identify where that material originated.

This is particularly important for palm oil and palm kernel oil. These renewable raw materials are widely used to manufacture ingredients for personal care products, detergents, cleaning formulations and other chemical applications. Their supply chains can extend from small agricultural producers to mills, processors, traders and chemical manufacturers.

In 2025, BASF traced 97.7% of its palm-based raw-material volume back to the oil-mill level. The company describes this level of visibility as an important contribution to supply-chain transparency and risk management.

Traceability does not automatically prove that every stage of the supply chain is sustainable. It does, however, give procurement teams a clearer foundation for identifying high-risk locations, examining sourcing practices and engaging suppliers where problems are found.

Without that visibility, sustainability commitments remain difficult to verify.

Certification Progress Has Not Followed a Straight Line

BASF’s 2025 results also reveal the practical difficulty of meeting responsible-sourcing targets when certified raw materials are not available in sufficient quantities.

The company reported that 79.2% of the palm oil and palm kernel oil it purchased during 2025 came from certified sources. That was significantly below the 98.1% recorded in 2024.

BASF attributed the reduction partly to limited availability of Roundtable on Sustainable Palm Oil-certified palm kernel oil. The company also said that the implementation of the EU Deforestation Regulation had contributed to shortages of materials suitable for the European market.

Rather than abandoning its commitment, BASF says it intends to continue increasing certified volumes where commercially and technically possible. RSPO certification remains its preferred standard.

The company has also moved its target for sourcing fully certified key palm derivatives to 2030. Those derivatives present an additional challenge because the palm material may have passed through several processing and transformation stages before entering a chemical formulation.

This is a useful reality check for the wider industry.

Responsible sourcing targets are necessary, but they cannot be achieved through procurement policy alone. Availability, regional regulation, supplier capability, segregation systems, certification capacity and commercial viability all influence whether a company can obtain compliant material at the required scale.

Responsible Sourcing Extends Beyond Palm Oil

BASF’s Care Chemicals business also uses other renewable oils, including coconut and castor oil, in products for cosmetics, personal care, detergents and industrial applications.

Each raw material presents a different set of sourcing risks.

Coconut supply chains often involve smallholders and geographically fragmented production. Responsible sourcing therefore requires more than checking the country of origin. It may involve certification, agricultural practices, working conditions and the economic resilience of farming communities.

Castor oil presents another distinct challenge.

India is central to the global supply of castor beans, while the oil itself is used across plastics, coatings, cosmetics, pharmaceuticals and several specialty chemical applications. BASF has participated in the Sustainable Castor Initiative, known as Project Pragati, since 2016.

The initiative brings BASF together with Arkema, Jayant Agro-Organics and the civil-society organisation Solidaridad. Its purpose is to address social, environmental, health and safety risks associated with castor cultivation.

Participating farmers receive training in areas such as:

• Safer use of crop-protection products

• Improved agricultural and cultivation practices

• Soil protection and crop management

• Occupational health and field safety

• Personal protective equipment

• Social and labour-related expectations

The work contributed to the development of SuCCESS, or Sustainable Castor Caring for Environment and Social Standards. The independently auditable framework covers 11 principles related to responsible castor production.

This approach goes beyond demanding compliance from farmers. It attempts to increase the capability of the supply chain to meet the required standard.

That difference matters.

A procurement system based only on supplier exclusion may remove an immediate risk from one company’s portfolio, but it does not necessarily improve conditions at the source. Supplier development, technical support and farmer training can create a more durable improvement.

Supplier Expectations Must Be Built Into Procurement

BASF’s wider procurement model requires suppliers to comply with applicable laws and internationally recognised environmental, social and governance standards.

Its Supplier Code of Conduct covers areas including:

• Environmental protection

• Human and labour rights

• Child and forced labour

• Occupational and social standards

• Anti-discrimination

• Anti-corruption

• Expectations for subcontractors and upstream suppliers

BASF states that suppliers are evaluated on more than price and commercial performance. The company also examines environmental, social and governance factors and expects suppliers to promote similar principles within their own supply chains.

The company uses a risk-based approach rather than treating every supplier identically. Country risk, industry risk, material criticality and BASF’s ability to influence the supplier can all affect the level of scrutiny applied.

Evaluations are conducted through mechanisms including EcoVadis assessments, Together for Sustainability audits and selected Responsible Care audits. When weaknesses are identified, corrective-action plans and follow-up reviews are used to track improvement.

In 2025, BASF reported that 100 sustainability audits were conducted at raw-material supplier sites on its behalf. It also received EcoVadis assessments for 257 suppliers considered to have potential sustainability risks.

The model combines four elements:

Define the expectation. Suppliers need clear environmental, ethical and social requirements.

Identify the risk. Procurement teams must know which materials, countries and suppliers require greater attention.

Verify performance. Questionnaires alone may be insufficient where the exposure is significant.

Correct or escalate. Findings must lead to improvement plans, commercial consequences or, in serious cases, termination of the relationship.

Collaboration Can Reduce Repeated Supplier Assessments

BASF is also a founding member of Together for Sustainability, an initiative created by chemical companies to improve and standardise sustainability assessments across the industry.

Under the model, suppliers can be assessed using a shared framework rather than repeatedly completing different questionnaires and audits for every customer. Participating procurement teams can use recognised assessment information through a common system.

This offers two potential benefits.

First, it reduces duplicated work for suppliers serving multiple chemical companies.

Second, it creates greater consistency in how environmental, social, labour and governance performance is evaluated.

For responsible sourcing to scale across the chemical industry, this type of shared infrastructure may be essential. Thousands of suppliers cannot practically respond to entirely different assessment methods, evidence requests and audit expectations from every customer.

Standardisation does not remove the need for company-specific due diligence, but it can make the underlying process more efficient.

Circular Feedstocks Are Becoming a Sourcing Decision

Responsible sourcing is also beginning to influence the type of carbon and feedstock entering chemical manufacturing.

In 2024, BASF and Encina Development Group announced a long-term agreement for the supply of circular benzene produced from post-consumer plastic waste. BASF intends to use the chemically recycled material within its Ccycled product portfolio.

This expands the procurement question beyond whether a conventional feedstock was sourced responsibly.

Companies must now also consider:

• Whether recycled or renewable feedstocks are available

• How their origin and chain of custody will be verified

• Whether the material meets process and purity requirements

• How circular content will be allocated and documented

• Whether supply is sufficient for commercial production

• How sustainability claims will be supported

Circular sourcing therefore requires close coordination among procurement, R&D, production, quality, sustainability and regulatory teams.

A material may appear attractive from a sustainability perspective but still require extensive technical qualification before it can enter a chemical process. Similarly, a technically suitable material may not support a defensible sustainability claim if its sourcing and allocation records are inadequate.

Procurement Partnerships Can Enable Lower-Carbon Production

BASF’s work with Siemens Energy provides another example of procurement supporting a wider production transition.

In March 2025, BASF commissioned a 54-megawatt proton-exchange-membrane water electrolyser at its Ludwigshafen site. The system was built in cooperation with Siemens Energy and has an annual production capacity of up to 8,000 metric tonnes of hydrogen.

The electrolyser is integrated directly into the site’s chemical-production infrastructure. Hydrogen produced using renewable electricity can be supplied through the existing hydrogen network and used as a raw material for chemical products with a reduced carbon footprint.

BASF estimates that the project has the potential to reduce greenhouse-gas emissions at the Ludwigshafen site by up to 72,000 metric tonnes annually.

Although this is a manufacturing project, it also illustrates the strategic role of sourcing and supplier collaboration.

The transition to lower-carbon chemicals depends not only on laboratory innovation. It requires companies to procure new technologies, secure alternative energy and feedstock inputs, establish qualified partnerships and integrate them into existing production systems without compromising continuity.

What Other Chemical Companies Can Learn

BASF’s approach does not suggest that responsible sourcing has become simple or that every target has been achieved.

Its 2025 palm-certification result shows the opposite. Even a large global organisation can face shortages, regulatory complications and limited availability of materials that meet the preferred sustainability standard.

The more important lesson lies in how responsible sourcing is being managed.

It is increasingly treated as a continuous operating system built around:

• Supply-chain traceability

• Material-specific sourcing policies

• Supplier codes and contractual expectations

• Risk-based assessments and audits

• Corrective-action management

• Smallholder and supplier development

• Cross-industry assessment frameworks

• Circular and renewable feedstock qualification

• Collaboration between procurement and technical teams

For chemical companies, the challenge is no longer deciding whether sustainability belongs in procurement.

The challenge is converting broad commitments into repeatable sourcing decisions that can survive technical review, supplier disruption, regulatory scrutiny and commercial pressure.

Turn Industry Developments Into Better Technical Decisions

Responsible sourcing, circular feedstocks and supply-chain transparency are changing how chemical companies select materials, qualify suppliers, manage compliance and plan future products.

Keeping up with these developments is useful. Knowing how to apply them within R&D, procurement, regulatory, quality and manufacturing decisions is what creates business value.

OnlyTRAININGS provides expert-led technical training for professionals across the chemical and allied industries. Its training portfolio covers sustainability and green chemistry, formulation, materials, regulatory compliance, processing, industrial problem-solving, artificial intelligence and emerging technologies.

Whether your team is responding to new sourcing requirements, evaluating alternative raw materials, strengthening regulatory capability or preparing for the next shift in chemical manufacturing, the platform is designed to help turn industry knowledge into practical action.

The Chemical Industry Is Changing. Is Your Team Keeping Up?

Explore Expert-Led Chemical Industry Trainings at OnlyTRAININGS

BASF responsible sourcing, sustainable procurement, responsible sourcing, chemical supply chain, supplier sustainability, sustainable sourcing, green procurement, supply chain sustainability, ethical sourcing, circular feedstocks, sustainable raw materials, supplier audits, supplier compliance, palm oil sourcing, castor oil sustainability, renewable raw materials, chemical procurement, ESG procurement, supply chain traceability, Together for Sustainability, circular economy chemicals, low-carbon chemicals, sustainable suppliers, procurement transformation, how BASF is advancing responsible sourcing, BASF sustainable procurement strategy, responsible sourcing in the chemical industry, how chemical companies manage responsible sourcing, BASF supplier sustainability requirements, sustainable sourcing of chemical raw materials, green procurement strategies for chemical companies, sustainability in chemical industry supply chains, ethical sourcing practices in the chemical industry, how BASF uses circular chemical feedstocks, sustainable raw material sourcing strategies, supplier sustainability audits in the chemical industry, BASF Supplier Code of Conduct requirements, BASF sustainable palm oil sourcing strategy, BASF sustainable castor oil sourcing initiative, responsible sourcing of renewable raw materials, sustainable procurement practices for chemical companies, ESG requirements for chemical industry suppliers, raw material traceability in chemical supply chains, BASF Together for Sustainability supplier programme, circular economy sourcing in chemical manufacturing, procurement strategies for lower-carbon chemical production, how to evaluate sustainable chemical suppliers, how sustainability is transforming chemical procurement

If you manufacture or supply food contact materials for more than one market, you already know the frustrating truth: passing FDA review doesn't mean you're compliant in the EU, and EU compliance doesn't carry over to China. Each system evaluates safety differently, uses different documentation, and fails materials for different reasons.

This isn't a regulatory footnote. It's the single biggest reason multi-region product launches get delayed — not because the material is unsafe, but because a team optimized for one region's logic runs into another region's blind spot.

Here's how the three major systems actually compare, and where the real compliance risk sits in each one.

The Three Systems at a Glance

United States: A Pathway Problem, Not a Listing Problem

Most teams assume FDA compliance is about checking whether a substance appears on a list. In practice, the harder decision is choosing the right pathway — Food Contact Notification (FCN), GRAS, Prior Sanction, or Threshold of Regulation (TOR) exemption — and defending that choice if it's challenged.

Each pathway carries different assumptions about migration, different timelines, and different levels of regulatory scrutiny. Choosing the wrong one, or leaning on an exemption that doesn't actually apply to the use case, is one of the most common causes of late-stage compliance failures — often after testing is already complete.

See the full FDA compliance framework, including FCN vs. GRAS decision logic →

European Union: Self-Declaration Puts the Burden on You

Unlike the US notification model, the EU runs on self-declaration. There's no "EU approval" — manufacturers must produce a Declaration of Compliance (DoC) demonstrating that a material meets Regulation (EU) 10/2011 (for plastics) and the broader Framework Regulation 1935/2004.

The EU's positive list (the "Union list" in Annex I) is under continuous revision — Commission Regulation (EU) 2026/245, which entered into force in February 2026, added and revised authorizations for several substances used in polyolefins, polyamides, PET, PLA, and PVC materials. Bisphenol A rules have also tightened substantially under Regulation (EU) 2024/3190, with transitional provisions running through September 2026 for products already on the market.

The technical risk that trips up most non-EU manufacturers isn't the positive list itself — it's Non-Intentionally Added Substances (NIAS): breakdown products, impurities, and reaction by-products that aren't deliberately added but still have to be risk-assessed. NIAS evaluation is where "compliant on paper" and "compliant in practice" most often diverge.

China: Fast-Moving Standards, Material-Specific Rules

China's GB 4806 series governs food contact materials by material type — separate standards exist for plastics, coatings, rubber, silicone rubber, paper, adhesives, and more, each with its own positive list under GB 9685.

This system has moved quickly in the past two years. Revised standards for coatings (GB 4806.10-2025) and silicone rubber (GB 4806.16-2025) take effect September 2026, expanding the approved substance list for coatings from 105 to 346 entries and cutting the BPA specific migration limit tenfold, from 0.6 mg/kg to 0.05 mg/kg. A draft standard for food-contact regenerated cellulose materials was also opened for consultation in 2026.

For international manufacturers, the practical risk in China isn't unfamiliarity with the concept of a positive list — it's the pace of standard revisions and the requirement to track which GB standard version applies to a specific material category, since older versions are explicitly invalidated once revisions take effect.

Why This Matters for Multi-Region Launches

The teams that get burned aren't the ones who don't know the regulations exist — they're the ones who apply one region's compliance logic to another region's system. FDA's notification model, the EU's self-declaration burden, and China's material-specific GB standards each require a different verification approach, different documentation, and different technical justification.

Getting this right up front — before formulation is locked and before submission timelines are set — is what separates a smooth multi-region launch from a late-stage scramble.

Go Deeper

This overview is intentionally high-level. For a full, decision-focused breakdown of FDA compliance specifically — including FCN vs. GRAS vs. Prior Sanction decision criteria, migration and exposure assessment logic, and the functional barrier assumptions that most often fail under review — see the complete training:

FDA Compliance for Food Contact Materials: Migration Risk, FCN/GRAS & Informed Compliance Decisions →

Sources: European Commission Food Safety Directorate; UL Solutions; SGS Safeguards; knoell regulatory updates; SESEC; ChemLinked; Kelley Drye & Warren (KHLaw) regulatory analysis.

Top 10 Global Chemical and Materials Companies Leading Publicly Verified AI Adoption

Published by: OnlyTRAININGS
Survey version: 2026 evidence-based edition
Evidence review date: 8 July 2026 (P.S. report includes only publicly verifiable claims)
Scope: Global chemical, specialty materials, polymers, petrochemicals, coatings, catalysts, water solutions, and advanced materials companies

1. Foreword and Purpose

Artificial intelligence is becoming one of the most important competitive forces in the chemical and materials industry. But it is also one of the most easily exaggerated topics. Many companies now mention AI in strategy documents, investor presentations, or innovation stories. That does not automatically mean they are leading the AI race.

For this report, OnlyTRAININGS has taken a stricter approach. We have focused only on companies where there is public, cross-checkable evidence that AI is being applied to real chemical or materials industry workflows: R&D, formulation development, materials discovery, plant operations, process optimization, catalyst performance, customer-facing technical support, internal knowledge systems, or AI-ready laboratory infrastructure.

The purpose of this report is not to produce a hype-driven ranking. The purpose is to provide chemical industry professionals, R&D leaders, technical managers, plant teams, formulation scientists, regulatory teams, and business leaders with a practical view of what credible AI adoption looks like in the chemical sector today.

This report also supports a broader workforce message. AI will not replace chemical expertise. But companies that organize their technical knowledge, structure their data, train their people, and apply AI to real industrial problems will move faster than companies still depending on disconnected spreadsheets, scattered technical reports, manual trial-and-error, and individual memory.

2. Methodology: How the Companies Were Evaluated

This survey ranks companies by publicly verified AI adoption in the chemical and materials industry, not by revenue size alone.

The ranking gives higher weight to evidence that is:

1.    Published by the company itself through official reports, press releases, product pages, or investor material.

2.    Supported by reputable third-party sources such as technology partners, recognized industry publications, or established media.

3.    Directly connected to chemical, materials, manufacturing, R&D, formulation, catalyst, polymer, or industrial operations.

4.    Specific enough to show what the AI system actually does.

5.    Supported by measurable deployment evidence where available.

This report avoids unsupported claims such as estimated patent counts, assumed financial savings, unverified yield improvements, or “AI leadership” statements that cannot be traced to a credible source.

Evidence Grades Used in This Report

3. Ranking Overview

4. Detailed Company Profiles

4.1 BASF

Why BASF Ranks First

BASF ranks first because its public AI evidence is broad, technically relevant, and directly connected to chemical-industry work. BASF states that it began exploring machine-learning models more than ten years ago and has developed and applied AI across production, engineering, R&D, management, and other areas. The company also frames AI as a way to improve safety, quality, productivity, cost efficiency, growth, sustainability, and circular-economy development.

One of BASF’s strongest public examples is QKnows AI, a research knowledge platform built for BASF’s global R&D community. BASF says QKnows allows researchers to search literature, patents, and BASF research reports in one application, using natural-language questions across a database of more than 400 million documents. This is highly relevant because chemical R&D depends heavily on buried internal knowledge, prior experiments, patents, and historical technical reports.

BASF also gives public examples of customer-facing AI-enabled formulation tools. Its Rediso platform uses machine learning and AI to support digital customer requests in home care and industrial cleaning, including a simulator that helps customers optimize hand dishwash formulations in less time. BASF also describes ZoomLab as an AI-supported pharma formulation platform that uses machine learning and other technologies to help formulators identify ingredients and predict and optimize formulations.

The company is also applying AI to plant knowledge. BASF’s PlantGPT is described as an AI-powered digital assistant focused on operations and site services, trained with thousands of documents, safety procedures, and materials to provide plant-specific knowledge in real time.

Another important example is BASF’s AI-supported groundwater leaching risk prediction in agricultural research. BASF states that it ran about one million simulations on its Quriosity supercomputer to develop the underlying model for predicting groundwater leaching risk early in the research process.

Why This Matters

BASF’s public evidence shows a serious AI pattern: AI is being used not only for generic productivity but also for research search, formulation support, plant knowledge, agricultural prediction, and customer-facing technical tools. That is why BASF deserves to be treated as one of the most credible AI leaders in the chemical industry.

What Other Companies Can Learn from BASF

The biggest lesson from BASF is that chemical AI becomes more powerful when it is connected to proprietary knowledge. Most chemical companies already have valuable data, but it is often trapped inside old reports, lab notebooks, PDFs, formulation files, customer documents, and individual memory. BASF’s example shows that searchable, AI-supported technical knowledge can become a competitive asset.

4.2 Covestro

Why Covestro Ranks Second

Covestro ranks second because it provides unusually clear public evidence that AI has moved from pilot projects into measurable operational deployment. In its 2025 annual financial report, Covestro states that AI and digital transformation have matured from pilot projects into systematic growth drivers with measurable business success.

The strongest evidence comes from production. Covestro reports automated monitoring of critical plant components and process steps supported by machine learning. The company says this system is being rolled out globally across nine sites and created increased production and cost-avoidance opportunities worth double-digit millions of euros in 2025 alone.

Covestro also reports that an autonomous production line in Dormagen has been running since June 2024 and that learnings from this line are being transferred to other production facilities, with pilot projects initiated at seven sites worldwide. Its 2025 innovation report further states that an entire production line in Dormagen has been working fully autonomously from production planning to finished-product provision since June 2024.

Covestro’s AI adoption is not limited to production. The company says AI is being used in procurement, production, sales, finance, and R&D, with direct EBITDA impact across several areas. It also reports that its Covestro Virtual Assistant, or CoVA, is becoming a central virtual team partner available to employees with company-terminal access.

In R&D, Covestro says it has implemented AI technologies to speed up simulation processes by combining machine learning with established computational chemistry. It also reports an AI Academy where around 100 employees are participating in a 12-month curriculum while working on concrete AI applications in their own areas.

Why This Matters

Covestro is one of the clearest examples of AI being used at enterprise scale inside a chemical company. Its public evidence covers plant monitoring, autonomous production, R&D simulation, virtual assistants, AI training, and measurable business impact.

What Other Companies Can Learn from Covestro

The most important learning is that AI must be tied to operational KPIs. Chemical companies should not measure AI success only by the number of tools launched. They should measure it by reduced downtime, improved production reliability, faster simulation, fewer manual steps, better customer response, and documented business impact.

4.3 SABIC

Why SABIC Ranks Third

SABIC ranks third because it provides strong public evidence of large-scale AI deployment in manufacturing. In its Q2 2025 earnings call presentation, SABIC stated that it had deployed more than 490 AI-powered digital models and that 42% of manufacturing facilities were actively using AI-powered tools.

The same SABIC presentation states that the company launched AI Guidelines to ensure secure, ethical, and effective use of AI technologies. This is an important maturity signal because large-scale AI in petrochemicals cannot be treated only as a technical experiment. It requires governance around data, safety, cybersecurity, decision accountability, and responsible deployment.

SABIC’s AI deployment is also linked to operational excellence and sustainability. The same presentation connects digital transformation to manufacturing excellence, supply chain and business optimization, feedstock quality, stream synergies, and end-to-end digital integration.

Why This Matters

SABIC’s strongest AI evidence is scale. Deploying hundreds of AI-powered digital models across manufacturing systems suggests that AI has moved beyond scattered pilots. For a petrochemical company, this scale matters because small improvements in energy use, planning, yield, reliability, feedstock flexibility, or emissions can have large operational consequences.

What Other Companies Can Learn from SABIC

The main lesson from SABIC is that AI scale requires both digital infrastructure and governance. Chemical companies cannot safely scale AI across production without clear rules, integrated systems, and responsible-use frameworks. AI guidelines are not just corporate paperwork; they are part of making AI usable in industrial settings.

4.4 Dow

Why Dow Ranks Fourth

Dow ranks fourth because its AI evidence is strongly connected to formulation development and production optimization. Dow’s Predictive Intelligence capability, created in collaboration with Microsoft, was developed for Dow Polyurethanes and recognized with a FutureEdge 50 award. Dow states that this capability combines materials science expertise with Microsoft’s AI and machine-learning experience to transform polyurethane product development.

The most important confirmed claim is the formulation-development acceleration. Dow states that Predictive Intelligence reduced the discovery phase for polyurethane formulations from a 2–3 month process to 30 seconds, described as a 200,000x acceleration.

This matters because formulation development is one of the chemical industry’s most time-consuming activities. In polyurethane systems, coatings, adhesives, sealants, elastomers, foams, construction materials, and many other formulated products, technical teams often work through repeated lab trials. AI that narrows the candidate space can reduce wasted experimentation and improve customer response speed.

Dow also has confirmed AI evidence in production optimization. Its Advanced Modeling of Oxygenated Solvents tool is described by Dow as a machine-learning platform deployed at 13 plants for more than 50 products, delivering value through improved efficiency and reduced waste.

Why This Matters

Dow’s strongest AI examples are commercially practical. They show AI helping with formulation speed, product development, plant efficiency, and waste reduction. That makes Dow especially relevant for specialty formulation businesses.

What Other Companies Can Learn from Dow

Dow’s example shows that AI should reduce the number of weak experiments. A realistic AI goal for formulation-led businesses is not “AI writes the perfect formula.” It is “AI helps technical teams decide which formulas are worth testing first.”

4.5 Syensqo

Why Syensqo Ranks Fifth

Syensqo ranks fifth because it has positioned AI directly around materials discovery, sustainable materials, generative AI, and chemical-industry partnerships. In June 2025, Syensqo and Microsoft signed a memorandum of understanding to explore strategic collaboration in AI, cloud computing, research, manufacturing, and product development.

The Syensqo-Microsoft collaboration specifically includes scaling secure cloud systems, applying generative AI across operations, and discovering sustainable materials through Microsoft Discovery, with Syensqo described as the main developing partner in the chemical industry for that platform. A Syensqo.ai version of the announcement states that Syensqo became the main chemical-industry partner for Microsoft Discovery and that the collaboration aims to apply generative AI at scale across research, manufacturing, and workplace systems.

Syensqo also states that the partnership will support next-generation sustainable materials, including bio-based polymers, circular composites, and clean-energy solutions. The company further describes internal AI tools such as SyGPT, a secure generative AI assistant for employees, and SyGrow, an AI-driven lead-generation tool for commercial operations.

In October 2025, Syensqo and Mohammed VI Polytechnic University launched a joint AI Lab. The lab is intended to build next-generation agentic AI technologies for chemistry and materials science, linking academic AI expertise with Syensqo’s materials and chemical-science experience.

Why This Matters

Syensqo is not only applying AI inside existing workflows. It is building an AI-centered innovation ecosystem around materials science, generative AI, partnerships, and sustainable industrial applications.

What Other Companies Can Learn from Syensqo

The key lesson is that AI partnerships matter. Most chemical companies do not have enough internal AI infrastructure to build frontier materials-discovery systems alone. The winners will combine domain expertise, proprietary data, secure cloud systems, and specialist AI partners.

4.6 LG Chem

Why LG Chem Ranks Sixth

LG Chem ranks sixth because public reporting shows broad AI adoption across manufacturing, quality prediction, smart factories, and employee analytics. The Korea Times reported in May 2025 that LG Chem was deploying AI across quality prediction, process optimization, contract review, and exchange-rate forecasting.

LG Chem’s battery-materials business is building a cathode-materials plant in Tennessee that is expected to operate as a smart factory equipped with AI and digital transformation platforms. The plant will adopt digital technologies used at LG Chem’s Cheongju factory to improve cathode-material quality.

LG Chem has also introduced a deep-learning image-analysis system at its Yeosu petrochemical plant to detect abnormalities in the flare-stack process and improve operational efficiency. The company also uses AI to predict properties of superabsorbent polymers and digital-twin technology to detect issues in regenerative thermal oxidizers and other equipment before problems occur.

A particularly important adoption signal is LG Chem’s citizen data scientist platform. The company launched an internal AI analytics tool accessible to employees regardless of coding or data-analysis skills. According to the report, about 40 employees used the platform over three months and identified more than 20 improvement opportunities, including RO membrane output and battery separator quality.

Why This Matters

LG Chem’s AI story is important because it is not confined to one elite R&D group. It includes quality prediction, process optimization, smart factories, digital twins, contract review, translation, and employee-accessible analytics.

What Other Companies Can Learn from LG Chem

The main lesson is that AI adoption must reach ordinary technical teams. A chemical company cannot rely only on a central data-science group. R&D, QA, production, regulatory, technical service, and commercial teams need safe AI tools that fit their daily work.

4.7 Evonik

Why Evonik Ranks Seventh

Evonik ranks seventh because its AI evidence is strongly connected to specialty-chemical formulation, coatings, additives, and laboratory support. Evonik’s COATINO platform is one of the clearest public examples of AI-supported formulation assistance in specialty chemicals.

Evonik states that COATINO is a digital formulation network that helps users discover, understand, and use coating additives. The platform includes specialized calculators, AI-driven tools for assessing coating defects, and an interactive learning environment for coatings professionals.

Earlier Evonik material also states that COATINO recommendation algorithms rely on AI-based technology and machine-learning principles, and that Evonik operates high-throughput equipment at its Essen site that tests and evaluates up to 120 coating formulations per day.

Evonik has also discussed AIChemBuddy, a laboratory AI support system designed to support researchers with relevant data, measurements, and experimental results.

Why This Matters

Evonik’s strongest AI relevance is domain depth. Coatings, additives, and specialty formulations are highly complex because performance depends on resin chemistry, pigment systems, wetting, dispersion, rheology, substrate interaction, drying behavior, film formation, regulatory constraints, and end-use conditions. A domain-specific AI platform can be more valuable than a generic chatbot.

What Other Companies Can Learn from Evonik

The lesson is to build narrow technical AI systems before building broad ones. A coating additive assistant, adhesive failure assistant, polymer compatibility assistant, or defoamer-selection assistant is often more valuable than a general company chatbot.

4.8 Arkema

Why Arkema Ranks Eighth

Arkema ranks eighth because it publicly connects AI with specialty-material innovation, formulation optimization, experimental-data analysis, and digital operations. Arkema’s digital strategy page states that digital technologies such as molecular modelling, digital twins, and AI help accelerate the design of innovative and sustainable specialty materials.

Arkema specifically states that it invests in digital tools to simulate the behavior and properties of molecules, polymers, and materials; optimize formulations using AI algorithms that predict material performance; automate experimental-data analysis; and collaborate with academic and technological partners.

Arkema also links these capabilities to markets such as batteries, adhesives, bio-based polymers, and electronic materials. In operations, Arkema says it uses data analytics, automation, AI, and real-time visibility tools to improve reliability, safety, sustainability, and supply-chain agility.

Why This Matters

Arkema’s AI positioning is credible because it is tied to specialty materials and formulation work, not just generic office productivity. The company’s evidence is broader than it is numerical, so the ranking keeps Arkema below companies with more detailed deployment data.

What Other Companies Can Learn from Arkema

Arkema shows that AI should be embedded across the innovation chain: molecular modelling, formulation optimization, experimental analysis, industrial performance, predictive maintenance, and supply-chain visibility.

4.9 Mitsubishi Chemical Group

Why Mitsubishi Chemical Group Ranks Ninth

Mitsubishi Chemical Group ranks ninth because it has long-standing public evidence around materials informatics and AI-integrated materials design. In 2018, Mitsubishi Chemical Holdings established a Materials Informatics Center of Excellence to promote the proactive use of materials informatics across the group.

The company stated that the center would use internal technologies and external partners, including academia, to speed up materials development and support early-stage new material design. It also stated that materials informatics was being introduced across the chemical industry as a new approach using data science and AI.

Mitsubishi Chemical’s technology-platform page states that the company is developing advanced materials design technologies integrating computation, measurement, and AI, using technologies such as high-performance computing, quantum computing, and experiment automation.

In 2026, Mitsubishi Chemical and AIST also partnered with Connected DMV on a Global Industry Challenge exploring AI-enhanced quantum eigensolvers for a Quantum Materials Informatics platform, with a focus on more efficient exploration of molecular and materials design spaces.

Why This Matters

Mitsubishi Chemical Group’s AI relevance is strongest in materials informatics and advanced materials design. It may not disclose the same level of operational AI detail as Covestro or SABIC, but its long-term materials-informatics foundation is significant.

What Other Companies Can Learn from Mitsubishi Chemical Group

The main lesson is that AI in materials science must connect computation, measurement, experiment automation, and domain knowledge. A model alone is not enough. The company needs a data-and-experiment system around the model.

4.10 DuPont

Why DuPont Ranks Tenth

DuPont ranks tenth because its strongest verified AI evidence is newer but highly relevant to chemical and materials R&D. In April 2026, DuPont announced a collaboration with Uncountable to advance its AI-ready labs strategy across its R&D organization.

DuPont states that the collaboration will help scale digital lab workflows, expand access to high-quality experimental data, and turn insights into faster, more efficient innovation across its R&D organization. DuPont’s CTO stated that high-quality, structured data is critical to achieving innovation excellence at scale and enabling advanced analytics, machine learning, and AI to accelerate delivery of solutions to customers.

The collaboration also focuses on complex formulations. DuPont states that the partnership enhances how the company designs, tests, and optimizes complex formulations by standardizing data and optimizing R&D workflows through Uncountable’s platform.

DuPont also launched an AI-enabled RO Operations Advisor in April 2026. The tool analyzes historical reverse-osmosis system operating data and applies advanced analytics and AI models based on DuPont’s water-science expertise to give system-specific guidance on cleaning and membrane replacement. DuPont states that the tool can estimate potential operating-expense savings and can help enable up to 20% reductions in operating expenses in customer applications.

Why This Matters

DuPont’s strongest lesson is data readiness. The company’s AI-ready labs initiative recognizes that AI in R&D depends on structured experimental data, standardized workflows, and consistent access to high-quality datasets.

What Other Companies Can Learn from DuPont

The learning is practical: before building advanced AI, structure the lab data. Chemical companies should standardize how experiments are captured, how results are stored, how formulations are versioned, and how failed trials are preserved for future learning.

5. Companies Not Included in the Final Top 10

Some companies are highly relevant to chemicals, digitalization, circularity, or manufacturing, but were not included in the top 10 because the AI-specific public evidence was either less direct or not strong enough for this report’s evidence standard.

LyondellBasell

LyondellBasell is an important circular-plastics and advanced-recycling player, especially through MoReTec, but the AI-specific claims often made around sorting purity, AI reactor modeling, and algorithm licensing require stronger direct public evidence before inclusion in this ranking.

Air Liquide

Air Liquide is a strong digital-operations and industrial-gas company, but the AI-specific savings and plant-monitoring claims often circulated publicly require stronger direct source support before being used in an AI leadership ranking.

Wanhua Chemical

Wanhua is a major global chemical producer, but the strongest AI claims around autonomous MDI process control, yield improvement, AI video safety analytics, and rapid AI model replication need direct annual-report, sustainability-report, or technology-partner evidence before publication.

Clariant

Clariant was close to inclusion. Its CLARITY Prime platform is clearly AI-relevant because Clariant describes it as an AI-powered digital service for catalyst operation, with automated catalyst health alerts and machine-learning-based performance-projection tools. Clariant also worked with AWS Partner Chaos Gears to build a generative AI foundation, and AWS reports that Clariant rolled out an internal chatbot to about 1,000 employees after developing the platform in less than three months. Clariant also partnered with Uncountable to replace a legacy ELN and build a structured scientific backbone across more than 1,000 users and 35+ global sites.

Clariant may deserve inclusion in a future “AI-ready specialty chemicals” ranking, but for this broader global chemical-company survey, DuPont was retained because of its confirmed AI-ready labs and AI-enabled water-operations evidence.

6. Cross-Cutting Findings: Why These Companies Are Ahead

6.1 They Do Not Treat AI as a General Chatbot

The strongest companies are not only deploying AI for writing emails or summarizing documents. BASF is applying AI to research knowledge, plant-specific support, formulation platforms, and agricultural prediction. Covestro is applying AI to production monitoring, autonomous production, R&D simulation, and internal assistants. Dow is applying AI to polyurethane formulation and plant optimization. Syensqo is applying AI to materials discovery. Evonik is applying AI to coatings formulation.

The pattern is clear: chemical AI becomes valuable when it touches real chemical work.

6.2 They Build Around Proprietary Data

The most important AI asset in chemicals is not the model alone. It is the company’s proprietary technical data: formulations, process conditions, lab results, failed experiments, patents, customer complaints, raw-material properties, equipment data, and technical reports.

BASF’s QKnows example, DuPont’s AI-ready labs initiative, Clariant’s ELN replacement, and Mitsubishi Chemical’s Materials Informatics Center of Excellence all point in the same direction: companies are trying to make technical knowledge structured, searchable, and usable by AI.

6.3 They Use AI to Reduce Technical Waste

The clearest value proposition is not “AI replaces scientists.” It is “AI helps experts test better.” Dow’s polyurethane example shows AI reducing formulation discovery time. Covestro’s R&D example shows machine learning being combined with computational chemistry to speed simulation. Arkema says it uses AI algorithms to predict material performance and reduce formulation-development cycles.

6.4 They Connect AI to Plant Reality

Manufacturing evidence matters. Covestro’s machine-learning-supported plant monitoring, SABIC’s 490+ AI-powered manufacturing models, LG Chem’s deep-learning image analysis at Yeosu, and Arkema’s use of AI and real-time visibility tools in operations all show that AI is moving into production environments.

6.5 They Build AI Governance and Skills

SABIC’s AI Guidelines, Covestro’s AI Academy, LG Chem’s citizen data scientist platform, and Syensqo’s secure internal generative AI tools show that AI adoption is organizational, not only technical.

7. Lessons for Other Chemical Companies

Lesson 1: Start With Pain, Not Technology

The most useful AI projects usually start with repeated technical pain:

·      Too many failed formulation trials

·      Slow raw-material substitution

·      Repeated batch deviations

·      Long customer complaint investigations

·      Poor historical knowledge retrieval

·      Slow patent and literature review

·      Unstructured supplier documents

·      Manual regulatory comparison

·      Plant downtime and process instability

·      Energy and yield variability

·      Repeated quality issues

The companies in this survey are not using AI because it is fashionable. They are using it where prediction, search, automation, or decision support can reduce technical waste.

Lesson 2: Build a Technical Data Foundation First

AI-ready companies organize their knowledge before expecting AI to deliver value. That means structuring:

·      Formulation history

·      Raw-material specifications

·      Analytical data

·      Stability results

·      Batch records

·      Process parameters

·      Failure investigations

·      Customer complaints

·      Supplier change notices

·      Regulatory records

·      Patent references

·      Lab and pilot-plant reports

Without this foundation, even advanced AI tools will give weak or generic outputs.

Lesson 3: Use AI to Narrow the Search Space

In formulation and materials science, AI is most valuable when it reduces the number of weak options. It can rank candidates, identify similar past cases, flag likely risks, suggest missing tests, or recommend the next experiment.

But final decisions still require technical judgment. Chemical systems are affected by storage, humidity, substrate variation, supplier differences, contamination, processing conditions, packaging interaction, regulatory constraints, and customer-use conditions.

Lesson 4: Build Focused AI Assistants

A narrow chemical assistant is usually more valuable than a broad generic chatbot. Practical examples include:

·      Adhesive failure troubleshooting assistant

·      Coating defect and additive selection assistant

·      Polymer grade substitution assistant

·      Cosmetic stability risk assistant

·      Food-contact compliance screening assistant

·      SDS classification support assistant

·      Batch deviation triage assistant

·      Supplier document comparison assistant

·      Technical report drafting assistant

·      Patent landscape review assistant

Lesson 5: Train Technical Teams, Not Only IT Teams

AI adoption in chemical companies must include R&D scientists, formulators, process engineers, production teams, QA/QC teams, regulatory teams, application engineers, and technical service professionals.

The most important future role may not be a pure data scientist. It may be the chemical professional who understands formulation, processing, testing, customer application, regulatory constraints, and enough data science to work effectively with AI systems.

8. Scope for Mid-Sized and Smaller Chemical Companies

The AI race is not closed. Smaller and mid-sized chemical companies may not have the infrastructure of BASF, Dow, or SABIC, but they often have a practical advantage: shorter decision chains, closer links between R&D and production, and deep niche knowledge.

A realistic 12-month AI roadmap could look like this:

First 90 Days: Select Two Use Cases

Choose two high-value, narrow use cases. Good starting points include formulation troubleshooting, batch deviation review, supplier document comparison, customer complaint triage, or technical knowledge search.

Months 4–6: Structure the Data

Organize the documents and datasets needed for the selected use cases. This may include old reports, formulas, lab data, batch records, quality complaints, supplier documents, and regulatory notes.

Months 7–9: Build Controlled AI Workflows

Create a controlled AI workflow with human review. Do not let AI make final technical decisions. Use it to retrieve, compare, summarize, flag, and suggest.

Months 10–12: Measure Real Impact

Measure whether the AI workflow reduced trial cycles, improved complaint closure time, reduced repeated deviations, improved documentation, shortened regulatory review, improved first-pass formulation success, or reduced plant issue investigation time.

9. Workforce Imperative: Skills for the AI-Ready Chemical Professional

The AI-ready chemical professional will not only know chemistry. They will know how to work with data, ask better technical questions, validate AI outputs, and protect confidential information.

The most important skills include:

·      Technical data structuring

·      Formulation data interpretation

·      AI-assisted troubleshooting

·      Prompt design for scientific workflows

·      Literature and patent search validation

·      Statistical thinking

·      Process-data interpretation

·      Digital twin awareness

·      AI governance and confidentiality

·      Human review of AI-generated outputs

This is where industry training becomes important. Chemical companies will need professionals who understand both the technical reality of chemical work and the practical use of AI in R&D, production, quality, regulatory, and technical service environments.

10. Final Conclusion

The AI race in the chemical industry is not being won by companies with the loudest AI slogans. It is being won by companies that connect AI to real technical work.

BASF is using AI in R&D knowledge systems, formulation platforms, plant support, and agricultural prediction. Covestro is scaling AI into production, R&D simulation, virtual assistants, and employee training. SABIC has deployed hundreds of AI-powered manufacturing models and created AI governance. Dow has strong AI evidence in polyurethane formulation and production optimization. Syensqo is building AI-powered materials discovery partnerships. LG Chem is applying AI across quality prediction, smart factories, employee analytics, and process monitoring. Evonik is using AI for coatings formulation and specialty-chemical support. Arkema is applying AI to material-performance prediction, formulation optimization, and digital operations. Mitsubishi Chemical Group has a long-term materials informatics foundation. DuPont is building AI-ready labs and AI-enabled water operations.

The clear message for the chemical industry is this: AI will not replace chemical expertise. But it will reward companies that organize their expertise, structure their data, train their people, and apply AI to real industrial problems.

Companies that continue relying only on scattered spreadsheets, undocumented trial-and-error, disconnected technical reports, and individual memory will find it harder to compete with organizations that can search, learn, simulate, predict, and act faster.

About OnlyTRAININGS & How We Support The Transition

For over 10 years, OnlyTRAININGS has been a trusted provider of online technical and corporate training, serving professionals across the US and Europe. Our mission is to deliver practical, immediately applicable skills that help chemical companies and their employees thrive in the digital age.

What we offer:

·       Instructor‑led live and self‑paced trainings on AI in chemical processing, sustainability, and Industry 4.0.

·       Custom online trainings designed to reskill entire teams in data‑driven decision making.

·       Access to a global network of expert trainers with decades of plant and R&D experience.

To discuss how we can support your organization’s AI readiness journey, visit www.onlytrainings.com or contact us directly.

© 2026 Onlytrainings.com. All rights reserved. Proper attribution required when citing this survey.

chemical companies using AI, top 10 chemical companies leading AI adoption, AI in chemicals, AI in chemical manufacturing, AI in materials science, AI for formulation development, materials informatics, AI in specialty chemicals, chemical industry digital transformation, AI in R&D chemistry, AI for process optimization