Bringing New Materials to Market: AI, Scale-Up and Adoption

A new material can attract attention before anyone knows how to manufacture it reliably or introduce it into an existing process. The challenge for a commercial leader is to help establish a credible first application and organise the work needed to reach it.

That work includes understanding customer requirements, coordinating technical trials, developing the right partnerships and budgeting for a period in which effort may be substantial and sales limited.

AI can assist parts of material discovery and experimentation. Industrialisation and market adoption still need their own evidence and resources.

A case that connects performance and practical requirements

The Le Ferrand research group blog reports that the test tiles were installed at NTU on 12 May 2025. The published abstract of a subsequent study compares two fungal species and reports differences in performance and resistance to weathering. The wider bioSEA project compares mycelium, clay and recycled concrete, and reports durability constraints for the mycelium version outdoors.^[1][2][3]

These findings do not establish annual energy savings in an occupied building. They do show why the performance of a material needs to be considered alongside its intended environment and service life.

The next step is to identify an application whose requirements match the material’s demonstrated capabilities. That may mean a protected location, a different material or a different balance of requirements than originally expected.

Begin with a specific application

“Construction” or “the furniture industry” is too broad a starting point for an adoption plan.

Describe the application precisely enough to identify the existing solution, the problem to solve and the people who must accept the change. In a building project, these may include the specifier, installer, owner and maintenance team. In a factory, they may include purchasing, production, quality and engineering.

The first customer needs an appropriate reason to participate. That reason might be a performance problem, a design requirement, an opportunity to recover waste or a cost the existing process cannot adequately address.

Record the benefit as a hypothesis until it has been measured. “Could reduce rejects under these conditions” is a useful trial objective. It becomes a commercial claim only when the supporting evidence justifies it.

Develop the evidence in parallel

When assessing a first commercial application, I would organise the work around six questions.

These workstreams should exchange findings throughout development. Waiting for a technically finished product before involving customers can leave expensive requirements undiscovered. Allowing a customer preference to override an unresolved technical limit creates a different problem.

A useful review asks what has changed in the evidence and which decision that change enables.

Give AI a defined task and a validation method

MatterGen is one documented example of AI-assisted materials design. The 2025 study concerns inorganic crystal structures generated towards property targets, with experimental validation of a selected candidate.^[4]

It does not establish that AI can deliver a finished product across all material families. Mycelium growth, polymer formulation and inorganic crystal design involve different data, models and experimental methods.

Before allocating resources to an AI application, define the decision it should improve. Candidate selection, experimental prioritisation or analysis may each require a different approach.

Specify the inputs, the range over which the model has been tested, how uncertainty will be handled and which physical measurements will validate the output. A model useful within a known range of conditions may be unreliable when raw materials, equipment or scale change.

Autonomous laboratory systems also combine algorithmic planning with physical synthesis and measurement.^[5] That integration is useful to understand: a simulated result and a measured result are distinct forms of evidence.

The commercial contribution is to make application requirements available early. Researchers need to know which properties matter to the customer, under which conditions and within which production constraints.

Budget the transition

A development programme needs a view of cash use before repeat orders become predictable.

Include pilot production, testing, equipment changes, technical visits, customer trials, rejected material and channel training where relevant. Track working capital separately from expenditure: material in a long production cycle can tie up cash even before it is sold.

On the supplier’s side, assess expected contribution after the support required to serve the application. On the customer’s side, assess the full cost of adoption and the measured benefit. A product can create value for one party while leaving the other with an unattractive return.

Use ranges where information is uncertain. A single precise revenue forecast can conceal dependence on yield, approvals or the willingness of a first customer to proceed.

Choose partners for the work ahead

A distributor’s existing reach is useful, but technical adoption can require more than access to purchasing departments.

Assess whether the partner can recognise the application, gather relevant information, coordinate trials and support the product after delivery. Agree when the manufacturer joins the customer discussion and how feedback returns to development.

A local processor, installer or equipment partner may also influence whether adoption is practical. Their role should be defined before a launch depends on capabilities nobody has agreed to provide.

The appropriate route to market can therefore differ by application and geography. An established distributor may be suitable in one market; a small group of specialist pilot customers may be more useful in another.

Track learning as well as sales

Separate a sample request, a trial agreement, an accepted result, a first order and a repeat order. They reflect different commitments and should not carry identical weight in a pipeline.

At each review, ask:

• What has been demonstrated, under which conditions?
• Which uncertainty now prevents the next commitment?
• Who owns the next action and what will it cost?
• What would justify continuing, changing the application or stopping?

A successful trial without a route to customer adoption is incomplete. A slower project may still deserve support if it is resolving a significant uncertainty within an agreed budget.

Define the first application before promising the whole market

New materials can create opportunities for industrial suppliers and distribution networks. Realising those opportunities requires a practical connection between research, production and the people who will use the product.

My view is that AI increases the importance of this coordination when it expands the range of candidates worth investigating. It does not remove the need to decide where to invest or to demonstrate that a solution works in use.

For companies developing a new material, a useful starting point is a tightly defined application with a customer prepared to test it, a technical plan and a funded route to the next decision.

Explore my professional experience, or get in touch to discuss a relevant international market or technical sales challenge.