How AI will evolve (and hopefully enhance) US manufacturing

Summary: US manufacturing is growing, and so is the hazardous waste that comes with it. Manufacturing already generates about 90% of the nation's hazardous waste, and some companies are closing the facilities they used to run for their own waste. There's enough treatment capacity today, but new sites take years to permit. The good news is that AI in manufacturing is helping plants make less waste and turn what's left into a resource, including rare earth elements and lignin. The manufacturers that come do more than buy the tools – they put the right leaders in charge of them.

If you follow AI manufacturing news, most of the headlines are about robots, humanoids and lights-out factories.

The less exciting story is what comes out the back of the plant. And like most operational problems, whether it turns into a crisis or an advantage depends on who's running it.

The challenge: more production means more hazardous waste

Every new line, new site, and reshored product adds to the waste stream. A meaningful share of that waste is hazardous. It's regulated, it's expensive to move, and only so many places can legally take it.

Picture this: a plant adds a second shift to keep up with demand. Output goes up 30%. So do solvent use, spent catalysts, contaminated rags and wastewater sludge – none of that was in the business case.

Here's the part most companies miss: In a lot of plants, nobody at the leadership level owns that number. It sits with EHS as a compliance item, or it shows up buried in a vendor invoice. When it's nobody's job to reduce waste, it only grows.

Fewer places to send it

At the same time, the places that handle this waste are shifting.

While many large manufacturers used to run captive facilities (treatment sites that handle only their own waste), some are getting out of that business. 3M closed a Minnesota incinerator in 2021, and Alcoa sold an Arkansas site to Veolia in 2020. EPA's national capacity assessment confirms that several companies have shut down onsite and captive operations, pushing more volume to commercial facilities that are costly and difficult to permit.

To be fair about the risk, EPA says the country has adequate hazardous waste capacity through 2049. But that assumes active planning from states, industry and commercial operators. New facilities take years to approve, and demand can (and often does) shift much faster. Veolia estimates that even a 1% capacity shortfall could put at least $27.5 billion in economic output at risk through 2033.

Planning ahead is a leadership job. The companies getting ahead of this aren't waiting on regulators. They've made someone senior responsible for how much waste they generate and where it goes.

How manufacturers are addressing it

The good news is that the industry isn't standing still. The work is happening on two fronts: making less waste, and getting more value out of what's left. AI plays a growing role in both.

Making less waste with AI

The best waste is the waste you never make. According to Jackson Lewis, manufacturers are using AI to predict equipment failures, tune production schedules, improve quality, and reduce waste. Each has a direct line to fewer hazardous byproducts:

• Predictive maintenance catches a failing pump or seal before it leaks or forces an emergency cleanout.
• Computer vision spots defects earlier, so fewer bad parts get made and scrapped.
• Process optimization adjusts temperatures, flow rates and inputs in real time, so less material becomes waste.
• Digital twins let teams test changes virtually before touching the real line. At a Climate Week NYC panel, Unilever said it plans to build more than 40 of them over the next 18 months.

But buying the tools isn't the same as getting results. Grant Thornton's 2026 survey found that 48% of manufacturers are still piloting AI, and only 10% have fully integrated it into operations.

The tools work. What's usually missing is a leader who owns the result and can get the plant floor to use them.

Turning waste into a resource

This is where it gets optimistic. Researchers and companies are finding ways to treat waste as a feedstock.

Rare earth elements. The US needs more rare earths for electronics, EVs, defense and data centers, and there aren't enough in the ground here to meet demand – so companies are recovering them from scrap.

In September, Cyclic Materials opened a $20 million facility in Mesa, Arizona that uses automated separation to pull rare earth magnets from discarded electronics, robotics and medical devices. It can process up to 25,000 metric tons of scrap a year, and the company picked Mesa partly for its access to talent. AI helps here too: machine learning sorting models can raise sorting efficiency from about 15% to 80%, according to the Ellen MacArthur Foundation.

Lignin is the tough polymer that holds plants together. Pulp mills and biorefineries produce it by the ton, and most of it gets burned for heat. Researchers at the University of Tennessee and Oak Ridge National Laboratory recently published a one-step method to recover lignin from industrial black liquor and turn it into a material that cleans water, then gets reused as fertilizer. One waste stream, two products.

Neither is fully mature yet. But the direction is clear.

Why this is good news for US manufacturers

This is the approach that keeps manufacturing growing on US soil, instead of overloading the infrastructure that supports it:

• Less waste per unit means growth doesn't automatically strain disposal capacity.
• Recovered materials like rare earths reduce dependence on foreign supply.
• Lower waste costs make US production more competitive, which matters to the private equity firms backing US manufacturers.
• New recovery facilities create US jobs and local supply as more production comes back to the US.

Here's the catch. None of this happens because a company bought the right software. It happens because someone owns the result: the waste number, the recovery program, the rollout that has to survive the second shift. The manufacturers that win will be the ones that put that person in charge first.

The leaders who make it work

In most industrial companies, four seats decide whether waste reduction and AI turn into real savings:

When you're hiring for these seats, "AI experience" is the wrong filter. The tools will keep changing. What you want is someone who's made change stick inside an industrial operation. A few questions that help:

• "Tell me about a pilot you took to full production. What almost killed it?"
• "How did you get operators and supervisors to trust a new system?"
• "What did it save, and how did you measure it?"

Then define what success looks like before the offer goes out. And keep leadership style in mind: the person who launches a pilot isn't always the one who scales it. Our guide to the 10 leadership styles in manufacturing covers where each one fits.

Where Index Search fits

Any search firm can find you someone with "AI" on their résumé. That's not the hard part.

The hard part is finding the leader who's already taken a pilot to full production in a plant like yours. Who knows what a waste stream actually costs. Who can get a skeptical second shift to change how it works. Those people are rarely looking, and a generalist recruiter usually can't tell them apart from the people who just sat near the project.

That's the gap we close.

We come from the operation. Our CEO spent 25 years running manufacturing, from plant turnarounds to COO of a business with 25 sites worldwide. Our team has worked together for a decade. We ask the questions an operator would ask, because we've been the operator.

We only work in industrial markets. Manufacturing, distribution and industrial technology. Nothing else. We spend every day in this market, so we know who's actually done the work and who's only talked about it.

We use a proactive search process. We don't post and wait. We go directly to the operations, engineering, EHS and supply chain leaders who've already solved your problem somewhere else.

We assess and prove fit before you interview. Every leadership search includes our Leadership Style Assessment. You see how a finalist leads under pressure, measured against what your operation needs, not how well they interview.

The result is a leader who can own the problem from day one, instead of a year spent finding out they can't.

Have a waste or AI initiative that nobody's owning yet? Let's talk about who should.

Frequently asked questions

How is AI used in manufacturing?

AI in manufacturing is used for predictive maintenance, quality inspection, production scheduling, supply chain forecasting and process optimization. More plants are also using it to cut energy use, reduce waste and recover materials like rare earth elements from scrap.

How much hazardous waste does US manufacturing generate?

Manufacturing accounted for about 90% of US hazardous waste generation in 2023. Total generation is projected to rise from roughly 32.4 million tons to 37.4 million tons by 2033, with manufacturing activity driving most of the increase.

Can AI reduce hazardous waste in manufacturing?

Yes. AI reduces waste by preventing equipment failures that cause leaks, catching defects earlier to cut scrap, and tuning processes so fewer materials become waste. It also improves sorting, which makes recycling and material recovery practical at scale.

Can rare earth elements and lignin be recovered from industrial waste?

Yes. Companies are recovering rare earth magnets from discarded electronics and robotics at commercial scale in the US. Researchers have also developed methods to recover lignin from pulp mill waste and turn it into useful materials like water treatment media and fertilizer.

What leaders do manufacturers need to reduce waste and scale AI?

Usually four key team members: an operations leader who owns the financial result, a process or manufacturing engineering leader who standardizes it, an EHS leader who manages compliance and data, and a supply chain leader who turns recovered materials into value.