Why More Defense Spending Doesn't Automatically Build Manufacturing Capacity: Rush LaSelle on Contracting, CMMC, and the Shop Floor

Open on YouTube ↗
Overview

Defense budgets are rising, so why isn't manufacturing capacity rising with them? On this episode of Manufacturing Happy Hour, host Chris Luecke sat down with Rush LaSelle, CEO of Fathom Digital Manufacturing, at Fathom's facility in Hartland, Wisconsin. LaSelle's answer centers on visibility. Manufacturers can't commit capital, people, and factories on the strength of headlines. They need a demand signal they can see and trust. Along the way, LaSelle draws on a career spent in industrial robotics, early mobile robotics, healthcare automation, and high-volume contract manufacturing. The conversation also covers the recent pause of CMMC Phase 2, the talent shortage, and why LaSelle thinks physical AI is the next big opportunity.

21 min read

The conversation was recorded, as the host noted on air, in mid-to-late July 2026, so the policy details reflect that moment.

2:28

What Fathom Does

LaSelle describes Fathom as a contract manufacturer focused on precision manufacturing. The company has about 450 employees whose work is taking new product introduction (NPI) designs and turning them into high-quality product. He emphasizes in particular getting those products into rate production. Luecke frames the company as spanning the spectrum from classic subtractive machining to additive methods. LaSelle adds that the facility also does molding and CNC work.

LaSelle pushes back on the idea that additive manufacturing is only for prototyping. In his telling, additive is roughly a 30-year-old technology that for a long time couldn't deliver the quality that demanding applications required. Aerospace, for example, was working through material "allowables," and that was difficult. So the industry took a "crawl, walk, run" approach. The crawl stage meant making trinkets, things with no critical mechanical requirements. LaSelle says process control has improved a great deal since then. That matters now because product iteration is fast and customers want smaller, more customized batches. He sees value in a facility that can start a part in additive and then, at a certain volume, move to molding or CNC. For an engineer or designer, he says, that combination is "like being in Disneyland."

5:10

Defining Digital Manufacturing

Asked for a simple definition of digital manufacturing, LaSelle said there isn't one. "There's so many layers to that onion." He walked through how it evolved. When he studied mechanical engineering, design was done with pencil and paper, which didn't scale and couldn't easily move between locations. CAD started the digital shift by "democratizing" the movement of designs through a manufacturing ecosystem.

When Fathom came online about 15 years ago, LaSelle says, digital manufacturing for Fathom and peers like Xometry was mostly about quoting from digital designs. A customer could quickly learn what a part would cost to run on a CNC machine, or what a mold would cost and how it would perform at a given volume. Over roughly the last ten years, he argues, the definition has expanded. It now includes getting the part made in the right location, with the right certifications and the required quality. "Digital manufacturing went from design to supply chain pretty quick." He links this to safety-critical components. As someone who flies a lot, he wants aircraft parts made with full attention to mechanical properties and quality flowdowns, not just geometry. He also flagged security as the outermost layer of the onion, which he returned to later in the conversation.

7:36

Early Robotics: Pioneering Outside the Rust Belt

LaSelle began his career at Fanuc. Even there, he says, he felt like a pioneer, because he wasn't serving automotive. His team was putting robots into unusual applications such as packaging. They also had to overcome people's fear of large robots that might hurt them. He notes there is a real psychology behind working alongside a robot, and that each new technology brings its own fear. Back then, the fear was about safety.

Around 2011 to 2013, he worked in mobile robotics, which he calls "a classic technology looking for a home." AGVs already existed, moving goods around warehouses without someone driving a forklift. Mobile robots promised to do that safely without the line-following infrastructure AGVs required. They also promised to eventually get outside factories, for uses like autonomous home delivery. In his view, the technology was "wildly promising." What it lacked was everything around it that makes it useful, which he sees as a foreshadowing of challenges with today's technologies. He mentions that his group acquired a technology from New Hampshire that included optics and mapping software, the pieces needed to make a robot navigate.

9:51

The Kiva Lesson: The Value Was in the System

LaSelle then turned to Amazon's acquisition of Kiva Systems for $775 million. He recalls that people in the logistics space, companies doing cranes and forklifts, considered the valuation "unheard of" in industry. But he argues the "why" is what matters. The valuation, in his view, "had little to do with those orange robots running around a factory." It was about how the robots integrated into a system and the software that Mick and the Kiva team had built. That software is why a tube of toothpaste and a toothbrush ordered together come from the same facility. LaSelle says that before Kiva's heat maps told robots where to go, Amazon didn't have that kind of coordination across its warehouses. The value lay in the software, the implementation, and how the physical product interacted with Amazon's ordering systems.

Luecke suggested this connected robotics to the supply chain and the consumer. LaSelle agreed, but called Kiva one piece of a broader ecosystem. He applied the same point to the CNC machines on Fathom's floor. "The CNC itself doesn't do anything." A machine needs a great deal around it to become a useful tool. He says that was true of early robots, of mobile robots, and of additive manufacturing. The technology is novel, but making it useful takes much more.

12:13

How "Accidental" Pioneering Shaped His View

Asked how being a pioneer for the first decade and a half of his career shaped him, LaSelle said it was "largely accidental at every turn." He describes himself as a mechanical engineer, "not a great one," who loved physical things. He studied in the Bay Area, where nearly all his classmates went into tech companies like Cisco. Luecke related to this from his own time working for Rockwell Automation in San Francisco around 2015, surrounded by people working at Meta and Google.

LaSelle argues that everyone in manufacturing is something of a pioneer, because every product is new. He recalls that even when he entered the field, the industry was already trying to figure out how to do more manufacturing at home, with automation as part of the answer. Today, he says, manufacturers are still trying to satisfy new demand signals. They are also trying to make the physical world of manufacturing as efficient as the digital world that came out of Silicon Valley.

14:15

Jabil: At Scale, Exceptions Are the Job

LaSelle later worked at Jabil, a contract manufacturer serving mostly consumer electronics. He describes the efficiency required to run such a large organization at the cost levels demanded by customers like Apple. His assignment was to answer how to add more flexible automation to operations running at that rate.

His main lesson was that at scale, the problem is the exceptions. His team evaluated robotics, vision systems, and other technologies. If a system did something right 99 times and wrong once, that one failure could take down highly optimized lines. So companies have to be judicious about which technologies they deploy and how mature those technologies are. "When you get to volume, exception management is the job." He shared a saying from those implementations: "If you're 95% of the way there, you've still got 95% to go."

16:22

Healthcare and What "Regulated" Means

Between mobile robotics and Jabil, LaSelle worked in healthcare automation at Swisslog. There, he says, information became central. In factory logistics, many moved items aren't high value, and the metrics are yield and on-time delivery. In a hospital, you're dealing with people's health. "Regulated" meant don't lose critical medication and get it there on time.

He also sat on a steering committee in the early 2010s that discussed which environments mobile robots could safely operate in. Autonomous cars weren't yet driving around. Factories were considered safe because the people in them were trained. Hospitals were full of people who weren't trained to work around robots, so part of the regulated challenge was moving safely among them. But LaSelle says the core value was traceability. The goal was tracking medication from pharmacy to patient, with a full record from the pharmacist's hand to the nurse's hand. He calls this his first understanding of what "regulated" means.

18:25

Why Defense Spending Doesn't Simply Become Capacity

Asked to describe what it's like for manufacturers supporting the defense industrial base right now, LaSelle answered in one word: "crazy."

He first explained why domestic production matters for all regulated industries. Doing work at home supports security and traceability. It also speeds iteration, because you can partner with a manufacturer down the street.

On defense, he says the US industrial base has eroded, and rebuilding it requires visibility. He uses steel mills from decades ago as an example: "You need to have 20, 30, 50 years' worth of visibility to revenue to put hundreds of millions of dollars in place." In his view, the government doesn't always provide that visibility. In advanced defense technology, only a few large primes tend to participate, which he describes as a partial "cornering" of the market. The rest of the industrial base can't see that demand. Without a visible demand signal, manufacturers can't justify building a new factory or borrowing for new equipment.

He says the nature of the demand is also changing. He cites a general he identified as "General Caine," who spoke at an innovation conference in Philadelphia earlier that week. LaSelle says the general urged the industrial base to "go faster, much faster, and be bold." LaSelle connects this to what he says recent conflicts overseas have shown about iteration cycles for drones. The challenge, he argues, is no longer steel mills. It's making drones in small quantities, cost-effectively, with design iterations in days or weeks rather than months or years. At the same time, he says, the US is trying to modernize its fleet and replace depleted munitions, including interceptors like THAAD and Tomahawks. He cites lead times of 25 to 51 weeks in the current industrial base. For companies that can meet the quality requirements and manage data safely, he calls it a great time. But demand signals arrive so quickly that deciding where to spend and how to support the government is itself a challenge.

22:14

How Manufacturers Should Prioritize

Asked for advice on prioritizing, LaSelle compared it to Maslow's hierarchy. First, know what you do well and where you add value. Two machine shops don't necessarily handle the same materials or geometries the same way. Next, accept that delivering defense parts carries a high cost of quality, including documentation and traceability. That is a requirement. Then you need CMMC compliance or NIST scores. When you stack these factors, he says, it becomes fairly clear whether you belong supporting a land-based program, drones, or something else. Because most operators know their strengths, they can plug into the right place in the supply chain.

23:29

What It Takes to Invest: Contracting First

Luecke asked what manufacturers need to make the required investments: capital, equipment, supply chain, workforce, or all of these. LaSelle said he would start with contracting. To hire people, add equipment, or build factories, manufacturers need some longevity in their contracts or demand.

He credits the defense department with improving its use of long-term agreements (LTAs) over the last couple of years. He also mentions agencies like the Defense Innovation Unit (DIU), which he says is trying to figure out how the government can work better with the industrial base. LTAs, he says, used to be mostly limited to primes, the very large contractors like Lockheed Martin that build big aircraft. Over the last five years, he notes, non-primes and startups have emerged that are trying to move fast and match the new demand signals.

After contracting, LaSelle points to the right level of oversight to ensure a robust supply chain. Sensitive data has to stay out of the wrong hands. Materials need traceability so you aren't sourcing from an unreliable part of the world and suddenly losing access. He mentions that a policy change around CUI (controlled unclassified information) had just been announced that week. Even so, he says, data security remains essential "just as a business thing beyond just a government thing."

26:32

Workforce: The Seven-Year Machinist

On workforce, LaSelle said, "It's always people." He argues that because many people like him chose not to go into manufacturing and went to the Facebooks and Metas of the world instead, manufacturing talent has been depleted over the last couple of decades. The result is a talent war.

He says that despite all the automation, and the AI Fathom uses for programming to take rote tasks off its machinists, skilled people are still essential. He refers to his machinists as "creatives." He puts the path to a level three machinist at about seven years on the job, not counting trade school, which he's glad to see gaining prominence. He calls machinists "the folks with the paint brushes that make all this happen." He says many have left the industry or retired, and Fathom focuses on that single role every day.

27:56

Manufacturing as Art

Luecke, who describes himself as a musician, noticed the artistic language and asked about it. LaSelle said he does equate manufacturing with art. From executives to the shipping dock, he says, people need to bring passion to work, and in manufacturing that passion tracks closely with curiosity. Curiosity, in his view, has an artistic side however it's applied. He says some of Fathom's people can put a hand on a machine and know it needs tuning, "just like your guitar."

He sees this creativity as most valuable at the front of the product life cycle. By the millionth part, the process is dialed in. Going from the first part to the hundredth is "where these folks, they create magic."

Asked whether Fathom's high-tech operations help with hiring and retention, LaSelle said they do. He describes the conference room where they sat, where customers' designers and engineers work directly with Fathom's teams. Watching that collaboration attracts good people, he says, because they want to contribute to design. He notes that healthcare was Fathom's number one market in the most recent quarter. Employees, he says, value facing a new challenge every day instead of making the same widget a million times. That makes recruiting easier, though he acknowledges every other company is chasing the same people. When Luecke joked about the "sell" in LaSelle's surname, LaSelle replied that he gets that joke one way or another and doesn't have to sell too hard there.

32:17

The CMMC Phase 2 Pause

LaSelle explained CMMC in lay terms. It ensures that sensitive digital designs, sent to manufacturers for quoting and production, are protected. Some designs are intricate and exotic, and the government or primes like Lockheed consider them "special sauce." Phase 1 is self-assessed compliance. Phase 2 brings in third-party auditors and many more steps.

He says manufacturing at its core "hates change." Its people like new challenges, but the business itself does not. Policy shifts that affect investment and production decisions are especially disruptive. The week's announcement was a big change for Fathom. The company had spent nearly nine months preparing the processes, policies, and evidence for auditors to certify it. Now, he says, there's a high probability the requirement will stand down within 60 days, though he added, "don't bet on this." He calls that "a lot of money we just spent."

He says the effort won't be wasted entirely, because the company needs to be secure anyway. But much of the preparation was paperwork for the auditors. It's problematic, he argues, to invest in infrastructure you then can't monetize, because certification would have been a real differentiator. He cites a report he read a few weeks earlier saying 10% of the industrial base would drop out if the policy were enacted, because those companies wouldn't choose to or couldn't afford to do the work. That's why he calls the pause "almost an obvious choice." He also guesses that among ten experienced people in the room, about half would have predicted the pause and half, including himself, would not have.

For companies that didn't prepare, he expects the reaction to be "Told you." For those that did, he says the core point stands even without CMMC 2.0. The government will still need a way to entrust top-secret designs to manufacturers with confidence, with "belts and suspenders." Some of the documentation may be wasted, he says, but the principles and processes that teach employees how to handle data matter for everyone. He compares the situation to the medical field. Over the last 50 years, he says, there have been many initiatives hailed as "the best thing since sliced bread" and then dropped as too expensive. The intent was always better outcomes for patients, and he believes you have to see it that way or "you drive yourself crazy."

37:38

Physical AI and What Excites Him

LaSelle calls this an extraordinary time to be in manufacturing, after more than 30 years in the field. He says AI is affecting Fathom's business every day, not just in accounting or sales, by removing repetitive work people shouldn't spend time on. In his view, technologies celebrated in Silicon Valley are making their way to the heartland.

Asked what would excite a new entrant, he said the baseline is ending the day holding something beautiful you made, ready to go into a complex machine. Beyond that, he argues that AI will eventually hit diminishing returns in communications. "You can only stare at Instagram and Facebook so long." The next frontier, he says, is physical AI. He cautions against jumping straight to humanoid robots. He says they do illustrate a long-standing barrier: for the last 25 years, a big limit on robot adoption was the need for programmers, and he sees the same issue with CNC. Applying the data in large data farms to physical problems could, in Fathom's case, mean better control of additive machines. It could also mean automating CNC programming from 80% to 98%. That would leave the last 2% to a person putting a hand on the machine and turning the dial slightly for more productivity.

He also argues that the US never lost design innovation, even if it lost steel mills and many factories. "If we aren't the most innovative country in the world, and I will lay my hat on that every day long." That innovation still has to be produced somewhere. Luecke connected this to an earlier interview with MIT guests, who argued the US mistakenly thought it could outsource production and lost the iteration and design-for-manufacturing learning that comes from seeing things made.

42:45

Closing Exchange

LaSelle turned the last question back on Luecke, asking what was similar and different compared to other guests. Luecke said the difference was how explicitly LaSelle connected art and manufacturing, which most guests do only subtly. The similarity was LaSelle's view that now is a great time for young people to enter the industry. Luecke says that theme comes up in more than half of his conversations. Still, he thinks it bears repeating, because many listeners only hear an occasional episode and might need to hear it just once.

The two noted that they'd meet again at IMTS, which Luecke described as the largest manufacturing show in North America, running the 14th through the 19th that year. LaSelle added one more reason to join the industry: its community. At a show like IMTS, he said, you realize how much other people know that you don't, so it's worth going there to learn and network.