Manufacturing isn’t a singular industry, but a complex network of systems and processes that collectively power every critical physical sector on earth. 

Any disruption —  from major global events like the pandemic and geopolitical conflicts, to everyday snags like malfunctioning machinery and labor shortages — reverberates across the entire ecosystem, affecting not just the direct and indirect industries that manufacturing underpins, but our national stability. In the United States, where our peak manufacturing days are long behind us, disturbances and inadequacies across the system have become harder to absorb — and with international trade relations in flux, impossible to ignore.

Forced to navigate a new landscape where reliable access to cost-effective (and thus long-outsourced) resources and services are no longer a given, organizations from multinational automotive manufacturers and electronics companies to small businesses selling directly to consumers are looking to on-or near-shored options for everything from raw materials to final assembly of products. 

Options are slim, but fortunately that is changing — because it finally has to, and the urgent reason for doing so goes far beyond mitigating tariff-induced price fluctuations. Manufacturing plays an outsized role in the U.S. economy: Although the sector represents just 10% of GDP in the U.S., it drives 20% of the nation’s capital investment, 35% of productivity growth, 60% of exports, and 70 % of business R&D expenditures. For each manufacturing job created, an additional seven jobs are created in adjacent industries, from shipping logistics and construction to restaurants and retail businesses around manufacturing facilities.

While it’s not economically necessary (or even feasible) to recreate the type of mass production-scale manufacturing that China has built up — and which the U.S. has never had —  it’s critical that we have the ability to manufacture the most important, high-value products we need the most. We need the capacity to reliably produce medical devices, vehicles, advanced computing components, defense technology, building materials, and so much more without relying heavily on foreign partners — especially adversarial ones — to do so. 

The question on everyone’s mind: Is this actually possible? 

At Eclipse, we believe that it is —  provided we don’t try to rebuild each and every element of the system from scratch. A modern, sustainable, and economical manufacturing capacity will come from leveraging what we already have while reinventing aspects of the industry to offset what we don’t. With labor shortages, supply chain constraints, and a lack of a sufficient number of manufacturing facilities, the only option is to do more with less. Technology is our greatest advantage — for now. It’s time to apply it to manufacturing while we still have the edge. 

Revitalizing domestic manufacturing has been one of our primary focus areas since we formed Eclipse in 2015. The multi-layered dynamics of manufacturing — and the U.S.-specific challenges they are addressing — are on display in the Eclipse portfolio: From startups that represent the next generation of contract manufacturing, like Bright Machines, which deploys software-driven, full stack robotic solutions that can mitigate labor shortages and drastically shorten the lead times from product design to assembly, or VulcanForms, which provides an additive manufacturing platform for rapid iteration and prototyping with limited raw materials; to companies like Cellares, which is developing a cell therapy manufacturing platform to deliver personalized medicine at scale; to companies like Augury using sensors and AI to monitor industrial machine health; and others like Starboard and Jitsu providing shipping logistics visibility; to our expanding portfolio of alternative energy sources like Tandem PV, Peak Energy, and The Nuclear Company that will be instrumental in delivering power to our increasingly digitized industrial base.

“Like many physical industries, manufacturing creates a circular economy,” says Eclipse Founding Partner Lior Susan. “From the tools you need inside, to the materials and mechanisms you need to build and supply power to the facility, to the visibility you need to get the products from one place to the other, there are multiple pieces that fit together and depend on one another.”

It’s an area Susan, the rest of the Eclipse team, and founders in our portfolio know well, having worked inside the previous wave of pioneering, manufacturing-driven startups like Tesla, Rivian, Apple, and Samsara, as well as at legacy ODMs and OEMs. Not only do we understand the challenges of traditional factory environments, we understand that innovation needs to happen at every stage of the process. We’ve been through the depths of production hell to understand what it takes to manufacture novel technology — from electric cars and trucks to robots and watches — to building factories.

Along with ensuring we are able to build critical products even in times of crisis, we believe reinventing domestic manufacturing will give rise to the next truly generational, multi-trillion-dollar-valued businesses. Consider the fact that Tesla and SpaceX (two of the most valuable public and private companies, respectively, in the world) are rooted in manufacturing. What’s more, as we mention later in the section on vertical integration, the success of those companies isn’t just due to their category-defining products, but because of the significant manufacturing innovations they created in pursuit of developing their products. With so many opportunities to either improve or reinvent technology throughout the entire manufacturing ecosystem, we are bullish that the world’s biggest companies will be built here.  

Here’s our perspective on how we can leverage our best tools into a domestic manufacturing system that will fuel the next wave of prosperity.

Mitigating Reality With Software-Driven Robotics and AI

At the most basic level, the decline of American manufacturing can be distilled to just a handful of factors: People, raw materials, and the feasibility of building factories.    

“The reality is, to ramp up manufacturing in the U.S., we’re starting from an industrial base that has eroded significantly over time” says Eclipse Partner Aidan Madigan-Curtis. “Even with advanced manufacturing capabilities, the challenge will be to scale adoption fast enough to serve the biggest market in the world.” 

With historic shortages of industrial workers including machine operators, welders, and electricians, revitalizing manufacturing has never been a simple matter of re-opening shuttered factories. Compounding this, existing U.S. manufacturers may be forced to operate with even fewer workers as mounting tariffs on many of the imported raw materials required to build their products drive up operational costs. Even trickier is the fact that few people actually want to work in manufacturing. 

We need technology that will compensate not just for today’s labor shortages with a compromised manufacturing system, but the expanded version we’re building for the future, says Madigan-Curtis. 

Previously, mechanical engineering advances drove manufacturing innovation. Software — particularly that used in conjunction with robotics to automate and amplify human efforts — is the force multiplier today. This brings us to Bright Machines, which has been on a mission to transform manufacturing with a software-driven, full stack automation solution since its founding in 2018. Initially aimed at producing any and all electronics, the company has since sharpened its focus on high-value, complex electronics including CPUs and GPUs. The company’s “microfactories” (known internally as Bright Robotic Cells) leverage robotics, automation, computer vision, and AI to handle everything from design recommendations and processing on through the assembly, testing, and disassembly of products. Through partnerships with major players like Nvidia and Microsoft, Bright Machines has become an integral force in the quest to strengthen the U.S.’s AI backbone, and exemplifies the next generation of contract manufacturing.

Robotics, computer vision, and AI are also playing a critical role in creating safer, more efficient systems workspaces for the humans who do work in industrial settings like warehouses and factories. Mytra, for example, has developed an AI-powered robotic pallet system that essentially turns every pallet of material into an intelligent, efficient elevator. This spells a significant productivity gain, considering that 80% of work in warehouses and 40% in factories is devoted to the simple tasks of finding, retrieving, and moving material. Founded by former Tesla engineers, Mytra aims to offset labor shortages while building resilience and efficiency across the supply chain. 

Other startups, like Voxel, are deploying computer vision and AI to improve safety in industrial settings. In 2023, there were nearly 2.4 million workplace injuries (not counting illnesses). Preventing injuries from human errors and hazardous working conditions is crucial when operating with limited staff, and Voxel leverages existing camera infrastructure to run advanced algorithms to detect these incidents or prevent them from happening in the first place. 

Enabling our existing human workforce to be more productive and efficient is crucial to managing the soaring cost of labor. At the same time, the enhanced precision and speed gleaned from automation and AI is imperative as we mitigate the costs and constraints of foreign supply chains. 

VulcanForms is addressing this with its additive metals manufacturing system (an industrial form of 3D printing). Founded out of MIT research, the company’s laser-based metal printing platform is 100 times more powerful than the conventional processes, which requires carving pieces from blocks of solid metal. With VulcanForms, businesses ranging from aerospace and defense companies to medical implant makers and semiconductor fabricators rapidly prototype, test, and build products at scale domestically with fewer raw materials. Other companies in stealth are also shortening the physical prototyping and testing phase with generative AI that provides modeling that obeys the laws of physics.

Software helps offset raw material constraints, but more can be done to remedy the shortages in the first place. Although the U.S. has domestic mineral deposits, we only sparsely mine a few of them, like lithium and cobalt. Others, including arsenic, tin, gallium, and scandium (which are necessary for semiconductor and high-performance alloy production) haven’t been mined domestically in more than 20 years. In the meantime, the U.S. is 100% dependent on imports for 21 out of the 50 most critical mineral commodities, and at least 50% dependent on imports for another 28. This means we are almost completely beholden to others for the raw materials to produce everything from smartphones to defense technologies like radar and missile guidance systems. That needs to change. Eclipse is actively looking to partner with founders who are developing the proper tooling to discover, extract, and process critical minerals. 

As the adoption of software-driven automation and robotics in manufacturing processes significantly cuts down costs, shortens lead times, and preserves precious resources through more precise operations, it changes the nature of what human manufacturing jobs entail in the process. Redefining manufacturing jobs as more tech-forward and dynamic will go a long way in attracting talent in the future. 

We only have room to grow. Shockingly, about 80% of U.S. manufacturers do not use any automation today. For every 10,000 manufacturing employees in the U.S., just 141 robots work alongside them. Meanwhile, China installs more robots in a single year than the U.S. has implemented in our manufacturing history.  

“Technology is not our blocker. Investment and adoption is,” says Madigan-Curtis.

Productivity Partners

The larger and more globally competitive our domestic manufacturing system becomes, the more sophisticated technology must become to mitigate our single biggest constraint of labor. 

This is where physical AI comes into play: The fusion of robotics, artificial intelligence, and real-world actuation. Robots equipped with advanced perception, manipulation, and learning capabilities are now able to perform complex tasks in factories, warehouses, and job sites.

It’s no exaggeration to say that physical AI could be the key to sustaining America’s manufacturing and construction boom in the face of demographic headwinds. Facing a critical shortage of machinists, welders, toolmakers, and technicians, every intelligent robot deployed in a factory or on a job site is essentially filling a job vacancy that might otherwise sit empty. 

Rather than viewing automation as a threat, it should be recognized as an enabler of growth that will ultimately help companies stay competitive globally, says Eclipse Partner Charly Mwangi.

“By embracing intelligent robots not as job-stealers but as productivity partners, we can unlock new levels of industrial capacity and resilience” says Mwangi. “Factories can run at full potential, construction projects can advance faster, and workers on the ground can be supported by tireless AI helpers.”

The success stories are multiplying, from AI robotic arms that can sort an infinite variety of products, to autonomous machines building infrastructure. As mentioned earlier, several physical AI startups in the Eclipse portfolio are making a significant impact, including Bright Machine’s software-driven, robotic microfactories, Mytra’s warehouse automation platform, and Reframe Systems’ robotic-powered construction services. 

Physical AI technology is still maturing, meaning there is plenty of room for new players. A particularly exciting challenge is developing more sophisticated foundation models for physical AI. Just like large language models are constantly improving through ongoing training on text and image-based data, AI-powered robots will need training on an increasingly larger set of parameters in order to get smarter and function as reliably as humans. Eclipse is actively looking for founders innovating in this area, and we’re thrilled to see more investors excited about this space and joining us on the journey we embarked on a decade ago. 

“Founders who can combine AI savvy with practical robotics stand to build the next generation of “builders” for our economy – the robotic workforce,” says Mwangi.

It’s an exciting, energizing vision: a nation building and making more than ever, powered by human ingenuity plus robotic muscle. For those with the entrepreneurial spirit, now is the time to build – quite literally – the physical AI systems that will define the future of work.

Power and Places

One of the biggest impediments to manufacturing our own products is the most basic: Lack of infrastructure. This is in both the actual facilities to manufacture products and the energy sources that power them. 

The push to reinvent and rebuild the U.S. industrial base is synonymous with digitizing many aspects of these critical physical sectors. From embodied AI and computer vision to additive metals manufacturing, the technologies integral to re-shoring our manufacturing capacity require a lot of energy. 

Fossil fuels alone aren’t enough to meet our exponentially-growing needs for energy, especially electricity in the form of data centers. This is why Eclipse backs founders who are dedicated to expanding our energy portfolio. This includes The Nuclear Company, which is pioneering the modernization of nuclear construction and leading fleet-scale deployment of nuclear power across the U.S.; Tandem PV, which is manufacturing its own perovskite-based solar panels domestically; and, again, Peak Energy, which is addressing a key bottleneck in renewable energy adoption by domestically manufacturing sodium-ion batteries, which are more cost-effective, safe, and better optimized for storage than lithium-ion batteries. 

As mentioned earlier, these companies highlight the importance of creating an ecosystem. For instance, since the Nuclear Company is trying to overhaul the construction and permitting process of building nuclear power plants, they are working lockstep with regulators and local governments. 

Construction of any facility is challenging and expensive. On top of lengthy permitting processes and skilled labor shortages, the engineering and design stage accounts for upwards of 14% of the costs.  Building factories can get even more complicated, which is why the construction industry is in dire need of technology if we hope to start building at a rate that can feasibly support a functional domestic manufacturing system soon. Companies like Augmenta are developing the next generation of computer-aided-design (CAD) tools, replacing legacy tools and manual processes with artificial intelligence and high-performance computing techniques to generate fully constructible, code-compliant designs of the most costly systems of a building: The mechanical, electrical, plumbing, and structural (MEP+S) systems. 

More companies are responding to the call to on-shore their facilities. In recent weeks,  Amazon, Novartis, and Nvidia all announced new plans for major U.S. expansion of factories. These are just the examples of major multinational organizations — many manufacturing companies in the U.S. are small and medium-sized businesses (74% have fewer than 20 employees), meaning they are even more incentivized to adopt technology to offset costs and labor of facilities construction.

Putting it All Together — Business Models

How you leverage new technology depends on the approach to building your business.  

In the pursuit of overhauling defense, healthcare infrastructure, and other critical physical sectors heavily reliant on manufacturing, it’s often necessary to take a verticalized approach. For companies building net-new, category-defining products, the unique innovations across the processes to create their goods often ends up being the most powerful determinant of their competitive edge. Tesla and SpaceX, two of the most valuable public and private companies, respectively, in the world today got where they are from creating bespoke, vertically-integrated manufacturing processes in-house. 

As Eclipse Partner and early Tesla employee Greg Reichow can attest, “Nothing beats the speed and ability to change like having your own manufacturing operation.”

The benefit of verticalization was something Tesla learned by doing, says Reichow. Now renowned for its Gigafactories, the company actually started out with the goal to do as little of their own manufacturing as possible, working with different contract manufacturers for each component such as batteries, motor parts, and the car body. It didn’t work. Not only were some suppliers unwilling to provide Tesla with the parts, most off-the-shelf components didn’t meet performance targets. Tesla realized that in order to build an entirely new, better type of electric vehicle that didn’t yet exist anywhere else in the market, they had to build each and every part of it from the ground up, in-house. 

“When you are trying to build a breakthrough product, it’s very difficult to innovate if you are just collecting different parts of an existing supply chain,” says Reichow. “To stand up your own manufacturing capacity in the U.S., vertical integration will be your core advantage.”

Building various components in-house was challenging, but the deeper Tesla vertically integrated, the better they functioned as a company. They gained in-depth knowledge that allowed them to rapidly innovate and continually improve the overall care faster than competitors.

Today, we’re seeing strong traction from startups employing a vertical approach as they develop category-defining products of their own. This includes Arc, which aims to electrify the entire marine industry; advanced propulsion systems maker Ursa Major; and Rivian-spinout Also, which is developing a full stack electrified micro mobility platform.

Total verticalization isn’t always possible (which is why companies like Bright Machines have a larger role to play as a next-generation contract manufacturer for a variety of electronics), and why other companies focus on manufacturing only the highest-value component of their products while creating an ecosystem around their industry. However, this means they must have a competitive edge in the other critical parts of production, such as supply chain, logistics, and — in some cases — an engaged ecosystem of the primary customers, especially if you are reliant on government contracts and regulatory approval.

Conclusion

Reinventing domestic manufacturing is clearly worth our time and effort. Beyond control over the quality, cost, and accessibility of goods, domestic manufacturing provides greater protection over intellectual property and is one of the largest drivers of economic growth and job creation across nearly every skill level. And, again, a robust domestic manufacturing system ensures that not only can we produce essential products and services in times of crisis, we won’t be beholden to trade partners who can weaponize access to some or all parts of their manufacturing capabilities during times of conflict. Effectively, our success, sovereignty, and security is directly tied to our manufacturing capacity. 

While government support and national initiatives to spearhead everything from factory building and on-shore materials sourcing is crucial, we need the speed and nimbleness of startups that take a full-stack approach to building, and which drive innovation with the urgency needed to quickly regain our footing in this sector. 

As we move into our second decade of building and investing in manufacturing, we are optimistic that transformative innovation will happen much faster than before due to the rapid advancement of technology to make this possible, but importantly, because we no longer have a choice.

Follow Eclipse on LinkedIn or sign up for Eclipse’s Newsletter for the latest on building the New Economy.