Note: The views expressed here are the author’s own and do not reflect the views of Energy Impact Partners.
Our newest portfolio company, Ayr Energy, is unlike any startup I’ve ever seen. On the one hand, it’s a grid equipment OEM. It sells transformers, circuit breakers, and switchgear to customers – a 100-year-old business model. On the other hand: in just over a year since founding, it has racked up over $500mm in purchase orders, secured manufacturing capacity comparable to that of $100b+ incumbents, and delivered units to happy customers. The company has speedrun the path from startup to powerhouse. You’re not supposed to be able to do that. But Ayr won’t be a one-off in retrospect. It’s at the leading edge of a new category of manufacturer.
Ayr exemplifies the neo-OEM. Like the neo-cloud and the neo-prime before it, the neo-OEM represents a fundamentally better way to build. Neo-OEMs see supply amid shortages and see speed amid stagnation. They build on manufacturing capacity that didn’t exist a decade ago, and they exploit rapid advances in generalizable technologies like robotics and AI. We will see more neo-OEMs emerge over the next five years; the market pull is too strong to ignore.
What is a neo-OEM? In short, a manufacturing business with a cheat code, built to serve a market that incumbents can’t keep up with. A neo-OEM sells manufactured products and associated services just like any other, but it can do at least one thing that incumbents can’t or won’t. Here are some of the cheat codes we’ve seen so far:
Little to no steel in the ground. There is abundant manufacturing capacity around the world, and in some industries the issue is customer access and factory utilization. In exchange for connecting supply with demand and ensuring quality, the neo-OEM gets to own the product without owning the factory, converting capex into COGS. Capacity scales nimbly in response to demand, not in anticipation of it.
Small, dense, reconfigurable factories. Where a neo-OEM does have its own facilities, they’re smaller than you might expect because production is dense. The factory floor is tightly coupled to engineering, and the team iterates on processes where it makes sense.
Orchestration between distributed assets to deliver a seamless customer experience. In the same way that CoreWeave abstracts away the physical complexity of distributed GPU clusters, a neo-OEM may coordinate dozens of partner factories behind a single brand and product experience.
Manufacturing defined by software rather than hardware. Think software-controlled robotic arms rather than huge, single-purpose machines, cutting capex per unit from millions to $100–200k. These tools can increasingly handle the imperfections of the physical world. And they travel: a neo-OEM can deploy them in partner factories, upgrading capacity it doesn’t own.
Operations purpose-built to shorten lead times. Riding the AI exponential to speed up design, quoting, and engineering is table stakes (though it can be surprisingly hard for incumbents to implement1). A neo-OEM may even vertically integrate to control a chokepoint in its supply chain.
This profile is a big deal for customers, investors, and of course the neo-OEMs themselves.
Customers get a value proposition that incumbents can’t match:
No compromise between lead time, quality, and price.
Flexibility from quote through aftermarket.
A first unit that pencils as well as the thousandth.
Customization that doesn’t compress margins.
A supplier that iterates at the pace of a startup.
Investors get exposure to scaling manufacturing businesses in areas facing severe shortages without putting up billions in equity capital to fund capex. Critically, neo-OEMs bypass the first-of-a-kind (FOAK) project finance problem that has so plagued clean energy and industrial tech startups. The AI boom has shown us that there is funding out there for capex, but it needs to have the promise of wild profitability, not a 10% IRR if the stars align.
The companies themselves benefit from this model. If the factory is the product, then neo-OEMs have a native cost advantage in building their products. If speed can reindustrialize America, then neo-OEMs have the inside track to winning favor in DC. If critical equipment costs too much, then neo-OEMs can capture arbitrage opportunities through orchestration – even pulling forward cash flow from customers desperate to reserve their place in line.
A few words of warning. First, the neo-OEM model isn’t defensible in the same way that a patent-protected widget company is. A neo-OEM’s moat is its reputation with customers and its supply chain. This means that many neo-OEMs will be “hustle businesses” and their initial market opportunity may be time-limited. Without next-gen products, a neo-OEM can become a PE buyout story rather than a runaway venture winner. Build your team and market strategy accordingly.
Second, a neo-OEM’s cheat codes are only as powerful as the market failure is large. In Ayr’s case, we think the shortage in grid equipment is structural rather than cyclical, and the size of the opportunity is in the hundreds of billions. How many more Ayr-sized market failures are out there?
Third, neo-OEMs are less capital-intensive than traditional OEMs, but they still need to maintain significant working capital to handle typical OEM issues: warranty claims, accounts receivable, supplier payments, and so on.
Zooming out: the interesting question isn’t whether neo-OEMs will work. Ayr is answering that as we speak. The interesting question is how many more $100b opportunities are hiding in plain sight. If you think you’ve found one, please get in touch.
Elsewhere:
Thanks for reading!
Please share your thoughts and let me know where I mess up. You can find me on LinkedIn and X.
For some color on how surprisingly hard it is for incumbents to adopt AI, see this WSJ article about Anthropic considering a $200mm investment to set up a consulting vehicle to help PE-backed companies do exactly this.
Awesome writeup, Michael!
You're first bullet on "little no steel on the ground" caught my eye. Could you say more about how neo-OEMs are able to nimbly scale capacity while ensuring quality and competitive cost? Are they working with multiple trusted partners with spare manufacturing capacity, and guaranteeing some longevity as a tenant in the factory?
Just pulling on parallel experience in software where varying workload serverless compute is one of the more expensive options vs. committing to a defined amount of compute over a longer time horizon with a cloud provider. So I'm wondering how neo-OEMs are threading the needle on making a longer term commitment (if that also has a cost advantage in manufacturing) and quickly scaling up or down when needed.