Latest / The Manufacturing Podcast with Fexingo: Factories, Supply Chains, and Industrial Business / How Factories Are Winning With On-Demand Capacity
Transcript
- Lucas: Look at the floor plans of the last decade. You see massive, monolithic buildings designed for one thing, one volume, forever. But today, if you walk into a forward-thinking facility, the biggest asset isn't the heavy machinery bolted to the concrete. It's the ability to turn that machinery on or off based on real-time orders. We are seeing a fundamental shift from owning capacity to renting it. Luna: Right, because the risk of being stuck with idle equipment during a demand dip is suddenly much more expensive than paying a premium for extra hours on someone else's line. What does that math actually look like on the balance sheet? Lucas: It changes the entire structure of your cost of goods sold. Traditionally, you have high fixed depreciation and low variable labor. Now, you're trading that fixed burden for a higher variable cost per unit. For a company like Tesla, when they ramp up Model Y production, they don't just build a bigger Gigafactory tomorrow. They optimize existing lines and use partner capacity for overflow components. It keeps their operating leverage intact while allowing them to hit volume targets without overbuilding. Luna: But doesn't that introduce a layer of complexity? If you're pulling parts from five different subcontractors instead of one internal warehouse, your coordination costs must spike. Lucas: That is the exact friction point. And it is where technology has to step in. You cannot manage on-demand capacity with spreadsheets. You need a digital thread that tracks inventory, machine availability, and quality metrics in real time across all those external partners. It is less about physical logistics and more about information logistics. If the data latency is more than a few minutes, the whole model collapses. Luna: So we are talking about a platform economy for industrial output. Like Uber for factory machines? Lucas: Exactly, but with much higher stakes. A ride-share app failing means you wait twenty minutes. An industrial capacity platform failing means a missed shipment deadline for an automotive assembly plant, which triggers penalties worth millions. The players in this space are not just brokers; they are integrating deep technical standards so that a part made on a rented CNC mill meets the exact tolerances required by the final assembler. Luna: I remember reading about how some consumer electronics brands handle peak holiday seasons. They don't own the injection molding machines. They lease the capacity from contract manufacturers who have spare cycles. Is that the standard model now? Lucas: It is becoming the standard for volatile demand sectors. Consumer electronics, seasonal goods, even medical devices post-pandemic. Companies are realizing that building for peak capacity leaves you underutilized for eight months of the year. By buying only what you need, when you need it, you improve your return on invested capital significantly. It is a pure financial engineering play backed by operational reality. Luna: What about the intellectual property risk though? If I am sending my proprietary designs to a shared network of factories, how do I ensure no one is reverse-engineering my product for a competitor? Lucas: That is the billion-dollar question. The answer lies in segmentation and contractual armor. You break your product down into modules. The core innovation stays in-house or with a deeply trusted partner. The commodity parts go to the open market. Plus, modern digital rights management can watermark production files so that any leak can be traced back to a specific node in the network. It is not foolproof, but it raises the barrier to theft enough to deter casual copying. Luna: It feels like we are moving away from the vertical integration dream that dominated the twentieth century. Ford used to make its own glass, steel, everything. Now we are embracing fragmentation for speed. Lucas: Ford still makes its own steel, largely, because margins are thin there. But for most mid-tier manufacturers, vertical integration is a trap. It locks you into legacy processes. On-demand capacity allows you to switch suppliers instantly if a better technology emerges. If a new battery chemistry comes out, you don't have to scrap your whole factory. You just source the new cells from a specialist who already mastered the process. Luna: So agility becomes the primary competitive advantage rather than scale. That changes how you hire, too. You don't need as many general managers overseeing production; you need more procurement specialists and data analysts. Lucas: Precisely. The workforce composition is flipping. You are seeing a decline in shop-floor supervisors and a rise in vendor relationship managers. These people need to understand both the engineering constraints of the parts and the commercial realities of the supplier. It is a hybrid skill set that is hard to find right now. Luna: Are there any industries where this model just doesn't work? Where you absolutely must own the asset? Lucas: High-volume, low-margin commodities. If you are making paper clips or basic fasteners, the margin is so slim that you cannot afford the transaction costs of coordinating external capacity. You need the sheer efficiency of scale that only comes from owning and running the line twenty-four seven. But for anything complex, custom, or high-value, the asset-light model wins. Luna: It reminds me of the software industry's move to cloud computing. Instead of buying servers, you rent compute power. This is the hardware equivalent. Lucas: It is literally the same trajectory. Cloud computing reduced the barrier to entry for startups because they didn't need a data center. On-demand manufacturing reduces the barrier to entry for hardware startups because they don't need a factory floor. You can prototype and produce small batches without writing a six-figure check for tooling. Luna: Does this mean the big contract manufacturers are losing power? Or are they adapting? Lucas: They are adapting by becoming platforms themselves. Foxconn, for example, is not just taking orders anymore. They are offering capacity as a service, providing the software infrastructure that lets clients monitor their production in real time. They are locking customers in with convenience and visibility, not just price. The ones who survive are the ones who become indispensable partners, not just passive vendors. Luna: What about the sustainability angle? Isn't shipping parts between multiple factories less efficient than making them in one place? Lucas: It can be, if you manage it poorly. But if you optimize the network geographically, you often reduce waste overall. Less overproduction means less scrap. Also, sharing capacity means fewer machines need to be built in the first place. That is a massive reduction in embodied carbon from steel and concrete. It is a net positive for the environment if the logistics are tight. Luna: I want to touch on the financial side again. When you rent capacity, your cash flow profile changes. You pay as you go. That sounds healthier, but does it hurt your ability to borrow money? Lucas: Banks love it. Asset-light companies have stronger balance sheets because they have less debt tied up in depreciating machinery. Lenders prefer steady, predictable cash flows over lumpy capital expenditures. So yes, financing becomes easier and cheaper for these firms. It is a virtuous cycle for growth-stage companies. Luna: It seems like the future belongs to the orchestrators, not the owners. The companies that can best coordinate the symphony of external capabilities. Lucas: Orchestrators. I like that. Because the complexity is growing, not shrinking. As products get more sophisticated, the number of suppliers increases. The orchestrator needs to maintain harmony among dozens of entities. That is where the moat is building itself. Luna: Before we wrap up this thought experiment on industrial orchestration, I want to pause for a second. If today was actually useful to you, the way these stay ad-free is listener support — buy me a coffee dot com slash fexingo. Lucas: Yeah, and if you find value in digging into these structural shifts, consider supporting us directly through buy me a coffee dot com slash fexingo. It helps keep the research sharp and the conversations honest. Luna: Thanks for listening. Now, back to the topic. Lucas: Let's look at a concrete example of this orchestration failing. There was a case last year with a major apparel brand that tried to shift entirely to on-demand knitting. They relied on a network of small, automated knitting machines scattered across three countries. They had great software, but they underestimated the maintenance variance. Different machines had different wear rates, leading to inconsistent stitch density. The brand had to recall thousands of units because the quality wasn't uniform across the distributed network. Luna: So the tech solved the coordination problem, but it couldn't solve the physical degradation of tools. That is a hard constraint. Lucas: Exactly. You can digitize the order, but you cannot digitize the metal fatigue of a spindle. The lesson was that you still need local quality assurance teams embedded in those external factories. You cannot be fully remote. There must be eyes on the ground to verify that the digital twin matches the physical reality. Luna: It brings up another issue: training. If you are sending workers to different external sites every week, how do you ensure they follow the same safety and quality protocols? Lucas: Standardization is key. The orchestrating company must provide the training curriculum and certify the external workers. It is essentially franchising your quality standards. The brand grants permission to operate only to those nodes that pass the audit. It creates a hierarchy of trust within the network. Luna: This feels like a significant cultural shift for traditional manufacturing CEOs. They are used to commanding, not collaborating. How do you change that mindset? Lucas: You start with the CFO. Show them the numbers. When the chief financial officer sees that variable costs align with revenue, the CEO listens. Once the financial model is proven, the operational team follows. It is top-down adoption driven by bottom-up economic pressure. Luna: Do you think this will lead to consolidation among the capacity providers? Will we see fewer, larger platforms? Lucas: Likely. Just like the cloud market consolidated around AWS, Azure, and Google, the industrial capacity market will consolidate. Small players will get acquired by the giants who need to fill out their geographic coverage. We are probably three years away from seeing the first major merger in this sector. Luna: Interesting. So the next wave of M&A won't be about buying factories, but buying networks of factories. Lucas: Precisely. It is a redefinition of what a manufacturing company is. It is no longer a collection of assets. It is a collection of relationships and algorithms. Luna: One final question on the horizon. With AI getting better at predictive maintenance, will we eventually reach a point where any factory can run perfectly regardless of who owns it? Would that kill the need for specialized orchestrators? Lucas: AI will lower the barrier to entry, but it won't eliminate the need for coordination. Even if every machine runs perfectly, you still need to decide which machine makes which part, when, and for whom. The allocation problem remains. Humans will always need to define the strategy, even if machines execute the tactics. The orchestrator role evolves, but it doesn't disappear. Luna: So the human element shifts from wrench-turning to decision-making. That is a hopeful note for the workforce. Lucas: It is, provided we invest in the education to make that shift happen. Otherwise, we just trade physical risk for cognitive burnout. But the potential for higher-value work is definitely there. Luna: We will leave it there. Thanks for tuning in.