Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Avoid Tooling Debt With MaaS Platforms
Transcript
- Lucas: We've covered a lot of manufacturing processes on this show — injection molding, die casting, CNC — but one thing we haven't talked about is how a startup decides whether to buy a mold or just rent the capacity. Luna: You mean like using a platform instead of committing to a full tool? Lucas: Exactly. manufacturing as a service — MaaS. Companies like Fictiv, Xometry, Protolabs. You upload a CAD file and get a quote in hours, parts in days. No minimums, no tooling investment. Luna: And it's not just for prototyping anymore. I saw a story about an ag-tech startup that made a portable soil sensor. They used Fictiv for their first 200 units — all CNC machined and urethane cast. Lucas: Right, that's the one. They needed a sealed aluminum housing and some internal brackets. Traditional route: inject a mold for the housing at $35,000, plus $5,000 for the bracket tool — $40,000 upfront. They didn't have that. Luna: So they paid per part instead. What did that look like? Lucas: About $12 per unit for the CNC housing, plus $4 for the machined brackets. For 200 units, that's $3,200 — no tooling. Compare to injection molding at maybe $3 per unit but with the $40,000 tool. They needed only 500 for their pilot, so MaaS saved them tens of thousands. Luna: That tradeoff is the whole game. Higher unit cost, but no upfront debt. When do you switch? Lucas: Usually around 1,000 to 5,000 units, depending on part geometry. The soil sensor guys ordered a second run of 300 via MaaS, then looked at demand. Once they had pre-orders for 3,000, they pulled the trigger on injection molding. Luna: And they used the same design files — just optimized for draft angles and wall thickness. So the DFM work wasn't wasted. Lucas: Exactly. That's the beauty. MaaS platforms let you iterate on a design without being locked into a tool. You can change the shape, add a boss, move a hole — next batch reflects it. No scrapping a $40,000 mold. Luna: What about lead times? I've heard Xometry can ship CNC parts in three days. Lucas: Typically 3 to 5 business days for CNC, 5 to 10 for urethane casting. Compare to 4 to 8 weeks for a new injection mold tool. That speed can be critical for market validation. Luna: And if you need to change the design after user feedback, you don't cry over sunk tooling costs. Lucas: Right. That's why I call it 'tooling debt.' When you pay for a mold upfront, you're betting that design won't change. MaaS keeps your options open. Luna: But there are downsides, right? Quality consistency? Scalability? Lucas: Definitely. MaaS parts are made on general-purpose machines, not custom production lines. Tolerances can vary more than production injection molding. And if you need 100,000 units, the per-unit cost of MaaS is prohibitive. Luna: So it's a bridge, not a destination. Lucas: Exactly. The smartest hardware startups use MaaS to get to first revenue, then switch to dedicated tooling once the design is stable and volume is proven. The soil sensor company went from concept to paying customers in 6 months instead of 12. Luna: And that's with a physical product that has an aluminum housing, a PCB, a sensor — real complexity. Lucas: Yeah. And now platforms are adding services like DFM feedback, auto-quote optimization, even automated quality inspections. It's getting easier to build hardware without a factory. Luna: Speaking of building things without a factory — these conversations are only possible because of listener support. If you've gotten something out of the show, a couple of dollars a month genuinely keeps it going. You can find us at buy me a coffee dot com slash fexingo. Lucas: And it really does make a difference. Every bit helps us stay ad-free and focused on the practical stuff. Anyway, back to MaaS — I think we're going to see more startups use these platforms not just for prototypes, but for regional micro-factories that combine local MaaS with final assembly. Luna: That's interesting. Like, produce the housing in the US via Xometry, then ship to an assembly partner in Mexico? Lucas: Exactly. Or produce different parts on different platforms — one for metal, one for plastic, one for electronics. The manufacturing supply chain is becoming as modular as the hardware itself. Luna: And that lowers the barrier for first-time hardware founders who don't have connections in Shenzhen. Lucas: Right. The soil sensor founder had never manufactured anything before. He found a DFM engineer through Fictiv's network, got his parts, and now has a product in the field. That's the real story — MaaS is democratizing hardware. Luna: So the key takeaway: start with MaaS to validate and iterate, but have a plan to migrate to dedicated tooling once you know what you're making. Lucas: And use the data from MaaS — cost per unit, defect rates, lead times — to build your business case for that tooling investment. Luna: Great angle. Next time we should talk about how to choose which MaaS platform for your specific product. Lucas: That's a whole episode itself. I'll start pulling together a comparison — Fictiv vs Xometry vs Protolabs vs maybe a new entrant. Luna: I'll look forward to that. For now, if you're building hardware, remember: you don't have to buy the factory to make the product. Lucas: Words to build by. Until next time.