Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / Hardware Startups Navigate Component Obsolescence
Transcript
- Lucas: There is a particular kind of panic that sets in at three in the morning when a hardware founder realizes the microcontroller on their bill of materials has been discontinued by the manufacturer. It happened to me early in my career covering the IoT boom, watching a well-funded startup in Austin burn through two million dollars in tooling costs because they bet everything on a single sensor from a mid-tier Asian supplier. Luna: That sounds like a nightmare scenario. Did they manage to pivot, or did the company just fold? Lucas: They folded, but it wasn't because the product was bad. It was because they treated components as interchangeable commodities rather than strategic assets with finite lifespans. The reality is that in hardware, unlike software, you cannot push a patch to fix a missing part. You have to redesign the board, re-spin the PCB, and re-certify the device, which can take six to nine months. Luna: So we are talking about a fundamental mismatch between the speed of software iteration and the sluggishness of physical manufacturing cycles. Lucas: Exactly. And as we sit here in September of twenty twenty-six, looking at the lingering effects of the global semiconductor shifts, this issue has only grown more acute. Companies are no longer just worried about short-term shortages; they are worried about long-term availability. A chip chosen today might be obsolete by the time the product hits mass production three years later. Luna: It feels like most founders only think about this once the first batch arrives and the factory calls them with bad news. Lucas: Right. They treat the bill of materials as a static list during the design phase, but it should actually be a dynamic risk map. The best hardware teams I have studied, like DJI in the drone space, spend significant engineering hours just mapping the lifecycle status of every single passive and active component before they commit to a final design. Luna: Does that mean they avoid using the newest, cutting-edge chips even if they offer better performance? Lucas: Often, yes. There is a concept called the maturity curve of components. New silicon is always faster or more efficient, but it also carries the highest risk of early failures and sudden discontinuation. Established parts, which might be slightly less powerful, come with decades of reliability data and guaranteed supply commitments from the fab. For a startup, reliability beats raw specs every single time. Luna: That makes sense. But what happens when you absolutely need that new tech to compete? You can't just ignore innovation forever. Lucas: You don't ignore it, you hedge against it. This is where the strategy gets interesting. Successful companies use a dual-sourcing approach, but not just for assembly. They qualify alternative components at the schematic level. So if your primary processor goes dark, you already have a pin-compatible or nearly compatible backup validated in your lab. Luna: Is that common among early-stage startups, or is that something reserved for bigger players with deeper pockets? Lucas: It used to be reserved for giants like Apple or Tesla, who could afford the testing costs for multiple variants. But now, there are tools and services that make this accessible to smaller teams. You can simulate compatibility digitally before you ever order a sample. The barrier is really cultural. Founders have to prioritize supply chain security over squeezing out every last percent of margin on the initial prototype. Luna: I think many founders see those extra tests as a waste of time until it's too late. It's a classic case of underestimating tail risks. Lucas: Precisely. And the financial impact is massive. Let us say a startup saves five dollars per unit by picking the cheapest chip, but then that chip is discontinued. Now they have to pay fifty dollars per unit for an emergency replacement plus the cost of reworking the existing inventory. That one decision wipes out years of profit margins. Luna: It really highlights why hardware is so different from SaaS. In software, you just update the code. In hardware, you are stuck with the physical reality. Lucas: And the physical reality includes things like lead times. Even if the part is not discontinued, the lead time can stretch from weeks to sixty weeks during peak demand. This forces companies to hold huge amounts of cash in inventory, which kills their burn rate. Luna: So holding inventory becomes a survival tactic rather than just a logistical detail. That changes the whole capital structure of the business. Lucas: It does. And this brings us to a critical point about forecasting. You cannot rely on standard sales projections to determine how many chips to buy. You need to model based on component lifecycle announcements. Most manufacturers publish an end of life notice eighteen to twenty-four months in advance. If you miss that window, you are buying at a premium on the gray market. Luna: The gray market sounds dangerous. Are we talking about counterfeit parts getting mixed in there? Lucas: We are talking about a minefield of counterfeits, refurbished parts sold as new, and components pulled from older devices. It is a massive problem in the industry. The reputable distributors act as filters, but if you go off-channel to save money, you are taking on enormous liability. One bad batch of fake capacitors can destroy your brand reputation overnight. Luna: Yeah, I would never trust a random eBay seller with my power delivery system. Stick to the authorized channels, no matter the price hike. Lucas: Absolutely. But let us pause here for a moment. If these conversations about navigating the tricky waters of hardware development are useful for what you're building or running, the way these stay ad-free is listener support — buy me a coffee dot com slash fexingo. It helps us keep digging into the details that matter to operators and builders. Luna: Thanks for the support. It really allows us to cover these niche but critical topics that big media often skips over. Lucas: Back to the technical side. Another layer of protection is modular design. Instead of integrating a specific sensor directly onto the main motherboard, you put it on a daughter card or a modular interface. This means if that sensor goes obsolete, you can swap out the module without redesigning the entire core system. Luna: That reminds me of how some industrial equipment handles upgrades. They keep the base platform stable and just swap out the functional heads. Lucas: Exactly. It is a trade-off. Modularity adds a little bit of weight and complexity to the interface, but it buys you flexibility. For consumer electronics, where size and cost are king, modularity is harder to justify. But for robotics or medical devices, it is often essential. Luna: Medical devices definitely have higher stakes. A failure there isn't just a refund, it's a regulatory nightmare. Lucas: Regulatory compliance is another huge factor. When you change a component, you often have to re-submit for FDA or CE certification. This can take months. So, even if you find a perfect substitute, the administrative burden might delay your launch by half a year. That delay can be fatal for a startup competing in a fast-moving market. Luna: So the hidden cost isn't just the part itself, it's the bureaucratic drag of changing it. That is something investors rarely ask about during due diligence. Lucas: They should. Smart investors are starting to ask about the bill of materials depth. They want to know if the engineering team has thought about what happens in year three, not just month three. If the answer is no, that is a red flag for me. Luna: It shows whether the team is thinking like operators or just like dreamers. Operational maturity is everything in hardware. Lucas: Let us talk about inventory buffering for a second. Some companies adopt a strategy of buying excess stock upfront. They might purchase three years of supply for critical components at the start of the project. This locks up a lot of capital, but it eliminates the uncertainty. Luna: That sounds risky financially though. What if the product doesn't sell as expected? You are left with millions in dead stock. Lucas: It is a real risk. The trick is to identify which components are truly critical. Not everything needs a three-year buffer. Usually, it is the custom ASICs, the specialized microcontrollers, or the unique mechanical housings. Standard resistors and capacitors can usually be sourced anywhere, anytime. Focus your risk management on the bottlenecks. Luna: So you categorize your BOM. High-risk, high-impact items get the heavy protection, while the commoditized stuff is left flexible. Lucas: Spot on. It is about prioritization. You cannot protect everything equally. You have to accept some risk on low-value parts to fund the protection of high-value ones. It is a portfolio management approach applied to engineering. Luna: I love that framing. Treating the bill of materials like a financial portfolio. It makes the abstract concept of supply chain risk much more tangible for founders. Lucas: And it forces them to have difficult conversations with their engineers. Engineers love novelty. They want to use the shiny new thing. Management has to be the brake, insisting on stability and provenance over excitement. Luna: That tension between engineering desire and business reality is probably the biggest friction point in any hardware company. Lucas: It is. And resolving it requires clear metrics. If you tie engineering bonuses to supply chain stability, not just feature completion, you will see a dramatic shift in behavior. People optimize for what they are measured on. Luna: So instead of rewarding the launch of a new feature, you reward the smooth ramp of production with zero disruptions. That aligns incentives perfectly. Lucas: Exactly. It moves the goalpost from 'can we build it' to 'can we build it reliably at scale.' That is the definition of mature hardware development. Luna: It seems like the gap between a good idea and a successful product is entirely filled with these kinds of boring, unglamorous decisions. Lucas: The boring decisions are the ones that keep the lights on. Innovation captures headlines, but supply chain discipline captures market share. Luna: I think that is a great takeaway. Success in hardware is less about the spark of genius and more about the endurance of planning. Lucas: Endurance and foresight. Looking past the next quarter to the next decade of component availability. Luna: Do you think this trend toward extreme supply chain vigilance will continue, or will AI eventually solve the prediction problems? Lucas: AI will help with prediction, but it cannot manufacture atoms. Physical constraints remain. As long as we are building physical objects, the laws of physics and geopolitics will dictate availability. Technology can optimize the process, but it cannot remove the risk entirely. Luna: Right. You can predict the weather, but you still need an umbrella. The physical world is stubborn. Lucas: Very true. So the question for founders in September of twenty twenty-six is not whether you can eliminate risk, but whether you have built the architecture to survive it when it inevitably strikes. Luna: Building for the storm, not just for the sunny days ahead. Lucas: That is exactly it. Because the storm always comes.