Latest / Hardware Startups with Fexingo: Devices, Robotics, and Manufacturing Tech Companies / How Hardware Startups Cut Costs with Panel-Level Packaging
Transcript
- Lucas: So I keep seeing this term pop up in my feed — panel-level packaging. And for a hardware startup, it sounds like a magic bullet for cutting chip packaging costs. But is it actually something a startup can use? Luna: Yeah, I've seen the hype too. Big players like Samsung and TSMC are investing in it. But for a startup shipping tens of thousands of units, does it even make sense? Lucas: Right, so let's break down what panel-level packaging actually is. Traditionally, chips are packaged on round silicon wafers — 300 millimeters in diameter. Panel-level packaging moves to a square or rectangular panel, like the ones used in PCB manufacturing. These panels are much larger — think 600 by 600 millimeters or even bigger. Luna: So more chips per panel means lower cost per chip. That's the basic economics, right? Lucas: Exactly. The panel has more usable area, so you can fit more packages per substrate. That amortizes the processing costs over a larger number of units. Some estimates say you can cut packaging costs by twenty to thirty percent compared to wafer-level processes. That's a significant chunk for a startup where every dollar of BOM matters. Luna: But what's the catch? There's always a catch. Lucas: The catch is that panel-level packaging is still largely a high-volume play. The biggest challenge is warpage — the panel bends and flexes during processing because it's not as rigid as a silicon wafer. That creates alignment issues and reduces yield. The equipment is also not as mature or as widely available as wafer-level tools. Luna: So for a startup that's maybe producing fifty thousand units a year, would they even get access to panel-level packaging lines? Lucas: Probably not directly. Most of the panel-level capacity is being built by the big OSATs — outsourced semiconductor assembly and test companies — and they're focused on high-volume applications like smartphones, RF front-ends, and power management ICs. They're not exactly knocking on the door of a startup that needs ten thousand units a month. Luna: So is it just a pipe dream for us, then? Lucas: Not necessarily a pipe dream, but you have to be strategic. Some startups are using panel-level packaging for specific high-volume products — think wearables or IoT modules where the volume justifies the investment. But for most early-stage hardware, the sweet spot is still wafer-level packaging or even traditional lead-frame packages. Luna: What about the design side? Does the chip have to be designed differently for panel-level? Lucas: That's a good question. Panel-level packaging often uses fan-out technology, which means you can design the package with more I/O pins in a smaller footprint. But that requires the die to be designed with that in mind — you need to have the right bump pattern and layout. If you're designing your chip from scratch, you could consider that, but if you're using an off-the-shelf die, you're limited to what the packaging house supports. Luna: So it really comes down to how much control you have over the chip design. Lucas: Exactly. And there's also the reliability question. Panel-level packages are relatively new, so there's less long-term field data. For a startup that's building a reputation for reliability, that's a risk. You might save money upfront, but if you get a higher failure rate in the field, that could cost you more in the long run. Luna: Have any startups actually made it work? I'm thinking of a company like Onyx, that makes AR glasses. They're probably shipping a few hundred thousand units a year. Did they use panel-level? Lucas: I don't know the specifics of Onyx, but there are examples in the wearables space. Some smartwatch modules are using panel-level fan-out because they need a compact package with lots of connections. But those are usually designed by the big chip companies, not the startup itself. The startup might buy a system-on-chip that already uses advanced packaging. Luna: So the startup benefits indirectly, without having to invest in the packaging technology themselves. Lucas: Right. And that's a key point — you don't have to own the packaging line. You just need to choose the right chip supplier who's using advanced packaging to make the chip smaller and cheaper. That's the indirect benefit. Luna: But what if a startup is really serious about cutting costs? Could they approach an OSAT with a volume forecast and see if they can get a slot on a panel line? Lucas: It's possible, but you'd need to bring some serious volume — maybe a million units a year — to get their attention. Otherwise, they'll steer you to their wafer-level or lead-frame options. The reality is that panel-level packaging is still a scale game. Luna: So what's the practical takeaway for a founder listening today? Lucas: The takeaway is to understand your packaging costs as a line item. Don't just assume panel-level is automatically cheaper. Do the math on your actual volume and yield. And talk to your OSAT early — they can tell you what they can offer at your volumes. Sometimes the cheapest option is actually a more mature process that has higher yields and lower risk. Luna: And maybe wait until panel-level becomes more mainstream, so the equipment is more flexible and the yields are higher. Lucas: Exactly. The technology is promising, but it's not yet a slam-dunk for startups. If you're shipping a million units a year, it's worth a serious look. But for most early-stage hardware, you're better off focusing on design for manufacturing and working with your OSAT to optimize what you already have. Luna: Which kind of ties into something we should mention — if you found this useful, these conversations are ad-free because of listener support. If today's episode helped you think about your packaging strategy, buying a coffee at buy me a coffee dot com slash fexingo is a simple way to keep that going. Lucas: Yeah, we really appreciate that. It helps us keep doing deep dives like this without any sponsors pushing their own agenda. So if you can, that's a big help. Luna: And we're back to the tech. I'm curious — what about the environmental angle? Does panel-level packaging reduce waste compared to wafer-level? Lucas: That's a great point. The rectangular panel uses more of the substrate area, so there's less wasted material compared to a round wafer that has edges cut off. Some studies suggest it could reduce material waste by as much as 30 percent. That's a sustainability win, and for startups that care about their carbon footprint, that could be another reason to consider it — as long as the volume makes sense. Luna: So it's not just about cost — it's about being more efficient with resources. Lucas: Exactly. And as the technology matures, those efficiency gains might trickle down to smaller players. But for now, the biggest opportunity for a startup is to stay informed, keep an eye on the OSAT capabilities, and make sure you're not overpaying for packaging that doesn't need to be premium. Luna: I think that's a solid takeaway. Thanks for digging into that. Lucas: Absolutely. And if you're curious about whether panel-level packaging might fit your product, the best first step is a conversation with your packaging partner. They'll give you the real numbers.