# How Chips Are Using On-Chip Drug Delivery for Implantable Devices

An episode of [The Hardware Podcast with Fexingo: Chips, Devices, and Electronics Engineering Conversations](https://invisiblenetwork.ai/podcast/1896799982), 8 Jul 2026, 9:33.

Audio: https://audio.fexingo.com/business/the-hardware-podcast/episode-0099.mp3

In this episode, Lucas and Luna explore the cutting-edge world of on-chip drug delivery systems used in implantable medical devices. They discuss how microelectromechanical systems (MEMS) technology enables precise, programmable release of therapeutic compounds directly at the site of need, reducing side effects and improving patient outcomes. The hosts examine a specific case study from MIT's Microsystems Technology Laboratories, where researchers developed a chip with 100 reservoirs that can be individually triggered to release drugs. They also touch on the power challenges, biocompatibility issues, and the potential for closed-loop systems that respond to real-time biomarkers. Lucas and Luna debate the timeline for commercialization and the impact on chronic disease management, from diabetes to cancer therapy. A must-listen for anyone interested in the intersection of semiconductor technology and healthcare. #OnChipDrugDelivery #ImplantableDevices #MEMS #MedicalChips #BiomedicalEngineering #DrugDelivery #MIT #Microsystems #PharmaceuticalTechnology #ClosedLoop #ChronicDisease #Diabetes #CancerTherapy #Technology #FexingoBusiness #BusinessPodcast #HardwarePodcast #ChipsAndDevices Keep every episode free: buymeacoffee.com/fexingo

## Transcript

Lucas: So there's this trend in chip design that I think a lot of our listeners might not have on their radar yet, and it's one of those intersections that feels almost sci-fi until you dig into the numbers. I'm talking about on-chip drug delivery — literally using semiconductor manufacturing techniques to build implantable devices that release therapeutic compounds with programmable precision.
Luna: Wait, so we're talking about a chip that stores drugs and then releases them on command? Like a pharmacy on a silicon wafer?
Lucas: Exactly that. And it's not just a concept — there are working prototypes. One of the most cited examples comes from MIT's Microsystems Technology Laboratories, where they built a chip with a hundred tiny reservoirs, each about the size of a pinprick, that can be individually opened by applying a small voltage. The reservoirs are sealed with a thin membrane of gold and titanium that dissolves when a current is applied.
Luna: How much drug can one of those reservoirs hold? Because a hundred pinprick-size volumes doesn't sound like a lot.
Lucas: It's on the order of nanoliters per reservoir — so total volume is maybe a few microliters. But for many potent drugs, especially peptides or hormones, that's enough for weeks or even months of therapy. The key is that you can release different drugs at different times, or stagger doses, all from a single implant. And because it's a chip, you can integrate control logic and telemetry, so a doctor could adjust the release schedule wirelessly.
Luna: So we're moving away from mechanical pumps and towards solid-state delivery. What are the advantages beyond miniaturization?
Lucas: Precision and reliability. Mechanical pumps have moving parts — valves, motors, syringes — that can fail or clog. A mems based chip has no moving parts; it's just a set of sealed cavities and electrodes. That means it can be more reliable over years inside the body. Also, because it's fabricated using standard semiconductor processes, you can integrate sensors on the same chip — glucose sensors, pH sensors, even temperature — and create a closed-loop system that releases insulin when glucose levels rise, for example.
Luna: So the chip becomes both the sensor and the actuator. That's a pretty big leap from where we are with current implantable pumps.
Lucas: Right. And there's a company called MicroChips Biotech — spun out of MIT — that actually got FDA approval for a microchip-based implant for osteoporosis. Their device releases teriparatide, a parathyroid hormone, once a day from an implant placed under the skin. The Phase 2 data showed it was as effective as daily injections, but with much better patient compliance because you don't have to remember to inject yourself.
Luna: I remember reading about that. The implant lasted about a year, right? And then you have to replace it surgically.
Lucas: Yeah, that's the current trade-off. But the chip itself is tiny — about the size of a small button — and the replacement procedure is a simple incision under local anaesthetic. Compare that to daily injections for a year, and most patients prefer the implant. The compliance rate in the trial was over 90 percent, versus something like 40 to 50 percent for injections in real-world settings.
Luna: That's a massive improvement. But I have to wonder about power — how do you power something like that for a year without a battery that's bigger than the chip itself?
Lucas: Great question. The early versions used a small battery, but that's not ideal because batteries have limited lifespan and can pose safety risks if they leak. More recent designs are exploring energy harvesting from the body — think piezoelectric materials that generate electricity from movement, or thermoelectric generators that use body heat. Some are even working on wireless power transfer using inductive coupling, similar to how you charge a smartwatch, but through the skin.
Luna: And the drug itself — how do you keep it stable inside the chip for months at body temperature?
Lucas: That's actually one of the cleverest parts. The reservoirs are hermetically sealed, so no moisture or oxygen gets in. Many drugs are formulated as a dry powder or a concentrated solution that's stable at room temperature. The chip is implanted under the skin where temperature is around 37 degrees Celsius, but because it's sealed, degradation is minimal. Some designs even include a small heating element to sterilise the drug before release.
Luna: So what's the holdup? Why aren't these chips everywhere yet?
Lucas: A few reasons. First, the manufacturing yield for these chips is still relatively low — you're combining semiconductor processes with biocompatible materials and precise drug loading, which is a complex supply chain. Second, regulatory approval is slow because these are combination products — part device, part drug — so they fall under both the FDA's device center and its drug center. And third, the economics: developing a new implant for each drug is expensive, so companies are focusing on high-value drugs where patient compliance is a big problem.
Luna: So chronic diseases like diabetes, osteoporosis, maybe even cancer therapy where you need frequent doses of a biologic.
Lucas: Exactly. And there's also interest in using these chips for contraception — releasing a hormone like levonorgestrel over several years, but with the ability to turn it on and off wirelessly. Imagine a woman could get a small implant that provides contraception for three years, but if she decides she wants to become pregnant, she just visits her doctor and they deactivate it remotely, with no need for a procedure.
Luna: That would be a game-changer. But I'm also curious about the closed-loop aspect — the idea that the chip could sense a biomarker and release drug automatically. How close are we to that vision?
Lucas: There are academic prototypes that have demonstrated glucose sensing and insulin release in animal models. The challenge is making the sensor stable for long periods — glucose sensors tend to drift over time because of biofouling, where proteins and cells adhere to the sensor surface. Researchers are working on coatings and self-cleaning mechanisms, but it's tough. I'd say we're probably five to ten years away from a fully closed-loop implant that's approved for humans.
Luna: So in the meantime, we have open-loop implants that release drug on a pre-programmed schedule, with the ability to adjust wirelessly. That's still a big step forward from daily injections.
Lucas: Absolutely. And the market is starting to recognise that. A few years ago, the global market for implantable drug delivery devices was around five billion dollars. Analysts project it could grow to over fifteen billion by the early 2030s, driven by these microchip-based systems. We're seeing partnerships between pharma companies and MEMS foundries — it's becoming a real industry.
Luna: You know, this conversation reminds me that we're able to have this kind of deep dive into niche technology topics because of listener support. We keep the show ad-free and focused on the substance, and that's only possible if people who find value in it chip in.
Lucas: That's a good point. If today's episode gave you something to think about — the future of implantable chips, or even just a clearer picture of where healthcare hardware is heading — and if it was worth the price of a coffee to you, there's a link for that. It's buy me a coffee dot com slash fexingo. No pressure, just an option to help us keep doing these deep dives.
Luna: And we really do read every message that comes through. So back to those glucose sensors — what's the most promising approach to solving the biofouling problem?
Lucas: One approach that I find really elegant is using a microfluidic channel that continually flushes a small amount of fluid across the sensor surface to prevent protein buildup. It consumes a tiny amount of power and a bit of the drug reservoir, but it could keep the sensor accurate for months. Another idea is to use a hydrogel coating that swells and releases an anti-fouling agent over time.
Luna: So there's still a lot of engineering to do. But the basic platform — a silicon chip with reservoirs, electrodes, and control circuitry — is proven. The rest is optimization.
Lucas: Right. And I think the most exciting thing is that this platform is fundamentally programmable. Once you have a reliable chip that can release any drug on command, you can imagine combining multiple drugs in one implant — say, a cancer therapy that releases a checkpoint inhibitor and a chemotherapy agent in a specific sequence, based on real-time tumor markers. That's the kind of personalised medicine that's only possible with hardware like this.
Luna: And it's all built on the same basic technology that gave us microprocessors and MEMS accelerometers. Just shows how versatile chip manufacturing can be.
Lucas: Exactly. And that's why we keep covering it — because the applications are only limited by our imagination. For next time, I want to look at how chips are being used in brain-computer interfaces, specifically for motor recovery after stroke. There's some new work using on-chip neural recording that I think will surprise people.
Luna: Sounds great. I'll start reading up on that. Until then, thanks for listening.
Lucas: Yeah, take care.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.