AI moving medicine to your wrist - top 3 movers and shakers

Host 1: Okay, so today let's do a deep dive into this uh massive structural shift happening in medicine, people are calling it Edge Health.

Host 2: Mhm. And what we're really talking about is moving the AI, the intelligence, from these big centralized cloud servers directly onto the wearable you have on your wrist.

Host 1: Right. And our mission today is to really unpack that. We'll look at the tech that makes it possible, the clinical breakthroughs we're expecting, and the, you know, the three big companies driving it all: Abbott, Apple, and Dexcom.

Host 2: And it's driven by necessity really. The old cloud-first model is just it's failing, especially for acute events.

Host 1: What do you mean by that?

Host 2: Well, look, if you're trying to catch something life-threatening like a heart arrhythmia or a sudden glucose crash, the latency is a killer.

Host 1: The time delay.

Host 2: Exactly. The time it takes to send raw data from your watch to a server, have it analyzed, and get an alert back. It's just too long in a clinical emergency.

Host 1: And I assume the energy cost is just astronomical. You can't be transmitting terabytes of raw health data all day.

Host 2: You kill the battery in hours. So the intelligence has to move to the edge, to the device itself.

Host 1: So how does that work? How do you fit that kind of computing power on a tiny chip?

Host 2: That's where this revolution in uh tiny machine learning comes in, or TinyML.

Host 1: Okay.

Host 2: Think of it as these incredibly complex diagnostic algorithms, these neural networks, but they're compressed, sometimes down to using 4-bit integers. They run on tiny microcontrollers with just kilobytes of RAM.

Host 1: So they're basically hyper-efficient, specialized little brains.

Host 2: That's a great way to put it. And the core logic becomes compute over transmit. The device is always analyzing, always thinking locally.

Host 1: But it's quiet.

Host 2: It's quiet. It only wakes up the high-power radio to send an alert when the AI detects an actual problem. That's how you get battery life that lasts for weeks, not days.

Host 1: And this must have huge implications for privacy, right? If my raw biometric data isn't leaving my wrist.

Host 2: It fundamentally changes the game. Your minute-by-minute heart rate variability, your glucose fluctuations, that sensitive stuff never leaves the secure enclave on your device.

Host 1: Only the final insight gets sent.

Host 2: Precisely. The message just says atrial fibrillation detected, not the raw data that led to the conclusion.

Host 1: And you can't do that with off-the-shelf chips. This requires custom silicon.

Host 2: It does. You need a dedicated Neural Processing Unit, an NPU, built right into the chip. It's a circuit designed to do one thing very, very well: run AI models constantly with very little power.

Host 1: Which is where Apple's big advantage comes in.

Host 2: It's their key advantage. Okay, so let's connect this tech to the clinic. Metabolic health is really at the forefront.

Host 1: Beyond just glucose monitoring.

Host 2: Way beyond. The immediate goal, sort of the Holy Grail right now, is a dual sensor for both glucose and ketones.

Host 1: Which for a type 1 diabetic is a huge safety net against DKA, diabetic ketoacidosis.

Host 2: A massive one. And at the same time it opens up this enormous wellness market for people on keto diets.

Host 1: Right. And we're seeing similar moves in cardiovascular health with cuffless blood pressure.

Host 2: We are, using the PPG sensor, the little light on the back of your watch. But there's a really critical regulatory nuance here.

Host 1: It's not actually a measurement, is it?

Host 2: Not yet. The current FDA clearances are mostly for notification. The watch can tell you if you have a persistent trend of high blood pressure, but it won't give you a specific systolic, diastolic number like a cuff does.

Host 1: Okay, that's a key distinction. Now let's talk strategy, the three big architects. Let's start with Abbott.

Host 2: Abbott is the biosensor sovereign. Their superpower is just massive high-volume manufacturing scale. They can produce these sensors incredibly cheaply.

Host 1: And they're pivoting hard into the wellness space with Lingo.

Host 2: Exactly, using their medical grade hardware for metabolic coaching for everyone, not just people with diabetes.

Host 1: Okay, then there's Apple. We mentioned their silicon advantage.

Host 2: It's the silicon and the ecosystem. But their secret weapon is regulatory. They use something called a Pre-determined Change Control Plan, or PCCP.

Host 1: What does that do?

Host 2: It lets them update their medical software features through software updates without needing a new, lengthy FDA review every single time. It's a massive competitive moat. They can iterate 10 times faster.

Host 1: Wow. And finally, Dexcom.

Host 2: Dexcom is the high-performance specialist. They're all about precision and uh untethering the patient.

Host 1: With their direct-to-watch connectivity.

Host 2: Right. Their new sensor can talk directly to an Apple Watch. It completely bypasses the phone. For someone on an automated insulin pump, that data continuity is everything.

Host 1: So what we're seeing is this total convergence. By 2026, the line between a consumer smartwatch and a medical monitor is it's basically gone.

Host 2: It will have completely disappeared. The tech is there and the regulatory pathways are now in place.

Host 1: So, to wrap up, what's the final, provocative thought here? Where does Edge Health go next? It can't just be about sensing things.

Host 2: That's the perfect question. The next frontier, and we'll see this by 2028, is moving from sensing to acting.

Host 1: You mean agentic AI.

Host 2: Exactly. Imagine this: your device doesn't just detect a major cardiac risk, it autonomously schedules a telehealth consult with a cardiologist, finds a slot in your calendar, and then maybe even adjusts your schedule to reduce stress triggers it has identified.

Host 1: So the future of your healthcare is literally a piece of code running on your wrist.

Host 2: That's exactly what it is. It's not in the hospital, it's not in the cloud, it's right there with you.