When a product manager asks “which chemistry should we use?”, the honest first answer is almost always a counter-question: how small, how long, and how safe does it have to be? Those three constraints collapse the choice faster than any spec sheet.
The three candidates, in one table
| Property | LiPo (LCO / NMC) | Li-ion cylindrical | LiFePO4 |
|---|---|---|---|
| Energy density (Wh/L) | 600–760 | 550–730 | 320–420 |
| Cycle life to 80% SOH | 500–800 | 500–1,000 | 2,000–3,000 |
| Nominal voltage | 3.7–3.85 V | 3.6–3.7 V | 3.2 V |
| Thermal runaway onset | ~150 °C | ~150 °C | ~270 °C |
| Minimum thickness achievable | 0.45 mm | ~4 mm (dia) | ~4 mm (dia) |
For most wearables, LiPo wins
Watches, TWS earbuds, fitness bands, smart rings, AR glasses — almost every device in this category uses a lithium-polymer pouch cell. The reason is mechanical: only a pouch can be made curved, stepped, rectangular or under 1 mm thick, which is what an industrial-design team needs when the cell has to disappear into the enclosure.
Within LiPo, the default cathode is LCO for sub-1 Ah cells. Above 1 Ah you see NMC more often, which trades a little volumetric density for better cycle life and easier fast charging.
When cylindrical Li-ion makes sense
- The enclosure has a natural cylindrical cavity (barrel-shaped handheld, some medical pumps).
- You need cells in parallel beyond ~2 Ah and want a mature supply chain.
- Mechanical robustness against drop/crush matters more than a few grams of weight.
When LiFePO4 is worth considering
LFP has one job: cycle life and thermal stability. For a wearable, its lower density usually disqualifies it — you give up 35–50% of runtime at the same volume. The cases where it still wins:
- Long-life industrial wearables (thermometers, logistics wristbands) that charge daily for 5+ years.
- Medical-adjacent devices where thermal runaway risk must be pushed as far as possible.
- Charging docks or hub batteries where volume is not critical.
High-voltage LCO: the new default for AR
The recent move to 4.45 V and 4.48 V charge cut-off extends the energy envelope by about 8% at the same cell volume. For AR glasses where the battery cavity is fixed, those 8 percentage points can be the difference between 4-hour and 4.5-hour runtime. The trade-off is cycle life — typical 4.48 V formulations retain 80% SOH for around 400–500 cycles instead of 600–800, which is acceptable if the device is charged nightly and replaced after 2–3 years.
A 3-question decision shortcut
- Can the cell fit inside a pouch under 2 mm thick? → LiPo.
- Does the device charge more than 1,000 cycles in its expected life? → Consider LFP for a packable product, or accept LiPo with over-sizing.
- Are you ready to trade ~30% cycle life for ~8% more runtime? → HV LCO.
The thing we see most often go wrong
Engineering teams pick a chemistry based on a single metric — usually energy density — and discover six months later that cycle life at their actual operating temperature is half the datasheet number. Always ask a vendor for cycle data at your C-rate, depth of discharge and ambient temperature, not the lab conditions.