GEO vs LEO vs MEO for IoT: Orbit Altitude Tradeoffs for Satellite Connectivity

Comparing GEO, LEO, and MEO satellite orbits for IoT connectivity tradeoffs

A coin cell battery has roughly 1,000 joules of energy. Reaching a geostationary satellite from a handheld device burns through that budget about 30 times faster than reaching a LEO bird. That single fact, more than any spec sheet or marketing claim, decides which orbit your IoT device should talk to.

If you’re an architect picking satellite connectivity for cattle collars, container trackers, or soil probes, orbit altitude is upstream of every other decision you’ll make: antenna size, transmit power, module choice, latency, cost per device. Get it wrong and you’ll be redesigning the BOM 18 months in.

The thesis here is plain. For low-power massive IoT in 2025, LEO is the right default. GEO and MEO still have homes, but they’re narrower than the hype suggests. Here’s the physics behind that call, and where the exceptions live.

The Three Orbit Regimes at a Glance 

                    GEO          MEO          LEO
Altitude            ~35,786 km   2k-20k km    300-1,500 km
One-way latency     ~240 ms      ~50-130 ms   ~5-25 ms
Path loss (S-band)  ~190 dB      ~175 dB      ~160 dB
Satellites needed   3            ~12-24       100s-1000s
Doppler shift       Negligible   Moderate     High (±48 kHz)
Handover frequency  Never        Minutes      ~5-10 min
View from device    Fixed        Slow-moving  Fast-moving

Altitude drives nearly every row in that table. Higher orbit means farther signal travel (more path loss, more latency) but a wider footprint per satellite (fewer birds needed for coverage). Lower orbit flips it: better link budget, worse handover and constellation complexity. Doppler, revisit time, and ground architecture all fall out of those two physics constraints.

What Low-Power IoT Actually Needs

The broadband satellite story (Starlink, OneWeb’s enterprise tier) is about throughput per user. Low-power IoT is a different problem with different priorities:

  • Power budget: years on a coin cell or a small primary battery, often with no solar option
  • Tiny payloads: tens to hundreds of bytes, sent minutes or hours apart
  • Mobile or remote deployments: livestock collars, shipping containers, pipeline sensors, soil probes
  • BOM pressure: the radio module needs to land under $10 in volume to make the unit economics work

When you optimize against that list, link budget dominates. Latency rarely matters (a soil moisture report doesn’t care about 200 ms). Throughput barely matters. What matters is how many joules it costs to deliver one packet, and how cheap the antenna and PA can be while still closing the link.

Why LEO Wins for Low-Power Massive IoT

Link budget is the decisive factor, and LEO has a ~30 dB advantage over GEO at S-band.

30 dB is a factor of 1,000 in power. A device reaching a LEO satellite with 100 mW of TX power and a small chip antenna would need roughly 100 W EIRP to close the same link to GEO, and a high-gain directional antenna to go with it. Battery-powered mobile IoT can do the first. It can’t do the second.

Latency is a bonus, not the driver. For an asset tracker pinging once an hour, 240 ms vs 20 ms is invisible. Where LEO’s low latency actually pays off is in protocol design: ack-based transports become viable, and over-the-air firmware updates stop being an overnight ordeal.

3GPP NTN (Releases 17 and 18) was built around LEO realities. Doppler pre-compensation is standardized. Extended timing advance handles long propagation delays. Conditional handover takes care of fast-moving satellites. NTN-capable silicon from Sony (Altair), Qualcomm (212S), and MediaTek (MT6825) handles all of it transparently. You’re not writing custom doppler compensation in firmware anymore.

The module ecosystem has consolidated around LEO. When NB-IoT NTN and LTE-M NTN modules sample, they target LEO constellations (Skylo via Viasat/Ligado, Iridium’s project Stardust, others) first. GEO L-band IoT modules exist, but the volume and price curve is on the LEO side.

Anchor use cases:

  • Asset tracking: global container telemetry where terrestrial roaming is broken, expensive, or both. A LEO link closes the gap when the ship is in mid-Pacific or the truck crosses a border into a partner network that charges $40/MB.
  • Precision agriculture: soil moisture probes, tank level sensors, gate actuators on farms that sit 30 km past the last cell tower. The economics only work if the device runs for 5+ years on its installed battery.

A more aggressive proof point: Hubble Network’s approach of receiving standard BLE advertising packets directly at LEO satellites only works because the altitude makes the link budget math close at all. Try the same trick to GEO and the numbers fall apart by orders of magnitude. The further detail on how BLE chirps survive that path is in the Hubble satellite network overview, but the takeaway here is general: aggressive low-power architectures need LEO to exist.

Where GEO Still Makes Sense

LEO isn’t universally better. GEO holds the line on a few real use cases:

  • Fixed, powered terminals: maritime VSAT, oil & gas SCADA at wellheads, broadcast video. When you have mains power and a parabolic dish, path loss stops mattering.
  • Broadcast and multicast: pushing a firmware image or weather alert to thousands of devices across a continent is cheaper from one GEO satellite than 50 LEO ones.
  • Mature L-band services: Inmarsat and Thuraya have decades of SLAs, certified module inventory, and regulatory clearance in places newer LEO operators don’t yet reach.
  • Simpler ground architecture: one satellite, one gateway, one orbital slot to license. For some deployments the operational simplicity is worth the link-budget penalty.

If your device plugs into a wall or sits on a ship’s bridge, GEO is still a serious option.

Where MEO Fits (and Why It’s Narrow for IoT)

MEO is the middle child. O3b mPOWER and the emerging MEO designs are built for enterprise broadband backhaul and government links, not battery-powered sensors.

The pitch is reasonable latency (50-130 ms), fewer handovers than LEO, and a smaller constellation than LEO needs. But the path loss is still roughly 15 dB worse than LEO at the same frequency, and there’s no IoT-specific MEO module ecosystem to speak of. For low-power massive IoT, MEO is rarely the optimal pick. If a MEO operator launches an IoT-specific service in the next few years with tailored modules, revisit. Until then, treat it as a backhaul orbit.

The Tradeoffs You Inherit With LEO

LEO isn’t free. Be honest about what comes with the territory:

  • Doppler shift: up to ±48 kHz at S-band as a satellite screams overhead at 7.5 km/s. NTN-capable silicon handles pre and post-compensation, but it constrains your choice of module.
  • Frequent handovers: a satellite is overhead for 5-10 minutes before another takes the link. Session continuity needs protocol support, or a store-and-forward design that sidesteps continuous sessions entirely (often the right answer for low-data-rate IoT).
  • Coverage gaps in early constellations: depending on the operator, revisit times can range from continuous to once-per-orbit. Continuous global coverage is fine for daily reports, painful for real-time alerts.
  • Constellation risk: you’re betting on the operator’s funding, spectrum filings, and longevity. Multi-orbit-capable modules are the obvious hedge where they exist.

These are boring engineering problems now rather than research questions. That’s the difference vs. five years ago.

A Short Decision Framework 

Battery-powered + mobile?
      |
   YES → LEO
      |
    NO
      |
  Fixed + broadcast?
      |
   YES → GEO
      |
    NO → Evaluate MEO or hybrid

Ask the questions in this order:

  1. Is the device battery-powered and mobile? If yes, LEO. Stop here.
  2. Is it mains-powered, fixed, with broadcast or multicast needs? GEO.
  3. Do you need sub-50 ms latency at low power? There’s no great answer; LEO is the closest, with caveats.
  4. What modules can you actually buy in volume in the next 12 months? Mostly LEO-targeted NTN silicon.
  5. Which operator can match your target markets on ground segment, spectrum, and regulatory footprint? That’s often the real deciding factor, not orbit per se.

Picking an orbit is increasingly picking a constellation. Two LEO operators with different spectrum, gateway coverage, and module partners will give you very different products even though they share an orbit.

Picking the Constellation, Not Just the Orbit

For low-power massive IoT, LEO is where the physics and the silicon both point. The link budget advantage is decisive, the 3GPP standards are caught up, and the module roadmaps from the big NTN vendors all converge there. GEO keeps its place in fixed broadband and broadcast. MEO is a backhaul orbit looking for an IoT story.

Pick the orbit and the constellation together. Evaluate the operator’s ground architecture, spectrum, and module ecosystem with the same rigor you give the RF link budget, because in practice that’s where deployments succeed or stall.

If you’re specifically exploring how BLE-class devices can talk directly to LEO satellites without a cellular NTN module in the BOM, Hubble’s developer docs walk through the link budget and integration path in more detail.

Hubble Network connects existing Bluetooth devices directly to LEO satellites, no cellular NTN module or custom silicon required. See how it works →