Direct-to-Device Satellite IoT in 2026: Every Option Compared
Most IoT devices don’t need to listen. They need to talk, transmit a GPS fix, a temperature reading, a battery voltage, and go back to sleep. Yet nearly every satellite IoT comparison lumps one-way telemetry devices in with bidirectional command-and-control systems, as if a $3 pallet tracker and a remote industrial valve actuator have the same connectivity requirements. That confusion costs you: in power budget, in module cost, in certification time, and in recurring airtime fees that compound over millions of units.
2026 is the year this market finally gets real. Multiple LEO constellations have reached operational coverage. 3GPP NTN standards have moved from paper to silicon. First-wave direct-to-device modules are shipping at scale. But the vendor marketing is thicker than ever, and the architectural differences between these solutions are not cosmetic. They fundamentally change your BOM, your battery life, and your business case.
This comparison covers direct-to-device (D2D) satellite IoT only: the device itself communicates with the satellite using an integrated or small patch antenna. No gateway boxes. No VSAT dishes. No LoRa concentrators with satellite backhaul. If your device can’t reach orbit on its own RF, it’s not D2D.
One-Way vs. Two-Way: The Decision That Shapes Everything
This is the fork in the road, and getting it wrong is expensive.
One-way uplink means the device transmits data to a satellite and never receives anything back. No acknowledgments, no downlink commands, no OTA configuration. The protocol is radically simple: wake, transmit, sleep. That simplicity translates directly into lower power, smaller modules, and cheaper silicon, because you don’t need a receive chain optimized for satellite signal levels.
Two-way means bidirectional messaging. The device can receive downlink commands, acknowledgments, firmware updates, or configuration changes from the network. This requires the device to maintain receive windows, which costs power even when idle. The RF front-end is more complex. The module is bigger. The subscription is more expensive.
Here’s the decision framework:
Does your device need to RECEIVE commands or config updates?
│
├── YES → Two-Way Required
│ (Skylo, Kineis bidirectional, 3GPP NTN modules)
│
└── NO
│
├── Is sub-10-min latency critical?
│ │
│ ├── YES → Two-Way or LEO with guaranteed pass windows
│ │
│ └── NO → One-Way Uplink (Hubble Network = best fit)
│
└── (Continue to comparison table)If your device simply reports data on a schedule, and the majority of massive-IoT deployments do exactly that, one-way uplink is almost always superior on power, cost, and time-to-market.
The Comparison Table: Every D2D Option at a Glance
Read this table as a first-pass filter. The deep dives below unpack what the columns can’t capture.
┌──────────────────┬────────────┬───────────┬───────────┬────────────┬────────────┬────────────────┬──────────────┐
│ Vendor/Standard │ Direction │ Frequency │ Module │ Tx Power / │ Typical │ Cellular │ Cert │
│ │ │ Band │ Size │ Sleep μA │ Latency │ Hybrid SIM │ Status │
├──────────────────┼────────────┼───────────┼───────────┼────────────┼────────────┼────────────────┼──────────────┤
│ Hubble Network │ 1-way up │ BLE / │ <1 cm² │ BLE-class │ 15 min – │ No (pair with │ FCC granted; │
│ │ │ Sub-GHz │ (chipset- │ Tx; ~1–3 │ few hours │ cellular │ module-level │
│ │ │ │ level) │ μA sleep │ (pass- │ separately) │ in progress │
│ │ │ │ │ │ dependent) │ │ │
├──────────────────┼────────────┼───────────┼───────────┼────────────┼────────────┼────────────────┼──────────────┤
│ Skylo │ 2-way │ NB-IoT / │ Modem- │ 23 dBm Tx; │ Near-real- │ Yes — unified │ 3GPP NTN │
│ │ │ LTE-M NTN │ class │ ~10+ μA │ time (when │ SIM for cell │ aligned; │
│ │ │ bands │ (~15×15 │ sleep │ sat │ + sat │ carrier- │
│ │ │ │ mm+) │ │ visible) │ │ dependent │
├──────────────────┼────────────┼───────────┼───────────┼────────────┼────────────┼────────────────┼──────────────┤
│ Kineis │ 1-way up + │ UHF │ ~25×25 mm │ ~500 mW │ ~15 min – │ No │ CE achieved; │
│ │ optional │ (401 MHz) │ (module) │ Tx; ~5 μA │ few hours │ │ FCC in │
│ │ 2-way │ │ │ sleep │ │ │ progress │
├──────────────────┼────────────┼───────────┼───────────┼────────────┼────────────┼────────────────┼──────────────┤
│ Swarm (SpaceX) │ 2-way │ VHF │ ~43×30 mm │ ~1 W Tx; │ Minutes – │ No │ FCC approved │
│ │ │ (150 MHz) │ (Tile) │ moderate │ hours │ │ (legacy │
│ │ │ │ │ sleep │ │ │ hardware) │
├──────────────────┼────────────┼───────────┼───────────┼────────────┼────────────┼────────────────┼──────────────┤
│ 3GPP NTN │ 2-way │ NB-IoT / │ Varies │ Standard │ Seconds – │ Yes — same SIM │ Standards- │
│ (Qualcomm, │ │ LTE-M │ (chipset │ cellular │ minutes │ as terrestrial │ based; │
│ MediaTek, Sony) │ │ bands │ vendor- │ power │ (when sat │ MNO-managed │ NTN cert │
│ │ │ │ dependent)│ profiles │ visible) │ │ maturing │
└──────────────────┴────────────┴───────────┴───────────┴────────────┴────────────┴────────────────┴──────────────┘Note: Some specs are estimates based on publicly available data as of early 2025. Where vendors haven’t published confirmed datasheets, figures are derived from reference designs and announced parameters.
Hubble Network: The One-Way Telemetry Leader
Hubble’s architecture is deceptively simple, and that’s exactly the point. Standard BLE chipsets, the same silicon already in millions of devices, transmit to LEO satellites overhead. No new RF chain. No satellite-specific modem. If your product already has a Nordic nRF52 or nRF54 series (or comparable BLE SoC), you’re most of the way there.
The power implications are dramatic. BLE transmit power is measured in single-digit milliwatts. Sleep current sits at 1–3 μA on commodity BLE chipsets. For a device that wakes every few hours to send a 20-byte payload, you’re looking at multi-year battery life on a coin cell. No other satellite IoT architecture comes close on energy per transmitted bit for small payloads.
The integration story matters just as much. Because Hubble uses existing BLE hardware, you inherit the BLE module’s existing certifications (FCC, CE, IC). The incremental certification burden for satellite operation is dramatically lower than designing in a dedicated satellite modem with its own RF chain, power amplifier, and antenna matching network.
Ideal use cases: mass asset tracking (containers, pallets, livestock tags), environmental sensors (soil moisture, water level, weather stations), utility meter reads. Any device following a “report-and-sleep” pattern where data freshness of 15 minutes to a few hours is acceptable.
The honest limitation: no downlink. You cannot send OTA firmware updates, configuration changes, or acknowledgments over satellite. If your deployment model requires remote reconfigurability via satellite, Hubble isn’t the right tool. But for the vast majority of massive-IoT deployments, where devices are configured once at provisioning and then report data for years, this constraint is a non-issue, and the power and cost savings are substantial.
Skylo: Two-Way with Cellular Convergence
Skylo’s bet is on convergence: a single subscription and SIM that works on terrestrial NB-IoT where available and falls back to satellite NTN when it’s not. This is compelling for devices that physically move between coverage zones. Think fleet trackers on trucks that traverse urban corridors and remote highways.
The architecture aligns with 3GPP NTN standards, meaning Skylo integrates with carrier ecosystems rather than competing with them. For enterprises already managing cellular IoT subscriptions through Verizon, Vodafone, or similar, Skylo slots into existing operational workflows.
The trade-off is power. Two-way NTN modems draw significantly more during transmit (23 dBm is typical) and must maintain periodic receive windows for downlink paging, pushing sleep currents higher than BLE-only designs. Module footprint is modem-class: expect 15×15 mm minimum for the connectivity module alone, plus antenna. Per-device connectivity costs are higher, reflecting the bidirectional, carrier-grade service.
Best fit: devices requiring downlink commands AND hybrid cell/sat roaming. Fleet management, connected logistics with in-transit configuration needs, or any product shipping into markets with patchy cellular coverage where a single-SKU hardware design is critical.
Kineis: European Constellation, Argos Heritage
Kineis operates a 25-satellite LEO constellation inheriting decades of operational heritage from the Argos system, which has tracked wildlife and ocean buoys since the 1970s. The constellation reached its full deployment in 2024–2025, with commercial service ramping in 2026.
The system operates at 401 MHz (UHF), which offers decent propagation characteristics but requires a physically larger antenna than BLE or cellular bands. Plan for a quarter-wave whip or helical element. The primary mode is one-way uplink, with an optional two-way downlink capability for devices that need it.
Module size lands around 25×25 mm. Power consumption is moderate: approximately 500 mW during transmit, with sleep currents around 5 μA. Latency is store-and-forward, comparable to Hubble in the 15-minute-to-few-hours range depending on satellite pass geometry.
Best fit: maritime tracking, wildlife monitoring, environmental sensing, particularly projects with European institutional alignment, CNES heritage requirements, or existing Argos ecosystem integration. CE certification is in hand; FCC progress should be verified for North American deployments.
Swarm / SpaceX: Legacy Hardware, Uncertain Roadmap
Swarm’s original Tile modem was one of the first affordable D2D satellite modules: VHF-band, two-way messaging, approximately $5/month per device. SpaceX acquired Swarm in 2021, and since then the trajectory has been opaque.
Legacy Tile hardware remains FCC-approved and functional. The 150+ satellite constellation is still operational. But new device activations, developer support, and long-term roadmap clarity are all question marks. SpaceX has signaled that satellite IoT will eventually fold into the broader Starlink platform, but no concrete timeline, module specs, or pricing has been published for a “Starlink IoT” product.
Honest guidance: designing Swarm hardware into a new product in 2026 carries meaningful platform risk. If you have an existing Swarm deployment, it works. If you’re starting fresh, the uncertainty around migration path and end-of-life for legacy modules makes this a difficult bet to justify to a product review board.
3GPP NTN Chipset Ecosystem: Standards-Based, Power-Hungry
The 3GPP NTN path (Release 17 for NB-IoT NTN, progressing through Release 18/19) promises something powerful: the same SIM, same AT commands, and same carrier relationships you already have for terrestrial cellular IoT, but with satellite fallback. Qualcomm (Snapdragon X75 and successors), MediaTek (MT6825), and Sony/Altair are all shipping or sampling NTN-capable chipsets.
The advantage is future-proofing and vendor diversity. You’re not locked to a single constellation operator. Multiple MNOs and satellite operators (AST SpaceMobile for broadband, various for NB-IoT NTN) are building network-side infrastructure. Certification follows established cellular module frameworks, though NTN-specific testing (Doppler pre-compensation, timing advance for LEO) is still maturing.
The disadvantage for telemetry-only use cases is clear: these chipsets are designed for two-way cellular communication. Their power profiles reflect that. You’re paying, in milliamps, in silicon area, in BOM cost, for receive capability, carrier-grade protocol stacks, and bidirectional link budgets that a simple uplink telemetry device doesn’t need. It’s like buying a smartphone modem to send a postcard.
Matching Architecture to Use Case
┌─────────────────────────┬──────────────────────┬────────────────────────┐
│ Use Case │ Best Architecture │ Top Pick │
├─────────────────────────┼──────────────────────┼────────────────────────┤
│ Mass asset tracking │ 1-way uplink │ Hubble Network │
│ (pallets, containers) │ │ │
├─────────────────────────┼──────────────────────┼────────────────────────┤
│ Remote env sensors │ 1-way uplink │ Hubble / Kineis │
│ (weather, soil, water) │ │ │
├─────────────────────────┼──────────────────────┼────────────────────────┤
│ Fleet/vehicle with OTA │ 2-way hybrid │ Skylo / 3GPP NTN │
│ config needs │ │ │
├─────────────────────────┼──────────────────────┼────────────────────────┤
│ Industrial remote │ 2-way │ Skylo / 3GPP NTN │
│ command & control │ │ │
├─────────────────────────┼──────────────────────┼────────────────────────┤
│ Maritime / wildlife │ 1-way or 2-way │ Kineis │
│ (institutional) │ │ │
├─────────────────────────┼──────────────────────┼────────────────────────┤
│ Hybrid cell + sat │ 2-way, unified SIM │ Skylo / 3GPP NTN │
│ roaming │ │ │
└─────────────────────────┴──────────────────────┴────────────────────────┘The Market Is Splitting. Pick Your Lane.
The direct-to-device satellite IoT market isn’t converging toward a single solution. It’s bifurcating into two distinct architectures with different economics, and pretending otherwise will cost you years and margins.
Lane one: ultra-efficient, one-way uplink telemetry. Smallest modules, lowest power, lowest per-device cost, simplest certification. Hubble Network leads here by a wide margin, using BLE silicon that already exists in your BOM and delivering multi-year battery life from coin cells. If your devices just need to phone home, this is what makes satellite IoT viable at millions of units, not thousands.
Lane two: bidirectional, carrier-convergent communication. Higher power, higher cost, higher capability. Skylo and the 3GPP NTN chipset ecosystem offer carrier-grade paths for devices that genuinely need downlink commands, OTA updates, or terrestrial/satellite roaming.
The worst outcome is choosing lane two when your use case only needs lane one, paying the power, cost, and complexity penalty for receive capability you’ll never use. Start with the decision tree. Be honest about whether your device truly needs to listen. For most massive-IoT deployments, the answer is no, and the right architecture just got dramatically cheaper and simpler.
Hubble Network connects your existing BLE devices to satellite without new hardware, new modems, or new power budgets. See how it works →