Satellite IoT is shifting its accounting ledger

The keywords for commercial spaceflight are shifting from "can we launch" to "can we use it." The September 5th Space Industry Daily mentioned that domestic satellite IoT commercial pilots are accelerating. The number of on-orbit satellites in constellations that have obtained commercial trial licenses is continuously increasing, expanding wide-area low-power connectivity services into fields such as intelligent connected vehicles, marine fisheries, low-altitude mobility, and energy/water resources. On the surface, it looks like there are more pilots; essentially, the revenue model is starting to change its calculation method.

In the short term, satellite IoT first targets scenarios where ground networks don't cover but data must return. Ocean-going fishing vessels, oil and gas pipelines, reservoir hydrology, low-altitude aircraft, and vehicles in remote areas may not need high bandwidth, but they need constant connectivity. Low power consumption, small message payloads, and periodic returns are actually better suited for constellations doing wide-area connections. In the past, many projects got stuck at launch, orbit insertion, and tracking/control operations. Now customers are asking more practical questions: How much does the terminal cost? Can it pass automotive standards? How long does it take to recover from a disconnection? Can bills be issued monthly?

In common delivery models of the past, commercial spaceflight was more like an engineering project: customers bought launch capacity, suppliers delivered launches, and it ended after acceptance. Now satellite IoT is more like telecom operations: providing links in the sky, and terminals, billing, O&M, and industry software on the ground. As the business chain lengthens, barriers shift too. Launch frequency is certainly important, but the key to determining long-term revenue is whether stable connection fees can be collected tomorrow.

The difference between launch services and connection services can be seen from aspects like revenue sources, customers, key capabilities, and risks. Launch services rely mainly on single launches and capacity contracts; customers are mostly governments, constellation operators, and satellite manufacturers; key capabilities lie in capacity, cost, frequency, and reliability; risks come from order fluctuations and capacity competition. Connection services rely mainly on subscriptions, terminals, data services, and O&M contracts; customers include automakers, fisheries, energy, low-altitude, and water resources sectors; key capabilities lie in terminal cost, link availability, SLA, and industry integration; risks come from homogenization, regulation, alternative networks, and O&M costs.

Benchmarking against overseas cases, this direction isn't new. Iridium initially opened the market with terminals and hardware, later gradually shifting focus to industry connectivity and subscription services. Starlink let many people intuitively see that satellite networks could be paid for monthly like broadband, transforming the experience from professional equipment to consumer electronics. More mature models often turn sky-based capabilities into ground-based bills, then find replicable paying scenarios.

In the long run, the real barriers for satellite IoT lie in three things: stable contracts, continuous O&M, and cross-industry replication. Stable contracts come from industry customers being willing to pay for connectivity; one-time subsidies won't hold up. Continuous O&M comes from the dirty work and hard labor of tracking/control, terminals, link quality, and fault handling. Cross-industry replication depends on whether, after making marine fisheries work, the same set of capabilities can be transformed into reusable versions for energy/water resources, low-altitude mobility, and intelligent connected vehicles. If these three things aren't achieved, no matter how large the constellation scale, it might just be assets piled up by capital expenditure.

Local industrial clusters will also benefit; satellite IoT will drive demand for test ranges, tracking stations, terminal OEMs, industry software, and data services. This chain is long, but profits aren't necessarily all in the sky. Many projects ultimately compete on the ground segment: who can make terminals cheaper, who can make industry software smoother, and who can turn data returns into customer-acceptable reports and alerts. Customers have weak willingness to pay for satellites being in orbit; they are more willing to pay for alerts when fishing boats cross boundaries, notifications when pipeline pressure is abnormal, and status returns for vehicles in areas without network coverage.

Here's some cold water to throw on. Satellite IoT isn't necessarily suitable for all scenarios. In cities, parks, near-shore areas, or along highways, if ground networks are already sufficient, satellites struggle to win on experience. It's more likely to win first in coverage gaps, extreme environments, regulatory requirements, and asset security—these rigid needs. Low-altitude mobility sounds like a big space, but safety certification, spectrum coordination, terminal airworthiness, and operational responsibility will stretch the cycle. Energy and water resource budgets are stable, but the project-based nature is strong, easily turning into large contracts with slow payments.

I recently used AI agents and agent harnesses to process materials for two weeks and looked through open-source projects on Hacker News. The feeling is very direct. When AI enters infrastructure industries, what lands first is often turning states, alerts, and O&M records into executable context. Satellite IoT is the same. Connectivity is just the entry point; behind it lies device management, trajectory analysis, anomaly detection, insurance risk control, and scheduling systems. Without this layer of industry software, no matter how stable the satellite link, it's just a more expensive return channel.

Commercial spaceflight shifting from launch fever to connection services means the harder part is just beginning. Launch tests engineering capability; connection tests operational capability. Engineering capability can be built through financing, talent, and testing; operational capability must be polished slowly through customer retention, unit economics, service standards, and cross-industry replication. Constellations with real barriers will ultimately settle on an operational system capable of generating sustained contracts and cash flow.

What satellite IoT competes on next is whether connectivity can become stable, operable, and replicable long-term revenue.