Every time a prospect tells us they are evaluating OmniCloud against AWS IoT Core or Azure IoT Hub, we have to unpack the same misunderstanding. The hyperscaler services and a full-stack IoT platform are not comparable products. They are in different layers.
This post is the long version of that conversation.
What AWS IoT Core actually gives you
A managed MQTT broker. Device identity with X.509 certificates. A rules engine that routes messages to other AWS services. Device Shadow for state sync. Jobs for simple OTA.
That is it.
It is a genuinely good MQTT broker with device provisioning. It scales. The pricing per message is reasonable. If you need to receive telemetry from millions of devices into your own AWS account, it works.
Azure IoT Hub is in the same category, with slightly different primitives. Same story.
What the hyperscaler does not give you
When you build a connected product, you ship something to a customer. The customer opens an app, sees their data, gets alerts, adjusts settings, pays you. That experience requires layers AWS IoT Core does not include:
A multi-tenant dashboard. Your end customers need accounts. Those accounts need scoped views. Customer A cannot see customer B’s devices. Roles within a customer (admin, operator, read-only) need to exist. Invite flows, password resets, SSO for the enterprise tier. None of this is in AWS IoT Core.
A mobile app. iOS and Android. Under your brand. Signed into both app stores. Push notifications. Offline-tolerant. Localized.
A device management UI. Which device is offline? Which version of firmware is it running? Which sensor is attached? When does its battery need replacing? You need this UI before you ship a second device.
A permissions model that scales. Multi-tenant means every data read has to be filtered through a permissions check. If you get this wrong early, the refactor later is catastrophic. AWS IoT Core has nothing here.
Billing and subscription. Your customers pay you. That requires Stripe or similar, plan tiers, grace periods, dunning, tax handling, invoicing. The IoT platform has to know who is allowed to access what based on subscription state.
Alerts and notifications. A sensor goes out of range. Who gets notified? By what channel? With what rate limiting so they do not get spammed? Rule triggers in AWS IoT Core can fire events, but turning that into a production alerting system your customer trusts is months of work.
Reports. Your customers want CSV exports. Weekly summaries. PDF reports for compliance. None of this is in AWS IoT Core.
OTA updates that do not brick devices. Jobs in AWS IoT Core can push firmware. But rolling out a firmware update to a fleet with health checks, version pinning, staged rollout, rollback on failure: that is its own subsystem.
The physical hardware. AWS IoT Core is software. You still need certified industrial hardware in your customer’s environment. Cellular. Battery. Sensors. CE, FCC. Years of work.
What it actually costs in engineer-months
We have seen this number in our own work and heard it consistently from teams we talk to. A connected product, built on hyperscaler primitives, with a mobile app and a multi-tenant dashboard, takes 18 to 24 months to reach a usable first release. That is with a team of four to six engineers: at minimum one embedded, one backend, one frontend, one mobile, one devops, one project lead.
At a fully loaded cost of $150k to $250k per engineer per year, that is $900k to $3M before a single customer is paying you.
And the product you get at the end is not obviously better than what a company who bought a platform shipped in 90 days. In most cases it is worse, because those four engineers were solving problems every IoT company has already solved instead of solving problems in your vertical.
Where a full-stack IoT platform fits
A full-stack platform starts further up the stack. Instead of a broker and primitives, you get:
- The hardware: industrial-grade, cellular, certified, with a modular sensor bus.
- The firmware: already handling deep-sleep cycles, secure telemetry, OTA updates, resilient connectivity over spotty cellular.
- The multi-tenant cloud: permissions, device management, reports, billing, alerting, all already running.
- The mobile apps: iOS and Android, already built, configured with your branding.
Your team’s job becomes: pick your sensors. Configure your branding. Design your customer-facing workflows on top of the platform. Sell.
When AWS IoT Core is still the right answer
If your competitive advantage is your infrastructure (you are a hyperscaler yourself, you are building an IoT platform of your own, or your product requires custom protocols at the transport layer) then hyperscaler primitives are the right tool. You are then one of the small number of companies solving the problem once, for yourself, at a scale that justifies the investment.
If your competitive advantage is your vertical (you understand cold-chain better than anyone, or aquaculture, or industrial HVAC) then the right strategy is almost always to buy the stack and focus on the vertical.
What we built
OmniCloud is the full stack. Hardware, firmware, multi-tenant cloud, mobile apps. Running in production across agriculture and aquaculture deployments, processing millions of sensor readings every month. We built it for ourselves first, and offer it to companies who do not want to spend 24 months rebuilding the same primitives.
If you are starting to price out a year of hyperscaler engineering, talk to us first.