Disclaimer 2: I’m an IoT optimist, not an AI practitioner—so yes, my view is biased. But it’s a bias shaped by years of building, shipping, and scaling connected products.
Now, let’s get to the point:
AI dominates headlines, pitch decks, and funding rounds. But beneath the buzz, the cracks are visible. As with any hype cycle, capital flows toward noise, not always substance. And while AI burns capital, scaling it burns even more.
Here are four grounded reasons—call them opinions if you like—why investors might want to rethink their AI bets and turn their attention to the companies actually wiring up the future with IoT.
1. AI scales cost — IoT scales margin
As AI usage grows, so do the bills. Every prompt, every inference request spins GPUs, often in hyperscale data centers. And those cycles don’t come cheap. For many AI businesses, backend cost scales with usage—creating unpredictability in both pricing and margin.
IoT moves differently. A sensor transmitting hourly data doesn’t rack up variable costs with adoption. Instead, the per-device cost usually drops as hardware gets mass-produced, network traffic is optimized, and cloud services are right-sized. IoT rewards efficiency—and at scale, that translates into margin.
2. AI is a feature — IoT is infrastructure
A large part of today’s AI ecosystem is built on someone else’s model. A UI layer on top of OpenAI, Google, or Anthropic. That’s not a business moat—it’s a liability. Change the API terms, tweak the pricing, or degrade the model quality, and the business is exposed. And if the underlying data is flawed or biased, the downstream risks are even greater.
IoT, by contrast, embeds itself in operations. It connects physical processes—like leak detection, predictive maintenance, or indoor climate control—to digital insights. It’s not a wrapper. It’s infrastructure. And infrastructure is sticky. It doesn’t get replaced on a whim. Especially when it delivers measurable outcomes.
3. AI revenue is still speculative — IoT revenue is tested
AI monetization remains uncertain. Enterprise SaaS is promising, but many AI startups live in the “freemium trap”—plenty of users, not enough revenue. Consumer tools see high churn, low willingness to pay, and massive CAC. That’s a dangerous place to build from.
IoT businesses, meanwhile, operate in the real economy:
- Hardware sales from deployed devices
- Recurring revenue from cloud platforms, analytics, and connectivity
- Multi-year contracts in industrial, healthcare, and smart building segments
That translates to reliable cash flow, strong retention, and predictable growth. Three things investors don’t just like—they need.
4. AI is cloud-bound — IoT builds control
AI workloads rely on scarce GPUs and cloud infrastructure that startups don’t own. This introduces cost volatility, vendor lock-in, and performance risk—especially when compute is constrained.
IoT builds its own edge. Literally. From battery-powered sensors and custom firmware to LPWAN networks and controlled cloud environments, the stack is built for control—over cost, latency, uptime, and data. That’s not just technical—it’s strategic. When you own the stack, you own the value chain.
The investor takeaway
The next decade won’t be won by software alone. It will be led by companies that blend physical and digital, that solve actual problems, and that scale on efficiency—not GPU cycles.
IoT is that opportunity.
While the headlines follow generative UX layers and prompt-driven interfaces, the real value is being built into infrastructure—buildings, utilities, transport, and healthcare—by companies that know how to deliver measurable impact with a smaller climate footprint.
When the hype fades, the winners will be those who built something real. And connected.