Hello readers!

This week we’re looking at the chemical industry’s move from scheduled to predictive maintenance, why eSIM is increasingly essential to scalability, and more!

The chemical industry has been servicing its equipment on a schedule for as long as it has had equipment worth servicing. Change the seals quarterly, inspect the pump every six months, overhaul the compressor on the calendar — regardless of how any of it is actually performing. That approach is now facing pressure from a different direction. Solvay, the Belgian specialty chemicals company, just extended a global IIoT sensor agreement with supplier IMI — and the numbers behind the deal tell the story: the company went from a few hundred connected sensors across its global production network in 2023 to more than 5,000 today, spread across 25 plants in 11 countries, with a target of 9,000 by 2027.

The approach being replaced is time-based maintenance: service machinery on a fixed calendar regardless of how it's actually performing. Change the oil every six months, inspect the pump every quarter, overhaul the conveyor belt every year. It's intuitive and easy to schedule, but it has a fundamental flaw. Equipment doesn't degrade on a schedule. A motor running under heavy load in a dusty environment might fail well before its maintenance date. A different motor in a climate-controlled facility doing light work might run indefinitely beyond it. Calendar-based maintenance is often inefficient because it treats all equipment as though it ages uniformly — which means you're either servicing too soon (wasting resources) or too late (incurring failure costs). Solvay's COO noted that expanding connected sensors allows faster, better-informed decisions on safety, reliability, and energy efficiency. The implicit acknowledgment: the old model doesn't actually give you that.

The chemical industry in particular has strong incentives to get this right. Chemical plants operate rotating equipment — pumps, compressors, agitators, conveyor systems — under demanding conditions involving heat, pressure, and corrosive materials. Unplanned downtime in that context isn't just expensive; it's a safety event. The U.S. Chemical Safety and Hazard Investigation Board has recorded more than 500 serious chemical incidents since 2020, resulting in roughly $1.8 billion in property damage. Connecting sensors to that equipment creates an early-warning system: the NEON vibration sensor, for instance, continuously monitors the vibration fingerprint of a rotating asset and triggers an alert the moment it detects a deviation — before the deviation becomes a failure. Early detection changes the maintenance calculus entirely. An engineer dispatched to investigate an anomaly is a scheduled cost. An unplanned shutdown in a chemical plant is an operational and safety crisis.

The broader shift underway is from condition-based monitoring — which is reactive, telling you something has changed — to genuinely predictive maintenance, which attempts to forecast failure windows before they arrive. It's a meaningful distinction. Condition monitoring tells you that a bearing's vibration pattern is abnormal. Predictive maintenance, using ML models trained on historical failure data from that asset type, can estimate how long that bearing has before it fails. The data infrastructure has to come first — you can't train models on sensor history you don't have yet. Those 5,000 sensors aren't just monitoring equipment today; they're building the dataset that makes true predictive capability possible tomorrow. The investment compounds over time.

If you're thinking about where this applies to your own operations, the entry point is simpler than it used to be. Battery-powered LoRaWAN sensors like Solvay is deploying are wireless, don't require major infrastructure investment, and can be retrofitted to existing equipment without production interruption. The harder question is organizational: maintenance teams that have run on calendar-based schedules for years need new workflows, new alert thresholds, and ultimately new decision models for when to act on sensor data. The technology has outpaced the processes in most plants. Solvay going from hundreds to thousands of sensors in roughly three years demonstrates that scale-up is achievable once the model is established — but the model itself takes time and data to mature. If your organization is still doing time-based maintenance on critical rotating assets, the gap between what's possible and what you're doing is growing.

The manufacturing industry is no longer just evolving; it’s being completely reimagined, and at the heart of this shift is the Internet of Things (IoT). Imagine a factory where machines don’t just operate but actively communicate, sharing real-time data about performance, efficiency, and potential issues.

Read more

4G is sunsetting, but full 5G is overkill for most IoT deployments. Where does that leave smart grid, SCADA, oil & gas, and other long-life industrial systems?

5G RedCap — the mid-tier 5G standard purpose-built for industrial and commercial IoT — is the answer. Omdia projects 963.5M RedCap connections by 2030.

Semtech's new whitepaper unpacks the technical architecture, real-world use cases, and 10 critical questions to plan your migration.

You prompt. The agent builds. Then it asks for a database.

Ghost is postgres made for this. Spin one up in seconds. Fork it like a branch. Delete it when you're done. Pay nothing when it's idle.

Your agent gets full sql, mcp support, as many databases. No dashboards. No provisioning. No forgotten dev databases draining your card at month end.

Build a weekend app. Fork the schema three different ways. Throw two of them out. Ghost doesn't care. The next prompt can spin up a fresh one.

You're already vibe-coding the app. Stop wiring up the backend.

Unlimited databases. Unlimited forms. 100 compute/hrs a month. 1tb of storage. Free.

Download Ghost

In this episode of the IoT For All Podcast, Matthias Wagner, Founder and CEO of Flux, joins Ryan Chacon to discuss AI-assisted hardware design for IoT. The conversation covers the historical challenges of hardware design, the current capabilities of AI tools, compressing the hardware iteration cycle, integration challenges, the limitations of AI, and enabling IoT innovation.

YouTube • Spotify • Apple • YT Music • Amazon

Interested in becoming an IoT For All Partner? Reach out here!