Hello readers!

This week we’re looking at the basics of private 5G, how smart sensors are streamlining commercial insurance, LLMs and the IIoT and more!

If you've been following the conversation coming out of MWC Barcelona, you'll have noticed private 5G showing up everywhere — in factory demos, connectivity announcements, and industrial AI showcases. But "private 5G" covers a lot of ground, and the decision to build or buy one is more nuanced than the vendor enthusiasm suggests. Before you commit to an architecture, here are the questions worth sitting with.

Do you need deterministic latency? This is the question that most cleanly separates use cases that need private 5G from those that don't. Public 5G is designed for best-effort delivery — it'll get your data there, but can't guarantee when. For a smart meter reporting every fifteen minutes, that's fine. For autonomous guided vehicles coordinating on a factory floor, or robotic arms operating in tight synchronization, latency variability is a safety and reliability issue. Private 5G networks, running on dedicated spectrum with no contention from outside users, can deliver sub-10ms latency with a consistency that shared public infrastructure can't match.

Where does your data need to live? Public 5G routes your traffic through the carrier's core network before it reaches your systems. For manufacturing environments handling sensitive production data — machine parameters, quality inspection results, proprietary process data — that introduces both security and compliance risk. Private 5G with a local 5G core means data never leaves the premises: processed, stored, and acted on entirely on-site. This is the architecture behind the on-premises AI deployments generating attention at MWC — they work precisely because the network and the compute are both local, and neither depends on an external connection to function.

What happens when the WAN goes down? Resilience is an underrated part of the private vs public conversation. A public network dependency means your operations are only as reliable as your carrier connection. Private 5G paired with local edge compute can be designed to run in island mode: if the outside connection drops, the factory floor keeps running, the AGVs keep moving, and local AI keeps making decisions. For mission-critical operations, this kind of autonomous local operation isn't a nice-to-have — it's a fundamental architectural requirement.

How complex is your device landscape? Private 5G networks require devices to be certified against your specific network configuration. The ecosystem is maturing — particularly around CBRS spectrum in the US and shared spectrum allocations in Europe — but it remains more constrained than public LTE or NB-IoT. A single device type deployed at scale is manageable. A heterogeneous mix of sensors, gateways, and industrial equipment from multiple vendors is a different proposition, and the certification overhead becomes a real project cost that needs factoring into any build vs buy decision.

Have you done the honest TCO calculation? Private 5G carries real upfront costs: radio units, a 5G core, spectrum licensing where applicable, and ongoing management. Those costs have come down, and managed service options reduce the operational burden considerably. But the honest comparison isn't private vs public on a per-device cost basis — it's private 5G vs the operational cost of the problems you're trying to solve. Unplanned downtime in a manufacturing environment typically runs tens of thousands of dollars per hour. The question worth asking isn't "can we afford private 5G" but "what is unreliable or insecure connectivity actually costing us right now."

Is hybrid the answer you've been avoiding? The private vs public framing is often a false binary. Most mature IoT deployments end up with layered connectivity architectures — private 5G for latency-sensitive operations within a defined perimeter, public 5G or LTE-M for devices that move beyond it, and LPWAN for low-power sensors where bandwidth is minimal. The GSMA's SGP.32 eSIM specification, gaining real traction in 2026, makes this significantly more manageable — devices can switch between network profiles over the air based on location, policy, or availability, without anyone physically touching them. The real planning question isn't which connectivity type to standardize on, but how to design a strategy that puts the right network under the right workload.

Verifying what happened before an entity made a commercial insurance claim is not always easy. Until recently, representatives gathering the relevant information had to rely on customer accounts of the circumstances, along with supplementary data like police reports.

Read more

Last month, 48% of visitors to documentation sites across Mintlify were AI agents—not humans.

Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.

This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.

Your docs aren't just helping users anymore—they're your product's first interview with the machines deciding whether to recommend you.

That means:
→ Clear schema markup so agents can parse your content
→ Real benchmarks, not marketing fluff
→ Open endpoints agents can actually test
→ Honest comparisons that emphasize strengths without hype

In the agentic world, documentation becomes 10x more important. Companies that make their products machine-understandable will win distribution through AI.

Make Your Docs Agent-Ready

In this episode of the IoT For All Podcast, Daniel Lang, Chief Marketing Officer at Toradex, joins Ryan Chacon to discuss navigating the complex and evolving landscape of embedded computing. The conversation covers trends in the embedded space, AI at the edge, recent cybersecurity regulations, transitioning from Windows CE to Linux-based solutions, and practical advice for companies navigating compliance.

YouTube • Spotify • Apple • YT Music • Amazon

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