Transition from Industry 4.0 to Industry 5.0, where the human being returns to the center of technological development. Why data, AI and IoT create real progress only when designed around people, processes, trust and meaningful adoption — including humanoid robots as one of the most visible examples of advanced technologies entering real working environments.

Every connected product needs four things to work: hardware, firmware, a cloud backend and a user interface. The hard parts are rarely the idea itself — they are the radio design, the certifications (CE, FCC, CRA and more), scaling to volume and managing devices once they are in the field. This talk breaks down where the real complexity and cost sit in a custom IoT project, then shows how the Walter and Walter Cat-1 bis modules fold hardware, radio and connectivity into a single certified module, and how the BlueCherry platform turns CRA compliance, device ownership, OTA updates and vulnerability management into a built-in feature.

Utility infrastructure is often out of sight and out of reach. Although many parts of geothermal, district heating and water systems have been monitored for a long time, parts of the systems have lacked real-time monitoring — critical checks done manually on site, requiring manpower, travelling and resources. Presenting deployments in Iceland, we share our experience in pressure monitoring in hot-water distribution networks, including critical supply pipes, filtration systems and key points across the grid, and how LoRaWAN connectivity enabled reliable data collection in challenging conditions.

Most IoT teams optimise sleep modes, duty cycles and radio settings to extend battery life. But the single biggest energy drain on a LoRaWAN or NB-IoT device is the transceiver — and nobody talks about what you're actually sending through it. This talk shares how we validated up to 84% lossless payload reduction on a production sensor platform, cutting transmissions in half and extending battery life from 9 to 17+ years with no hardware change. It covers the real-world integration challenges (f32 precision, prediction-model sync, heap constraints on 4KB MCUs), the honest benchmarks, and why structural intelligence on top of sensor data opens a new category of proactive monitoring.

Cloud-dependent IoT systems face growing challenges related to latency, bandwidth, energy consumption and data privacy as deployments scale across industry, utilities and logistics. This talk explores how TinyML — running machine learning directly on small, low-power devices — enables intelligent monitoring at the edge without constant cloud connectivity. Through practical examples, the session shows how this approach reduces latency, lowers communication costs, improves energy efficiency and strengthens data privacy.

Valencia is already full of sensors — traffic, air quality, bikes, trees — but most of that data stays invisible to the people living there. This talk shows what happens when you take those sensors already scattered across the city, open up the API, and turn them into a single live layer that anyone can look at, not just city departments or specialists.

Large-scale LoRaWAN deployments face increasing operational complexity as gateway fleets grow to thousands of devices. Gathering data is important but not sufficient on its own: engineers struggle to quickly diagnose the root causes of gateway incidents amid a flood of metrics, logs and system events. AI can help transform this overwhelming data into clear, actionable troubleshooting insights.

For years, eSIM has promised to simplify global IoT connectivity, but adoption has been slower than many expected. The arrival of the new SGP.32 specification could change that by introducing a framework designed specifically for IoT devices. This presentation examines how eSIMs can benefit enterprises, how SGP.32 addresses longstanding barriers to adoption, and how it will disrupt the IoT value chain.

As connected devices mature from telemetry endpoints into intelligent, autonomous systems, a new paradigm is emerging: Physical AI, where AI actively operates within the physical world. This session presents Physical AI as the natural evolution of IoT, and how AWS and partners are helping customers build a new generation of machines with multi-modal physical intelligence.

Technology moves fast. Teams don't always keep up — not because they lack talent, but because the system around them isn't clear. David Baldoví, industrial engineer, founder of Brandsummit and creator of Do it Clear, shares how a simple principle — clarity before speed — transformed the way his studio works, and why the same logic applies to any team navigating complexity, deadlines and digital transformation. A practical talk on culture, leadership and the kind of clarity that makes teams actually work.

As IoT deployments scale into safety-critical environments — industrial automation, connected vehicles, smart infrastructure — securing these systems demands more than monitoring network traffic. Physical-layer threats such as sensor manipulation, actuator tampering or anomalous timing often go undetected by conventional network-focused intrusion detection. ITI researches the intersection of Edge AI and cyber-physical security, developing lightweight AI models for resource-constrained embedded devices that enable autonomous, low-latency threat detection at the edge — covering the demand for holistic cyber-physical intrusion detection in IIoT and the role of the EU Cyber Resilience Act as a market driver.