Neal on Fire

Greg Klund of Fyre Shield Solutions joins Neal on Fire to explain how a grid-independent, high-volume, water-only sprinkler system can defend homes and properties from approaching wildfires. Learn how the Ember Active Intelligent Control System manages a 27-hour wildfire event, how backup power, water recycling, and communication redundancy work, and how 3D hydraulic and environmental modeling help homeowners and insurers evaluate protection before installation.

Show Notes

Episode summary
In this episode of Neal on Fire, host Neal Spungen talks with Greg Klund of Fyre Shield Solutions about active property defense: a high-volume, water-only suppression system designed to protect homes when a wildfire is on the move. Greg explains how the system works, what happens when power or communications fail, how it relates to insurance, and why zone zero and home hardening come first.

About the guest
Greg Klund is part of Fyre Shield Solutions, a company building automated, fully grid-independent residential wildfire suppression systems that can also be adapted for commercial properties. He draws on 35 years of experience in irrigation, site development, and construction, and began developing the concept after the 2018 Camp Fire affected his wife’s family in Paradise, California.

Topics covered
* What Fyre Shield does: a high-volume, water-only, grid-independent suppression system, mainly for medium to high-end residential properties
* The origin story: the 2018 Camp Fire and the question of whether something could have been done
* Applying large-scale irrigation components, such as those used for parks and schools, to a residential setting
* The Ember Active Intelligent Control System, which uses on-site weather and sensor data and AI to manage operations through a wildfire
* A 27-hour operating window, evacuation protocols, and managing the event until it is cleared
* Why there is no UL-style standard yet, and how each system is individually designed, engineered, and stamped
* 3D hydraulic modeling and 3D environmental wildfire simulation, including wind, ground fire, and canopy fire
* How insurers are starting to look for scientific evidence of risk reduction
* Fault tolerance: overlapping coverage that provides about 90% coverage even if components fail, and four levels of communication
* Backup generators, on-site reservoirs (typically 10,000 gallons), and recapturing up to 70% of water, so the system does not draw on the domestic water supply
* Flow rates of 100 to 150 gallons per minute and the “microclimate” effect on a property
* Pilot projects under NDA, including one over 80,000 square feet
* Why zone zero and home hardening should come first, and how the system fits in as a “special forces” layer
* Using LiDAR modeling to show homeowners how a wildfire event would play out before installation
* Synergies with larger-scale mitigation efforts, including Burnbot
* Future research into performance in 50 to 70 mph winds and beyond

Key takeaways
1. Active, automated suppression can complement, not replace, zone zero and home hardening.
2. Redundancy in power, water, communications, and sprinkler coverage is central to reliability.
3. Recycling water and using an on-site reservoir keeps the system from competing with first responders for municipal water.
4. Modeling and engineering data help homeowners and insurers evaluate protection without hearsay.
5. There is currently no UL standard for this category, so engineering, stamping, and transparent modeling matter.
6. Future development is focused on mega-fire wind conditions.

Resources
https://fyreshieldsolutions.com/
https://nealonfire.com/

Jonathan Tyshler, CEO of Semble AI, joins Neal on Fire to explain how his AI platform turns a fire alarm design workflow that can take hundreds of hours into a five-minute process. Learn how Semble AI generates code-compliant device layouts, bills of materials, and cost estimates from floor plans, how two Stanford sophomores took the idea to Y Combinator, and why the fire protection industry is ripe for automation.

Show Notes

Episode summaryIn this episode of Neal on Fire, host Neal Spungen talks with Jonathan Tyshler, CEO of Semble AI, about how artificial intelligence is changing the way fire protection and life safety systems are designed. Although Semble AI isn’t a wildfire detection company, it addresses a major bottleneck in the broader fire protection landscape: the slow, manual process of laying out and pricing fire systems for buildings.

About the guestJonathan Tyshler is the CEO and co-founder of Semble AI, which he started with Ethan Byers while both were sophomores at Stanford. Before founding the company, he worked at Convergent, one of the largest fire system integrators in the US.

Topics covered
* What Semble AI does: automating fire alarm and life safety system design from floor plans and project specs
* Outputs: code-compliant device placement drawings, a bill of materials, a project cost estimate, and a customizable proposal
* Replacing the manual workflow of dragging and dropping devices in Bluebeam, checking code, and reworking after jurisdiction review
* The origin story, including a co-founder whose home was lost in a 2017 wildfire and the first-hand industry exposure from working at Convergent
* Target customers: system integrators and companies that design, install, or manufacture fire systems
* How code differences across jurisdictions are handled, currently customer by customer
* How the platform works: upload plans, answer five to six quick questions in a chat-style setting, then generate the layout
* Why few competitors are applying AI to the full workflow
* Why the pre-contract bidding and estimating stage can take weeks and cause companies to miss other projects
* Going from PDF-based workflows first to the harder challenge of CAD integration
* Pricing: a per-seat monthly subscription at $500 per seat
* Building credibility as 20-year-old founders selling to industry veterans
* Skepticism in a field where AI hasn’t been widely adopted
* The longer-term vision: a platform for building systems including security, HVAC, electrical, and plumbing

Key takeaways
1. The biggest benefit is time: faster estimates mean more bids, not just lower labor costs.
2. Automating layout, bill of materials, cost estimates, and proposals in one place removes the jumping between tools.
3. Starting with PDF workflows gave the team a practical path while CAD integration is built.
4. Code compliance is handled jurisdiction by jurisdiction, with a goal of nationwide coverage.
5. Credibility and respect for veterans’ expertise matter when introducing new technology to a mature industry.
6. The same approach could extend to other building systems beyond fire.

Resources
https://www.sembleai.com/
https://nealonfire.com/

Subarna Bhattacharyya, founder of Climformatics, joins Neal on Fire to explain how predictive climate intelligence can forecast wildfire risk, fire weather, heat waves, and drought months to a year ahead. Learn how climate forecasting differs from weather models, how the interactive fire weather dashboard works, and how CAL FIRE, utilities, and wildfire mitigation companies are using long-range forecasts to prepare before fire season.

Show Notes

Episode summary
In this episode of Neal on Fire, host Neal Spungen talks with Subarna Bhattacharyya, founder of Climformatics, about how climate modeling can predict wildfire risk before ignition. She explains how long-range climate forecasts differ from weather models, how her team’s dashboards and pilots work, and how utilities, agencies, and wildfire technology companies can prepare months ahead.

About the guest
Subarna Bhattacharyya is the founder of Climformatics, a company building products that predict localized climate months to a year ahead. Her background spans catastrophic risk modeling in industry and climate modeling at Lawrence Livermore National Laboratory, where she worked on abrupt climate change scenarios for the IPCC.

Topics covered

• What Climformatics does: localized forecasts that flag wildfire, fire weather, heat waves, drought, and storms well in advance
• The origin story: from catastrophic risk modeling to climate modeling, sparked by California wine growers’ drought questions
• Weather versus climate: weather models lose skill after a week or two, while climate forecasts cover one month to a year and beyond
• How machine learning combines climate models, observations, and historical data to fill resolution gaps
• A walkthrough of the fire weather dashboard: timeline slider, location selection, and metrics like fire weather index, burning index, and energy release component
• Overlaying transmission lines to spot vulnerable infrastructure at least three months ahead
• Forecast accuracy in the last fire season, with predictions made May 1 for July through September fires
• Coverage for any location worldwide, with 4 km gridded products for California and the US and 25 km globally
• What can and can’t be forecast: conditions that let small fires become large, but not ignition itself
• Pilots with CAL FIRE, a Northern California community choice aggregator, and a Central Valley air technology company
• The Early Alert Warning System for compound events like heat waves followed by high winds, including power purchasing and public safety power shutoffs
• Solar and hydro capacity forecasting, and a healthcare use case predicting hospital surges during heat waves
• Opportunities with detection sensor and drone companies, challenges with insurers’ legacy models, and a five-year goal of removing the word “unforeseen” from climate disasters

Key takeaways

1. Climate forecasts can flag high fire risk conditions months to a year ahead, long before weather models can.
2. Predicting fire weather conditions is different from predicting ignition, and long-range forecasting helps most with the former.
3. Utilities can use early alerts to plan power purchasing and avoid hard tradeoffs during compound heat and wind events.
4. Sensors, drones, and cameras can be placed more effectively when guided by forecast risk areas.
5. The same forecasts support solar and hydro planning and hospital preparedness.
6. Climate forecasting is not weather forecasting, and that education gap is the biggest adoption barrier.

Resources
https://climformatics.com/
https://nealonfire.com/

Paul Schwartz, co-founder and CEO of Senarch, joins Neal on Fire to explain how solar-powered, off-grid LoRaWAN gateways create connectivity where there is no grid power or cellular coverage. Learn how up to 30 days of battery life, gateway relay, and built-in health telemetry support wildfire, flood, and utility monitoring, and why the “plumbing” of IoT matters for forest and climate resilience.

Show Notes

Episode summary
In this episode of Neal on Fire, host Neal Spungen talks with Paul Schwartz, co-founder and CEO of Senarch, about how off-grid, solar-powered LoRaWAN gateways make IoT monitoring possible in places beyond grid power and cellular coverage. Paul explains the technology, the origin story, how wildfire and flood monitoring use it, and what’s next for the company.

About the guest
Paul Schwartz is the co-founder and CEO of Senarch, a Copenhagen-based startup (DTU Science Park) building off-grid, solar-powered LoRaWAN gateways. The company is in its second year of operation, with customers in Europe and the US.

Topics covered
* How Senarch creates a “bubble of connectivity” where no grid or cellular infrastructure exists
* Solar and battery design: three battery options, with up to 30 days of operation on battery alone
* LoRaWAN as the primary protocol, plus support for the German Wirepas-style protocol growing in Europe (see note below)
* The origin story: a university project where four gateways ran 400% over budget on cabling and labor
* Smart connectivity: Wi-Fi, built-in LTE modem, and an optional satellite modem
* Gateway Relay: a “string of pearls” of up to eight gateways bouncing sensor data back to a connected site
* Utility and energy as key markets, working directly and through network operators
* Wildfire monitoring: air temperature, humidity, soil moisture, and tree moisture sensors feeding AI risk models
* Senarch’s role as the vendor-agnostic “plumbing” for sensor data
* Two-way LoRaWAN: reconfiguring sensors and updating firmware over the air
* Core innovations: ultra-low power use (about 1.1 to 1.2 W versus 9 W and up) and a built-in LoRaWAN health sensor
* Gateway telemetry: energy, enclosure temperature and humidity, tilt, and open-enclosure detection
* A plain-language explanation of LoRaWAN, including range, penetration, and acknowledgements
* Flood and water-level monitoring in Denmark, with sensor batteries lasting five to seven years
* Telcos, utilities, municipalities, and environmental agencies, including the Cleveland Water Alliance on the Great Lakes
* Climate resilience as the headline driver for new projects
* Challenges of scaling as a new entrant: building trust, competing on price, and SIM card compatibility
* What’s next: a compact unit combining electronics and battery in one enclosure, planned for the latter part of 2026

Key takeaways
1. Connectivity is the missing link: the best sensor is useless without a way to get its data out.
2. Off-grid solar gateways remove the cost and delay of running power and cables.
3. Gateway relay extends coverage across remote terrain.
4. LoRaWAN sensors can run for years on a battery, making them well suited to forests and remote areas.
5. Built-in health telemetry lets owners fix problems before a gateway fails.
6. Wildfire, flood, and utility monitoring all benefit from the same resilient network.

Resources
https://senarch.net/
https://nealonfire.com/

Madeline Hunt, PhD researcher at Boise State University, joins Neal on Fire to explain how infrasound, low frequency sound below human hearing, can detect and track wildfires without a line of sight. Learn how autonomous sensors work, how arrays triangulate fire location, how smoke and tree canopy don’t block the signal, and what field tests at prescribed burns and an Oregon wildfire revealed about this emerging wildfire monitoring technology.

Show Notes

Episode summary
In this episode of Neal on Fire, host Neal Spungen talks with Madeline Hunt, a PhD student and researcher at Boise State University, about a new frontier in fire monitoring that doesn’t depend on cameras, satellites, or line of sight. Madeline explains how infrasound can track fire behavior, how her team deploys sensors at prescribed burns and active wildfires, and where this technology could fit in the wildfire detection ecosystem.

About the guest
Madeline Hunt is a PhD student at Boise State University whose research uses infrasound to monitor wildfire behavior. Her lab builds its own low-cost infrasound sensors, and her dissertation work and code will be open sourced.

Topics covered
* What infrasound is and how fires generate it through cycles of air heating and rising
* The 4 to 8 Hz range where fire signal appears, well below human hearing
* How infrasound is already used to monitor volcanoes, avalanches, earthquakes, and nuclear tests
* Why infrasound travels kilometers without losing strength, and isn’t blocked by smoke plumes or tree canopy
* Telling smoldering from active flaming combustion, and how frequency drops as a burn shrinks
* What the sensors look like: lunchbox-sized, autonomous, press a button, set on the ground, and walk away
* Using direction-finding and sensor arrays to triangulate where a fire signal is coming from
* Sensor placement under bushes and trees, and windscreens to reduce wind noise
* The central research question: how many sensors are enough for good data at the lowest time and cost
* Baseline noise recording, plus noise from roads, cities, waterfalls, helicopters, and aircraft
* Filtering noise by frequency rather than machine learning
* Tracking a fast hillside burn at an Idaho prescribed burn, even in wet conditions and with patchy fire
* How terrain and atmospheric conditions affect infrasound travel
* Rapid deployment at a roughly 33,000-acre Oregon wildfire: three arrays, 39 sensors, 2 to 7 km from the fire
* Early adopters, including prescribed burn teams and the potential to catch reignitions
* Where infrasound could be in five years, including networks of always-listening sensors
* Possible community and residential applications

Key takeaways
1. Infrasound fills a gap that cameras, satellites, and aircraft leave, especially under canopy and through smoke.
2. Sensors can be placed kilometers from the fire, keeping people and equipment out of danger.
3. Arrays of sensors can triangulate fire location and may reveal the stage of a burn.
4. The sensors are low cost and easy to deploy, and they record continuously.
5. Prescribed burns, including reignition detection, may be a strong first use case.
6. Wind and helicopters are the biggest noise challenges, and both can be filtered by frequency.


https://nealonfire.com/

Markus Tarin, President & CEO of Movitherm, joins Neal on Fire to explain how thermal imaging and IoT cloud monitoring detect fires before smoke and flame. Learn how multi-vendor cameras, LoRaWAN sensors, AI hotspot filtering, and off-grid Starlink deployments protect industrial sites, battery facilities, and wildfire-prone areas. Includes a real case where an automated system stopped a recycling facility fire with no human intervention.

Show Notes
Episode summaryIn this episode of Neal on Fire, host Neal Spungen talks with Markus Tarin, President & CEO of Movitherm, about how thermal imaging has evolved into practical early-warning fire detection for industrial facilities, utilities, and wildfire-prone landscapes. Markus covers how the platform works, how it avoids false alarms, how it’s deployed, and a real incident where automation stopped a fire before it spread.

About the guest
Markus Tarin is the President & CEO of Movitherm, a company with 26 years in business that builds IoT cloud-based monitoring systems. Its gateway connects multi-vendor thermal cameras and sensors to a cloud dashboard that sends alerts when events occur.

Topics covered
* How Movitherm’s gateway and cloud platform integrate multi-vendor cameras and sensors
* Event-based alerting: warnings versus full alarms by text, voicemail, or email, plus fire alarm panel integration
* How the company started: building software for thermal cameras and creating one of the first automated thermal imaging systems for industrial use
* Why thermal detection gives an extra 10 to 20 minutes, sometimes an hour, before flame or smoke
* Setting alarm thresholds after a one-week baseline of the monitored area
* Exclusion and inclusion masks, scheduling, timing filters, and AI to reduce false alarms from vehicles and equipment
* How AI tracks moving hotspots, and why a stationary hotspot, such as an ember that lands and stays, triggers an alarm
* PTZ cameras versus fixed-mount cameras: resolution, cost, and blind spots
* Ideal wildfire use cases: elevated vantage points and hillside communities
* LoRaWAN sensors, battery life of three to five years, and range of miles outdoors
* Off-grid options: solar with battery backup, cellular, and Starlink
* Deployment: self-install V2 plug-and-play (similar to a Ring camera) and a national installer network
* Reliability: self-monitoring, edge gateway redundancy, and multi-carrier SIM cards on first responder networks
* Why the system is fully automated and doesn’t require 24/7 human monitoring
* Industries served: recycling, landfills, warehouses, food and wood processing, industrial laundries, and the EV battery chain
* Beyond fire: condition monitoring for substations, transformers, hydraulic equipment, and the cold chain

Featured story
A material recycling facility lost a tumble dryer to a battery fire, then installed two thermal cameras with an IR window. When the next battery ignited, the system stopped the conveyor, triggered the sprinklers, and sent a text with a facility map link to three people, all automatically. The fire department found it under control, and damage was minimal.

Key takeaways
1. Thermal cameras detect abnormal heat before smoke or flame, buying time to respond.
2. Lithium-ion batteries are a leading cause of fires in recycling and waste facilities.
3. Smart filtering (masks, scheduling, AI) is what makes thermal detection practical and reduces nuisance alarms.
4. Camera placement matters: line of sight and elevation determine wildfire detection performance.
5. Platform integration, from cameras to LoRaWAN sensors, is a major need in wildfire technology.
6. The same technology also monitors equipment health, such as transformers and cold chain.

Resources
https://movitherm.com/