Torch Systems Builds Intelligence for Outdoor Infrastructure

Torch Systems is building an observation layer for the physical infrastructure that sits beyond the reach of a control room. The Lafayette, California company combines solar-powered field devices, multi-sensor AI, and an operating platform to help utilities, renewable-energy operators, municipalities, and other asset owners understand what is happening at remote outdoor sites.

The company was founded in 2021 and is led by Vasya Tremsin, co-founder and CEO, and Anton Tremsin, co-founder and CTO. Vasya began working on early wildfire detection after the 2017 Northern California fires. Anton brought more than three decades of scientific research and sensor experience into the engineering. The original question was how to detect fire sooner. The larger opportunity became how to make an entire outdoor site observable.

That expansion is why Torch matters now. The company reports more than 300 TerraNodes deployed across seven states and over 80 protected sites. It has also announced $11.5M in equity and venture debt to move from hundreds of devices toward thousands. The next test is not only technical accuracy. It is whether a hardware company can make manufacturing, installation, connectivity, field service, and customer response work as one repeatable system.

About Torch Systems and TerraNode

Torch pairs its TerraNode field device with the Torch Intelligence Platform. TerraNode combines visible and thermal cameras with acoustic, gas, air-quality, atmospheric, and environmental sensing. It operates on solar power, analyzes signals on the device and in the cloud, and connects over cellular or supported private LTE networks. A Torch gateway can use satellite backhaul at sites without coverage.

The official product page says a TerraNode can be installed in roughly five minutes without wired power or a wired network. Its job is to keep watching conditions that would otherwise require a patrol, a site visit, or several separate monitoring products.

The software brings sites, devices, events, imagery, and environmental data into one view. Torch says its AI learns what normal looks like around an asset and flags meaningful changes. That could be smoke, unusual heat, a misaligned solar tracker, vegetation growth, unauthorized access, abnormal sound, or a damaged structure.

Why Outdoor Asset Intelligence Matters

Utilities and infrastructure operators have invested heavily in digital monitoring, yet the physical environments around their assets can remain surprisingly opaque. A substation, solar farm, transmission corridor, data center perimeter, or public site may cover a large area with limited power, connectivity, and staff nearby.

The traditional answer is often a patchwork. Cameras handle security, point sensors watch equipment, crews inspect vegetation, and weather services provide regional context. Each tool can be useful, but the operator still has to combine the signals and decide whether a change deserves action.

Torch's strategic claim is that several sensing modes in one autonomous device can create better context. A thermal change beside smoke and a visual cue is more informative than a single threshold. An image attached to an alert can help a team decide whether to dispatch someone. The goal is not to replace human judgment. It is to give the responsible person a clearer first look.

Evidence From Real-World Deployments

Torch's solutions page shows the product being used for fire and smoke, equipment health, vegetation, security, operational anomalies, and environmental conditions. The company names utilities, solar operators, data centers, cities, public lands, and other critical-infrastructure owners as target customers.

Dominion Energy is the clearest named customer. Torch says the utility uses its system at solar sites to monitor fire risk, overheating equipment, tracker alignment, and vegetation. The company also presents anonymized case studies involving a renewable-energy owner-operator, a regulated utility, and a national mulch manufacturer.

Torch reports more than 300 TerraNodes in the field, seven states with active deployments, and more than 80 protected sites. Axios reported, based on comments from Vasya Tremsin, that another 800 devices were backordered. These are company-reported operating figures, not independently audited results, but they show the nature of the pressure on the business. A larger order book turns deployment capacity into part of the product.

Leadership and the Company's Technical Roots

Vasya Tremsin's path to Torch began with detection systems built as a teenager. His wildfire work won top honors at the Intel International Science and Engineering Fair, and he later studied at the University of Pennsylvania and UC Berkeley before working at Microsoft and Visa.

Anton Tremsin has spent more than three decades conducting scientific research at UC Berkeley, according to Torch's official About page. The company credits him with more than 300 peer-reviewed publications and sensor deployments in research and aerospace settings.

That combination gives Torch a useful founding pattern: urgency from a visible climate problem paired with deep sensor engineering. It also creates a discipline test. Scientific range must become a product that a utility crew can install quickly, maintain across harsh conditions, and trust when an alert arrives.

Capital, Hiring, and the Operations Signal

Torch announced an $11.5M financing package in October 2026. The transaction includes $9.5M in equity through SAFE notes from Khosla Ventures, Abstract Ventures, Construct Capital, and Starship Ventures, plus $2M in venture debt from First Citizens Innovation Banking. Torch did not disclose a priced-round label or valuation.

The company plans to use the capital for manufacturing, deployment capacity, go-to-market expansion, field operations, and continued platform development. Its About page includes a general hiring contact, but no dedicated careers page or verified current role list is available.

The hiring signal is still meaningful. Torch's constraint is moving from invention to installation. People who coordinate hardware, logistics, customers, connectivity, and field teams determine whether a device reaches the correct location and remains useful after it arrives. In physical AI, operations is not a back-office function. It is part of the customer experience and a source of product feedback.

What Torch Signals for Physical AI

Torch reflects a broader shift in AI from digital content toward continuous interpretation of the physical world. The advantage will not come from attaching an AI label to a camera. It will come from combining useful sensors, reliable power and connectivity, field data, and an interface that fits the decisions operators already make.

The company now has to show that one outdoor intelligence layer can remain accurate and manageable across different assets, climates, risks, and response teams. If it succeeds, Torch will be more than a faster warning system. It will become infrastructure for seeing remote infrastructure before a small change becomes an expensive incident.