Pano AI

Snapshot

Pano AI is a wildfire-detection company that mounts panoramic networked cameras on telecom-grade towers, continuously scans the surrounding forest for the earliest visible plumes of smoke, and routes verified alerts to fire agencies and utility operations centers. The company has raised roughly $100 million total through a June 2025 Series B of $44 million led by Giant Ventures and corporate strategics including Salesforce Ventures, Latitude Capital and MassMutual Ventures. It matters now because the US utility sector — in the wake of PG&E's bankruptcy, Hawaiian Electric's Maui liability and the chain of Southern California wildfire-driven insurer withdrawals — has finally made wildfire detection a line-item in utility capital plans rather than a research-pilot curiosity. Pano is the operational pure-play.

Founding story

Pano AI was founded in 2019 by Sonia Kastner (CEO), who previously ran product at Nest-adjacent IoT companies and before that was a product manager at Google and McKinsey consultant. Kastner has described the founding insight as the gap between consumer-camera product quality and the needs of fire agencies: existing wildfire detection in 2019 was dominated by look-out towers staffed by seasonal human spotters, satellite tasking that was too infrequent for early detection, and hobbyist camera networks (AlertWildfire) with limited coverage and inconsistent 24/7 analyst backing. The Pano thesis was to build a vertically integrated hardware-plus-ML-plus-analyst-services product, operate the camera and analyst network itself, and sell detection-as-a-service to utilities and agencies.

The founding team pulled from Google, Lyft, Palantir and Verizon for engineering and operations, and from CAL FIRE and multiple state agencies for domain expertise. The company was initially headquartered in San Francisco; it has since added field operations across the Western US, Hawaii, Australia and parts of Western Europe.

How it works

Each Pano station is a 360-degree camera mast — typically installed on an existing telecom tower, utility ridgeline tower or purpose-built mast at elevation above the surrounding forest canopy. The cameras capture continuous panoramic imagery with roughly 10-second revisit intervals during the day and infrared imagery at night. On-device compute pre-filters the imagery for candidate smoke plume signatures; candidate detections stream to Pano's cloud, where a layered ML model classifies the detection as smoke versus cloud, dust, mist, controlled burn or fog. Detections above the confidence threshold route to a 24/7 Pano Intelligence Center where human analysts verify, triangulate, geo-fix the ignition point using data from multiple cameras, and dispatch alerts to the customer's incident-command system within minutes.

The physical setup is non-trivial. Each station requires line-of-sight to the forest canopy, power (grid or hybrid solar-battery), communications backhaul (fiber, microwave or LTE/satellite), permitting from the tower owner, and ongoing maintenance against wildlife, weather and lightning. Pano operates the station as a managed service rather than selling hardware. In the field this looks closer to a Zayo or Crown Castle network-operations model than a traditional sensor vendor.

Alerts land in the customer's SCADA, GIS or dispatch system. For utilities, this means operations staff can trigger a PSPS (public safety power shutoff) decision with 10-30 minutes of warning instead of trailing the fire. For fire agencies, it means a confirmed geo-fix on an ignition before a 911 report would typically arrive, cutting response time by 20-60 minutes in the field.

Product and business overview

Pano sells a managed service bundle: hardware deployment and operation, continuous imagery capture, ML-based detection, 24/7 analyst verification, geo-triangulation, and alert integration into customer dispatch or SCADA systems. Add-on products include historical imagery review for post-incident reports, vegetation management imagery for utilities doing right-of-way inspection, and smoke-forecasting fused with weather models.

Customers fall into three buckets. Utilities are the fastest-growing: Portland General, Idaho Power, Xcel Energy, Puget Sound Energy, Southern Company and Hawaiian Electric have been named in company press releases and reporting. State and federal fire agencies including CAL FIRE, Oregon Department of Forestry, Washington DNR, Hawaii's Department of Land and Natural Resources, and Texas A&M Forest Service. Private landowners including timber operators and large insurers writing wildfire-exposed property.

Business model and pricing

Pano prices per station per year on a multi-year contract, typically five to seven years, with the physical station bundled into the service fee. Public contract filings and utility PUC testimony have put individual station annual service pricing in the $25,000-$45,000 range depending on remoteness, communications requirements and SLA. A statewide utility deployment of 50-150 stations therefore runs single-digit millions of ARR, and a large multi-state utility can be a mid-eight-figure contract.

The economics are capital-intensive for Pano because the company owns the stations: ARR comes with roughly $75,000-$150,000 of installed equipment cost per station amortised over five to seven years, plus annualised maintenance. Gross margin at scale is reasonable but not SaaS-grade — probably in the 55-70% range as the field operations service model matures. The attractive feature of the P&L is that each station has decade-plus service life once installed, so the net-new-ARR curve compounds against a long-tail of installed-base ARR.

Traction over time

Pano has not published revenue, but a reasonable sketch:

Year	Cameras deployed (approx)	ARR (estimate)	Headcount
2021	<50	<$1M	~30
2023	300+	$5-10M	~90
2024	600+	$15-25M	~150
2025 (post-Series B)	1,000+	$25-40M	~200

(Headcount from LinkedIn; station count estimated from press announcements and PUC filings; ARR implied from station economics.) Pano has publicly described itself as deploying across all US Western fire-prone states, Hawaii, Australia and selected European jurisdictions. The June 2025 Series B included Giant Ventures as lead with participation from Salesforce Ventures and MassMutual Ventures (Insurance Journal, June 2025).

Market analysis

The US annual wildfire-suppression spend by federal and state agencies ran roughly $7 billion in the five-year rolling average through 2024 (Congressional Research Service, 2024); utility-driven wildfire liability settlements and insurer withdrawals have added tens of billions of balance-sheet exposure for electric utilities. Insured wildfire losses topped $12 billion in California alone from 2017-2022.

The addressable spend for detection-as-a-service is harder to size. A reasonable top-down: Western US investor-owned utilities (PG&E, SCE, SDG&E, PacifiCorp, Portland General, Idaho Power, Xcel) collectively operate across tens of thousands of circuit miles of wildfire-exposed territory. If each utility deployed 100-500 stations and paid $30k each annually, that is $500M-$1.5B of utility-side ARR potential just in the US. Add state agencies, timber and insurance, and the serviceable market is likely in the $2-5 billion range annually by 2030, with growth driven by regulatory pressure and insurer underwriting demand.

Competitive intel

• AlertWildfire (University of Nevada Reno + partner universities). Non-profit camera network, broadest spatial coverage in the US West, lower cost but without the integrated 24/7 analyst service and SLA.
• Dryad Networks (Berlin, raised ~€19M). Ground-sensor-based gas detection; detects smoke chemistry rather than plume, with faster theoretical detection but higher density of hardware required.
• Overstory (Netherlands). Satellite + ML vegetation management, with wildfire-ignition-risk product adjacency; stronger in prevention than detection.
• Zeitview (covered previously). Drone and inspection platform used by utilities for vegetation management; some adjacent detection functionality.
• Satellite-tasking vendors (Planet, Capella, Umbra, OroraTech). Space-based detection with shorter revisit intervals as constellations grow, but still minutes-to-hours latency and spatial resolution limitations.
• Buzz Solutions (covered previously). Utility vegetation and asset-inspection AI; adjacent but not the same product.

The specific attack angles: AlertWildfire undercuts Pano on price for state agencies that will accept volunteer-run analyst services; Dryad closes faster on detection latency where ground sensors can be densely installed; satellite constellations will eat into Pano's advantage once hourly revisit at sub-km resolution becomes standard; Overstory and ignition-prevention vendors reframe the problem as pre-ignition risk rather than post-ignition detection.

History and evolution

2019. Sonia Kastner founds Pano in San Francisco. 2020-2021. Early pilots with PG&E, Southern California Edison, and Oregon state agencies. 2022. $20M Series A led by Initialized Capital, with Congruent Ventures, Salesforce Ventures and Convective Capital. 2023. Expansion into Hawaii post-Maui wildfires; added utility customers in the Pacific Northwest and Mountain West. 2024. Expanded into Australia and Western Europe. June 2025. $44M Series B led by Giant Ventures, with participation from Salesforce Ventures, MassMutual Ventures, Latitude Capital, Initialized Capital, Congruent Ventures and Convective Capital, bringing total raised to roughly $100M. 2025-2026. Multi-year expansion contracts with utility customers; management has described the network as crossing 1,000 cameras.

What people say

The case for. Utility PUC testimony across Oregon, Washington and Idaho has specifically credited Pano detections with pre-ignition PSPS decisions and earlier containment on named fires. Portland General Electric's filings have referenced the Pano network as a core element of its wildfire mitigation plan. Insurance Journal and TechCrunch coverage have described Pano as the operational leader in the private detection-as-a-service category. Employees on Glassdoor describe mission-driven culture and reasonable engineering compensation.

The complaints. The capital-intensity of a station-owned-service model is the recurring investor concern: unlike a pure SaaS, Pano cannot grow ARR without proportionate capex, which caps capital efficiency metrics. Some utility customer PUC filings have noted detection-latency performance variability by terrain and weather. AlertWildfire proponents in academic circles have argued that public, open camera networks should subsidise rather than compete with private commercial detection, raising a policy risk that state agencies could standardise on free public tooling.

Outlook: the open question

Does Pano successfully re-architect from a station-owned managed service into a software layer on top of a hybrid network that includes satellite, drone and public cameras? The business works if Pano is the detection operating system — the alerting layer utilities trust, agnostic to which hardware captured the plume. The business is capped if Pano is only the owner of its own camera network, because capital intensity is structural and satellite constellation economics will erode its advantage. The specific answer-condition for the bullish case: by 2027-2028, Pano has fused non-Pano sensor feeds (satellite, drone, third-party cameras, ground sensors) into its alert product and signed utility contracts specifically for the software subscription separate from camera deployment. If instead Pano remains a camera-network operator priced per station, the TAM compresses and the capital intensity caps the valuation ceiling at venture-attractive but not category-defining scale.

How to attack it

Three attack vectors are real. First, satellite-first detection. OroraTech, Planet and Umbra are all building detection products with shortening latency; a wildfire-specialist startup building a thermal-plus-visible constellation (or riding a hosted-payload program) that promises sub-minute latency at national coverage sidesteps Pano's tower economics entirely.

Second, ground-sensor density. Dryad Networks' approach — distributed low-cost gas and smoke sensors — has latency advantages inside dense sensor coverage areas and could out-perform tower-based cameras in specific utility rights-of-way where sensor density is affordable.

Third, software-only detection layer. A vendor offering utility operations teams a fused-signal alerting platform that ingests AlertWildfire, satellite, drone and private cameras could commoditise Pano's camera network. The specific weaknesses in Pano's current position: station-owned capital intensity makes gross margins SaaS-inferior; the 24/7 analyst-center cost scales with alert volume; and the hardware-plus-service bundle makes it hard to compete on individual tower deployments where a utility already owns the tower.

Adjacent-segment play

The core capability is distributed outdoor computer vision at telecom-grade reliability with ML plus human analyst verification. The clearest adjacent is transmission-line and distribution-line monitoring — the same camera network can watch for vegetation encroachment, intrusions, animal-caused outages and physical security events. Zeitview and Buzz Solutions already compete there; Pano's installed tower network is a natural extension. Insurance underwriting imagery is another: historical and real-time footage from Pano stations is valuable to property insurers writing wildfire-exposed books, and that is likely the strategic logic behind MassMutual Ventures' Series B participation. Border and perimeter monitoring for critical infrastructure (water utilities, pipelines, military installations) uses the same technical stack with a different buyer; Anduril and Rebellion Defense are the competitors there. International expansion into Mediterranean Europe (Spain, Portugal, Greece, southern France) and Mediterranean Chile and Australia follows the same wildfire-threat footprint. The adjacency that does not generalise cleanly is urban security CCTV — the installed base is enormous and defended by Verkada, Flock Safety, Avigilon, Milestone and others; Pano's product is tuned for rural outdoor at scale, not urban density.

Sources and further reading

• Insurance Journal, Pano AI in California Raises $44M Series B to Scale Early Detection, June 2025.
• Fintech Global, Pano AI secures $44m to fight wildfires with AI, June 2025.
• ImpactAlpha, Corporate VCs power Pano AI's $44 million raise, 2025.
• Salesforce Ventures, How Pano AI Is Building the Front Line of Wildfire Defense.
• GlobeNewswire, Pano AI Secures Record-Breaking Investment, 2022.
• Coverager, Pano AI raises $44 million, 2025.
• Enlit, AI-powered wildfire detection startup Pano AI raises another $44M.