Gridware

Snapshot

Gridware clamps a solar-powered sensor the size of a shoebox onto electric distribution poles and uses on-device signal processing to detect the physical and electrical anomalies that cause outages and ignite wildfires. The company has raised approximately $97M across four rounds — a $5.3M seed in May 2021, a $10.5M seed extension in June 2023, a $26.4M Series A led by Sequoia in February 2025, and a $55M Series B led by Tiger Global and Generation Investment Management in November 2025. Co-founder and CEO Tim Barat is a former Australian lineman who was on service trucks during the 2009 Black Saturday fires; he and two Berkeley co-founders (Hall Chen, CTO, and Abdulrahman Bin Omar, COO) launched the company out of the CITRIS Foundry in 2020 and Y Combinator's W21 batch. By November 2025 the company's Gridscope devices were installed on tens of thousands of poles with utilities serving roughly 40% of U.S. electricity customers, and a UC Berkeley Haas benefit-cost study published February 4, 2026 reported a marginal benefit-cost ratio of 13.4 for adding Active Grid Response on PG&E's 80 riskiest circuits.

Founding story

Tim Barat grew up in Victoria, Australia. He dropped out of high school and became an electrical lineman as a teenager, which put him in service trucks during the 2009 Black Saturday fires that killed 173 people — many of the ignitions started on the same conductors his crews patrolled. He emigrated to the U.S. in 2013, enrolled in community college, and transferred to UC Berkeley for EECS (B.S. '19, M.S. '20). The 2018 Camp Fire — 85 dead, PG&E bankruptcy — convinced him this was a venture problem rather than one a utility would organically fix. At Berkeley's Sutardja Center for Entrepreneurship & Technology he met Abdulrahman Bin Omar, a chemical engineering master's student with industry and product-development experience, and recruited his undergraduate classmate Hall Chen (EECS '19) as CTO.

The three incorporated Gridware in 2020 through the CITRIS Foundry at Berkeley, won a $150,000 California Energy Commission grant, got into Y Combinator's Winter 2021 batch, and closed a $5.3M seed in roughly three weeks (announced May 24, 2021), co-led by True Ventures (Priscilla Tyler) and Fifty Years (Seth Bannon). The founding thesis is drawn directly from Barat's field experience: linemen find faults by driving miles of road and climbing poles. Put a smart listening device on every tenth pole, let it hear the acoustic signature of an insulator failure or a tree branch impact, and you can dispatch a troubleman to the exact structure before an arc ignites dry grass.

How it works

The Gridscope is a self-contained pole-mounted device that clamps around a conductor without de-energizing the line and installs in under 15 minutes per unit. It carries transducers and piezoelectric sensors picking up acoustic spectra, vibration, pole tilt and environmental conditions, sampling roughly 6,000 measurements per second (per EEPower, 2024). On-device analog and digital signal processing filters raw waveforms so only inferred events — a tree impact, a conductor slap, an incipient high-impedance fault, a cracked crossarm — hit the backhaul. Solar panels keep the device alive during outages, which is the point: traditional SCADA goes dark the moment a feeder opens, while Gridware's whole value proposition is maintaining visibility when the grid is down.

Backhaul is cellular with satellite as a fallback for remote spans. The cloud platform (Gridpatrol) feeds alerts into utility outage management systems through API integrations — PG&E wired AGR directly into its OMS to speed dispatch. A single device covers a span rather than a single pole, so a feeder can be instrumented at a fraction of pole count; Gridware's PG&E deployment is approximately 10,000 Gridscopes across 1,000+ miles of line (February 2025).

Product and business overview

The commercial wrapper is called Active Grid Response (AGR): hardware fleet plus edge ML plus cloud analytics plus utility-OMS integration plus a managed services component — Gridware installs and operates the devices. The company is explicit that the long-term defensibility is in the data, not the hardware. Four-plus years of waveforms labeled with the physical outcomes behind them produce fault-classification models that no new entrant can approximate from a cold start. International expansion into the U.K. and continental Europe is on the Series B roadmap, as is a dedicated Applied Research Complex in Richmond, California.

Business model and pricing

Published pricing does not exist. The commercial pattern — triangulated from utility filings, investor materials and industry comps — is a hardware CapEx component (bought outright by the utility or recovered through rate base) plus an annual per-device SaaS fee for analytics, firmware and operations. Industry chatter puts the fully loaded first-year cost in the low four figures per device, dropping to a few hundred dollars per device per year in steady state, which utilities can rate-base under wildfire mitigation plans. Because wildfire capital spending is overseen by state regulators (CPUC in California, PUCT in Texas, etc.), the actual lever on Gridware's growth is less an enterprise sales cycle than a wildfire mitigation plan approval cycle — a cycle Barat has been working deliberately since 2020.

Traction over time

• May 2021: $5.3M seed closed; devices deployed in two states.
• June 2023: $10.5M seed extension; operating in five states; 1M+ hours of field data; headcount up ~20 people.
• February 2025: $26.4M Series A led by Sequoia; 18 utility customers; PG&E expanded from pilot to approximately 10,000 poles across 1,000+ miles; trailing-12-month revenue up 7x; some customers reported outage durations down more than 70%.
• November 2025: $55M Series B led by Tiger Global and Generation IM; tens of thousands of devices in the field; utility partners collectively serving ~40% of U.S. electricity customers; Duquesne Light added 1,700+ devices after its pilot and won the 2025 Charles Steinmetz Top Innovator in Grid Operations award.
• February 2026: UC Berkeley Haas difference-in-differences study: median 16% reduction in outage duration (CAIDI) per circuit protection zone; 24% in outage-prone areas; 48% fewer wildfire ignitions on monitored circuits with fast-trip enablement; marginal benefit-cost ratio of 13.4 for adding AGR on PG&E's 80 riskiest circuits.

Named customers include PG&E, Xcel Energy (Colorado), Duquesne Light (Pittsburgh), Brigham City Public Power (via UAMPS), and Portland General Electric. The remaining 14 of 18 Series-A-era utilities are undisclosed.

Market analysis

In September 2025 PG&E announced a $73B capital plan through 2030 for grid hardening, with wildfire mitigation central to it. A 2023 Buchalter analysis of California's filed wildfire mitigation plans had PG&E and Southern California Edison combined spending $23B through 2025. Nationwide the market for distribution-grid intelligence — devices and software — is a conservative $5–15B TAM through the decade, concentrated in geographies with wildfire exposure (California, Oregon, Washington, Colorado, Utah, Texas, Hawaii, and increasingly the Southeast) and in regulators who approve higher ROEs for proactive mitigation. Importantly, insurers and reinsurers are now effectively indirect customers: utility liability from wildfire-triggered ignitions is nearly uninsurable without demonstrated monitoring, which is itself a tailwind for Gridware's buying motion.

Competitive intel

LineVision (Boston, ex-National Grid alums) sits one level up on the grid at transmission with non-contact line sensors and dynamic line rating; it closed a $37.5M Series D in September 2026, bringing it to roughly $87M+ raised, and counts National Grid, AES, Southern Company, TVA, NV Energy and Exelon as customers. Sentient Energy is a pole-mounted sensor incumbent now owned by Hubbell; it has an installed base but moves at Hubbell's cadence. Schweitzer Engineering Labs is the de facto standard in substation protection and could extend downward into pole-level sensing on a long time horizon. G&W Electric sells the reclosers and switchgear that execute fast-trip settings — more partner than competitor. Heimdall Power out of Norway raised ~$43M and is focused on dynamic rating and asset health. In-house utility pilots at Oncor, SCE and others round out the field.

Gridware's edge over the Sentient-class incumbents is on the latency axis — sub-cycle anomaly detection off the physical layer, not just power quality — and on the integration/managed-service side. Its exposure is that SEL or Hubbell can match the sensor hardware; what they cannot easily match is four-plus years of labeled waveforms around real wildfire-adjacent events.

History and evolution

• 2020: Incorporated through CITRIS Foundry at Berkeley; $150K California Energy Commission grant.
• Winter 2021: Y Combinator batch.
• May 2021: $5.3M seed (True Ventures / Fifty Years).
• Late 2022: PG&E pilot begins in high wildfire threat areas.
• June 2023: $10.5M seed extension (Fifty Years / Lowercarbon Capital); JPMorgan venture debt line. Operating in five states.
• September 2024: Brigham City Public Power averts a brushfire during Peach Days festival after a vulture strike ignites a 21 kV line; Gridscope flagged the exact structure.
• February 2025: $26.4M Series A (Sequoia); PG&E expanded to 10,000 poles.
• November 2025: $55M Series B (Tiger Global / Generation IM); HQ now 945 Front Street, San Francisco, with Applied Research Complex in Richmond, CA; U.K. / Europe expansion begun.
• February 2026: UC Berkeley Haas benefit-cost study published, marginal BCR of 13.4.

What people say

The case for. Trade press has been consistently positive. T&D World covered Gridware's smart-pole platform in September 2024 and gave Barat a speaking slot at T&D World Live. Utility Dive republished both the Series B announcement and the Berkeley study. Duquesne Light won the 2025 Charles Steinmetz Top Innovator in Grid Operations award for its Gridware deployment — about as close to independent utility endorsement as the industry produces. Impact Alpha and Time's Best Inventions 2022 both profiled the company favorably. Glassdoor reviews, admittedly a thin sample on a sub-100-person company, cite strong leadership and mission alignment; the CTO himself left a public review praising team quality.

The complaints. Independent third-party evidence is still thin. The February 2026 study was authored "in partnership with Gridware" by a Berkeley Haas project scientist on PG&E data — directionally credible but not independent in the way an EPRI or NIST study would be. Case studies on gridware.io are, by nature, vendor-selected incidents with vendor-estimated counterfactuals: the "3,200-acre fire that didn't happen" at Xcel is a model, not a measurement. The 7x revenue claim at Series A came off an undisclosed and almost certainly small base. And utilities are notoriously slow to generalize from pilots — the 18 Series A customers include many at single-digit-device pilot scale, not production. Finally, hardware durability over 10-plus years in UV, snow and wind is not independently verified.

Outlook: the open question

Gridware works if Gridscope reaches ~1M production poles across 20+ utility production deployments by 2028 before Hubbell/Sentient, SEL, or a utility-built alternative achieves functional parity on wildfire-adjacent anomaly detection. The hardware is reproducible; the moat is the labeled dataset of fault events across geographies and the regulatory relationships that let a utility rate-base monitoring at scale. If Gridware spends the Series B primarily on manufacturing capacity and dedicated utility integration teams, and the Berkeley BCR survives external audit, it has a credible path to grid-standard status in wildfire geographies. If it over-indexes on international expansion or lets Sentient/Hubbell catch up on acoustic anomaly detection before U.S. mitigation plans lock in multi-year contracts, it becomes a middling grid-intelligence vendor.

How to attack it

The specific wedge is distribution-side predictive maintenance on equipment failure, not wildfire. Gridware's dataset and story are oriented around wildfire-adjacent anomalies — vegetation strikes, high-impedance faults, pole tilt — because that is where regulatory urgency and insurer money is. But the ratepayer-financed wildfire mitigation plan window is a California/West phenomenon with a political half-life. Everywhere east of the Mississippi — PJM, MISO, East Texas, the Northeast — the economic driver is SAIDI/SAIFI (reliability metrics state PUCs use in rate cases) and aging transformer and switchgear fleets.

A credible challenger builds a pole-and-span sensor whose primary job is predicting transformer, splice and switch failure two to six weeks ahead so utilities prevent outages rather than respond faster to them. The hardware spec is similar: acoustic plus vibration plus harmonics plus temperature, cellular backhaul, solar power. The go-to-market is different: lead with Entergy, Dominion, Duke, FirstEnergy and Exelon in the Southeast and Mid-Atlantic; sell as a reliability investment under base rate cases, not under a wildfire mitigation plan. Hire from LineVision's commercial team (they already sell DLR to these utilities) rather than from PG&E's wildfire staff. Partner with Siemens or ABB to be stocked into standard substation and distribution builds, skipping a cycle that Gridware still has to grind through.

The second wedge is below-the-sensor: a smaller, cheaper, span-level device at roughly $300 all-in rather than Gridware's reported ~$1,000+. Grid-edge density matters for fault localization — if a lower price enables one-per-pole deployment vs. one-per-span, the data advantage can flip. The architectural bet is that compression at the edge beats raw sampling rate, and that modern Bluetooth-mesh backhaul off a few LTE anchor devices matches Gridware's per-device cellular at a quarter of the connectivity cost. Harder engineering problem, but viable, and a path to the ~millions-of-poles tier where Gridware's current cost base struggles.

Adjacent-segment play

The adjacent segment that remains underexploited is behind-the-meter resilience and microgrid orchestration for commercial and industrial customers under a PSPS regime. When PG&E de-energizes a line, every C&I site behind that line goes dark unless it has backup generation or a battery and the ability to automatically island. Today this is handled by a fractured market: Generac for diesel gensets, Enphase and Tesla for solar+storage, Enel X and AutoGrid for load management. Gridware's upstream sensors already know a shutoff is coming earlier than SCADA because they catch wind-driven anomalies in real time.

A sensible adjacent product is a C&I-side box that consumes Gridware's upstream event stream, predicts PSPS triggers six to twenty-four hours out at site granularity, and orchestrates on-site generation, EV charging, HVAC pre-cooling and battery dispatch to minimize operational impact. Pricing per site is far higher than per-pole SaaS, and the customer (a hospital, data center, cold-storage facility, cannabis grower) has budget authority individually rather than through a three-year rate case. The go-to-market lever is bundling with insurance: outage-related business-interruption insurance in California is nearly uninsurable today, and a device that provably reduces downtime moves premiums measurably. The California-only opportunity is a $500M+ annual recurring opportunity and a natural adjacency for a company that already owns the upstream signal.

Sources and further reading

• Gridware, "Gridware Powers Forward with $26.4M Series A Led by Sequoia" (Feb 4, 2025)
• Pulse2, "Gridware: $55 Million Series B Closed" (Nov 2025)
• T&D World, "Gridware Raises $55M to Expand Real-Time Grid Monitoring" (Nov 2025)
• Berkeley Engineering, "Energized for Real-World Impact" (Jan 2023)
• TechCrunch, "Gridware raises $5.3M seed to prevent wildfires" (May 24, 2021)
• PR Newswire, "Gridware Raises $10.5M to Improve Grid Reliability Nationwide" (June 13, 2023)
• EEPower, "Solar-Powered Line Sensors Predict Wildfires" (2024)
• Utility Dive, "Study Finds Active Grid Response Improves Power Reliability" (Feb 4, 2026)
• CITRIS UC Berkeley, "Gridware co-founders discuss CITRIS Foundry and the startup fast lane"
• Gridware Case Study: Brigham City Public Power averts wildfire on festival day (Sept 6, 2024)
• Insurance Journal, "PG&E Investing $73B in Capital Spending Through 2030" (Sept 30, 2025)