Gridsight, a software company founded in Sydney in 2020 that helps electricity distributors measure and release unused capacity on existing networks, said on September 1, 2026 that it had closed a US$26 million Series B led by Insight Partners, with participation from Galvanize and returning backers Airtree, Energy Transition Ventures and Aera VC. The capital is aimed at scaling deployments with US utilities including Xcel Energy and Avangrid subsidiary United Illuminating, while continuing work in Australia, where the platform underpins a flexible solar export program at Endeavour Energy expected to add 600 MW of solar hosting capacity.
Why Utilities Are Buying Visibility Before They Buy Copper
American electricity demand is rising again after roughly fifteen years of flat consumption, and the U.S. Energy Information Administration expects generation to keep climbing through 2027 to serve data centres and electrification. Utility capital plans have not kept pace with that curve, and new distribution and transmission assets are increasingly gated by permitting, supply chains and rate case scrutiny rather than by capital availability.
That gap is what the capacity-software category sells into. Gridsight’s central claim, repeated across its announcement, is that as much as three-quarters of grid capacity sits unused on average, varying by feeder, time of day and season.
The figure is a company estimate rather than an independently audited benchmark, and it describes average utilisation rather than headroom at the constrained points that actually block connections. The commercial argument survives the caveat: if even part of that headroom is dispatchable, utilities can connect load now and defer steel.
The Platform Rests On A Utility-Specific Grid Model, Not A General Purpose Assistant
At the centre of the product is what Gridsight calls its Foundational Grid Model, a continuously updated, utility-specific representation that pulls fragmented network data into one view. The company positions this against general purpose AI by pairing the model with electrical physics constraints, so that agents assisting planning engineers are bounded by what the network can actually do.
That framing matters commercially. Distribution engineers have been burned by black-box analytics that cannot be defended in a regulatory filing, and physics-constrained modelling is the standard answer vendors now give to that objection.
Gridsight has not published third-party validation of the model’s accuracy, and no independent benchmark for capacity-forecasting software currently exists across the sector.
Endeavour Energy’s Flexible Exports Rollout Is The Reference Deployment
The most concrete evidence in the round is in New South Wales. Endeavour Energy’s Flexible Exports service lifts single-phase household export limits from a static 5 kW to 10 kW for about 95 per cent of the year, curtailing only when local conditions require it, using technology from Gridsight and Newcastle-based SwitchDin.
The system draws on roughly 700,000 monitoring points across the network to forecast solar flows and adjust limits dynamically, and it uses the CSIP-AUS communications protocol pioneered by SA Power Networks. Endeavour estimates the program will unlock an additional 600 MW of solar exports by 2035, worth more than $100 million to its solar customers, and it becomes the default connection offer for new and upgraded systems from late 2026.
“We’ve doubled customers’ solar export capacity within our existing poles and wires,” said Danny Cooper, Chief Executive of Endeavour Energy, in Gridsight’s September 2026 funding announcement.
Xcel Energy And United Illuminating Anchor The North American Push
The two named US customers are structurally different bets. Xcel Energy serves eight states across a mix of high-solar and high-load-growth territories, while United Illuminating operates a dense, constrained Connecticut service area under Avangrid.
Xcel’s electric distribution leadership has publicly described the platform as embedded in how its distribution teams plan and operate, which is a stronger reference than a pilot but stops short of confirming network-wide deployment or contract value. Neither utility has published procurement documentation for the work.
Gridsight’s earlier disclosures put it in front of roughly half of Australia’s distribution businesses, with named users including Essential Energy and Horizon Power domestically, plus Northpower and Waipa Networks in New Zealand. The company also lists UK customers without naming them.
The Competitive Set Around Latent Capacity Is Filling Up Quickly
Gridsight is entering a market that has attracted substantial capital in the past year. Fellow Australian firm Neara reached a A$1.1 billion valuation after an A$90 million Series D, as Kurrantly News reported in February, selling physics-based digital twins that surface hidden network headroom from a different technical angle.
Further upstream, Piq Energy raised a $5 million seed to compress interconnection studies, targeting the same bottleneck through study automation rather than dynamic operations. Incumbent vendors are also in the frame, and Xcel already runs a Colorado virtual power plant on Itron’s DERMS platform alongside Tesla, a reminder that large utilities buy several overlapping tools rather than one.
Differentiation in this category is unlikely to come from model architecture. It will come from integration depth with ADMS, GIS and metering estates, and from whether regulators accept dynamic limits in connection standards.
What The Cap Table Signals About Stage And Currency
Insight Partners reported more than $90 billion in regulatory assets under management as of December 31, 2025, and has invested in over 900 companies, which places Gridsight in the firm’s growth-stage software pipeline rather than its early venture activity. Galvanize, an energy-focused asset manager, is new to the register.
The round follows an A$7.5 million Series A led by Airtree in April 2025, which included Australian cricket captain Pat Cummins among the investors and was worth roughly US$4.5 million at the time. Several outlets have reported total funding near $33.5 million by adding the Australian-dollar Series A to the US-dollar Series B, which overstates the figure; in single-currency terms the company has raised approximately US$31 million.
Third-party trackers put headcount at around 58 as of mid-2026, up from close to 40 at the Series A. Gridsight has said the new capital will fund US team growth over the coming year without specifying headcount targets.
The Benchmarks That Will Test The Thesis
The backlog Gridsight is selling against is real. Lawrence Berkeley National Laboratory counted roughly 2,060 GW of generation and storage actively seeking transmission interconnection at the end of 2025, with a median duration of more than five years from request to commercial operation for projects completed that year.
Two caveats matter for this specific investment case. Those queues cover transmission-connected generation, not the distribution-level load and DER connections where Gridsight operates, and large-load queues are tracked separately.
Demand forecasts are also less settled than the funding narrative suggests. The EIA cut its 2027 Texas load growth projection from 14 per cent to 6 per cent in its August 2026 outlook following a state pause on new data centre development, an indication that the load surge underwriting grid software valuations is policy-sensitive as well as technology-driven.
