A four-person startup in Seattle is making a bet that the next great source of electricity for America’s strained power grid isn’t a new nuclear reactor, a wind farm, or a massive solar installation. It’s the office building you’re sitting in right now.
Edo, founded in 2025 by a team of former Amazon Web Services engineers, has developed software that coordinates the energy systems already embedded in commercial buildings — batteries, solar panels, HVAC systems, electric vehicle chargers — and turns them into dispatchable power assets that can sell electricity back to the grid. The concept isn’t theoretical. The company emerged from stealth this month with $4 million in pre-seed funding led by Pioneer Square Labs and Wireframe Ventures, as first reported by GeekWire.
The timing is deliberate. And the opportunity is enormous.
The United States is facing an electricity supply crisis that has been building for years but has accelerated dramatically with the explosion of artificial intelligence data centers, the electrification of transportation, and the reshoring of manufacturing. Grid operators across the country are warning that demand is growing faster than new generation capacity can be built. The Federal Energy Regulatory Commission has flagged reliability concerns. Utilities are scrambling to add capacity, but permitting new power plants — whether natural gas, nuclear, or renewable — takes years. Sometimes a decade or more.
Edo’s founders looked at this problem and saw something most energy planners were missing: commercial buildings in the U.S. already contain an astonishing amount of energy infrastructure that sits idle most of the time. Battery storage systems installed for resilience. Rooftop solar arrays generating power that often goes unused on weekends. HVAC systems that can shift their consumption patterns without anyone inside noticing a difference. EV chargers that draw power at predictable, adjustable intervals.
“We’re not asking anyone to install new hardware,” said Edo CEO Arun Padmanabhan in an interview with GeekWire. “The assets are already there. What’s missing is the intelligence layer to coordinate them.”
That intelligence layer is what Edo is building. The company’s platform connects to a building’s existing energy management systems and uses real-time grid pricing signals, weather data, occupancy patterns, and utility rate structures to make second-by-second decisions about when to consume power, when to store it, and when to push it back onto the grid. The building owner gets paid for the electricity they export and saves money by shifting consumption away from peak pricing periods. The grid gets a new source of flexible capacity without a single new transmission line.
It sounds simple. It is not.
The technical challenge lies in orchestrating dozens of disparate systems — each made by a different manufacturer, each running different protocols, each with different operational constraints — into a single coherent energy strategy that respects both the building’s primary purpose (keeping people comfortable and productive) and the grid’s needs (matching supply with demand in real time). Padmanabhan and his co-founders — Anurag Gupta, Samantha Liu, and David Chen, all former AWS engineers — spent years building distributed systems at planetary scale. They see a commercial building’s energy infrastructure as fundamentally the same kind of problem: a distributed system that needs coordination.
The comparison to cloud computing isn’t accidental. Edo’s pitch borrows directly from the playbook that made AWS a $100 billion business. Amazon didn’t build new servers for every customer. It took existing computing capacity, made it programmable through software, and sold access to it on demand. Edo wants to do the same thing with electricity. Take existing energy capacity locked inside buildings, make it programmable, and sell it to the grid when it’s needed most.
The economic case is compelling on paper. According to the U.S. Energy Information Administration, commercial buildings account for roughly 35% of total electricity consumption in the United States. A significant and growing share of those buildings now have on-site generation or storage. The Department of Energy estimates that distributed energy resources — rooftop solar, batteries, controllable loads — could provide more than 200 gigawatts of flexible capacity by 2030 if properly coordinated. That’s roughly equivalent to the output of 200 large natural gas power plants.
But almost none of that capacity is being tapped today. The reason is fragmentation. Building owners don’t have the tools or expertise to participate in wholesale energy markets. Utilities don’t have the visibility into what’s happening behind the meter. And the regulatory frameworks governing how distributed resources can participate in grid markets are still evolving, varying wildly from state to state and utility to utility.
Edo isn’t the only company trying to solve this. The virtual power plant concept — aggregating distributed energy resources and dispatching them as if they were a single power plant — has been gaining traction for years. Companies like Sunrun, Tesla, and OhmConnect have built virtual power plants primarily from residential batteries and smart thermostats. Voltus and CPower focus on demand response from large commercial and industrial customers. What distinguishes Edo, at least in its founders’ telling, is the focus on the full stack of building energy assets rather than a single device type, and the use of AI-driven optimization to maximize value across all of them simultaneously.
“Most virtual power plant platforms are device-centric,” Padmanabhan told GeekWire. “They’ll manage your battery, or your thermostat, but not both together. We’re building-centric. We optimize the whole building as a single energy asset.”
The distinction matters because the value of coordination is nonlinear. A battery alone can shift a few kilowatt-hours from one time period to another. But a battery working in concert with an HVAC system that pre-cools a building before a peak pricing window, while simultaneously throttling EV chargers and exporting rooftop solar — that combination can deliver far more value than the sum of its parts. The optimization problem is complex, but it’s exactly the kind of problem that modern machine learning excels at.
Investors seem to agree. Pioneer Square Labs, the Seattle-based venture studio that co-founded Edo and led the pre-seed round, has a track record of backing enterprise software companies. Wireframe Ventures, based in San Francisco, focuses specifically on climate and energy technology. The $4 million raise is modest by venture capital standards, but pre-seed rounds for enterprise software startups typically are. The money will fund engineering hires and early customer pilots.
The broader investment climate for grid-related technology has been intensely favorable. Venture capital firms poured more than $12 billion into energy and climate tech startups in 2025, according to PitchBook data, with grid flexibility and virtual power plants emerging as one of the hottest subcategories. The Inflation Reduction Act’s investment tax credits for battery storage and distributed energy have accelerated deployment of the very hardware that companies like Edo need to exist in buildings. And FERC Order 2222, which requires regional grid operators to allow distributed energy resource aggregations to participate in wholesale markets, has opened a regulatory pathway that didn’t exist five years ago.
Still, the road from a $4 million pre-seed to a functioning business is long and littered with the remains of startups that had compelling technology but couldn’t crack the commercial real estate market’s notorious resistance to change. Building owners are conservative. Facilities managers are overworked. Energy procurement decisions are often made by people who have no incentive to optimize and every incentive to avoid risk. Selling software to commercial real estate is, in the words of one veteran energy tech investor, “a graveyard of good ideas.”
Edo’s founders are aware of this. Their go-to-market strategy, as described to GeekWire, focuses initially on buildings that already have sophisticated energy management systems and on-site generation — typically Class A office towers, university campuses, and large corporate facilities with sustainability mandates. These are buildings where someone has already spent millions on solar, batteries, and smart building controls, and where the incremental cost of adding Edo’s software layer is trivial compared to the potential revenue from grid participation.
The revenue model is performance-based. Edo takes a percentage of the energy revenue it generates for building owners, meaning the building owner pays nothing upfront and only shares value that wouldn’t have existed without the platform. It’s a familiar model in the demand response industry, and it removes the biggest objection in any enterprise sales cycle: the capital expenditure.
Whether Edo can execute on this vision will depend on several factors beyond the founders’ control. Wholesale energy market rules are still being rewritten. Utility interconnection queues — the backlog of projects waiting to connect to the grid — are clogged, and while Edo’s model doesn’t require new interconnection for existing behind-the-meter assets, regulatory clarity on aggregation rules varies by market. California, Texas, and the Northeast are furthest ahead. The Southeast lags significantly.
There’s also the question of scale. A virtual power plant is only valuable if it can aggregate enough capacity to be meaningful to grid operators. One building exporting 50 kilowatts during a peak event is a rounding error. Ten thousand buildings exporting 50 kilowatts each is a 500-megawatt power plant that can be dispatched in seconds — faster than any gas turbine can spin up. Getting from one building to ten thousand is the challenge that will define Edo’s next several years.
The competitive pressure will intensify. Google’s DeepMind has demonstrated AI-driven HVAC optimization that reduced cooling energy consumption by 40% in its own data centers. Schneider Electric, Siemens, and Johnson Controls — the incumbents that already sell building management systems — are all adding grid interaction features to their platforms. And the major utilities themselves are beginning to build virtual power plant programs, sometimes in partnership with startups, sometimes on their own.
But the incumbents move slowly. And the utilities have conflicting incentives — every kilowatt-hour a building doesn’t consume is a kilowatt-hour the utility doesn’t sell. The startup advantage, if there is one, lies in speed and in the willingness to treat the building not as a load to be served but as an asset to be optimized.
Padmanabhan draws an analogy to the early days of cloud computing, when enterprises were skeptical that they could trust their workloads to a shared infrastructure managed by software rather than dedicated hardware managed by their own IT staff. “The same mental shift has to happen in energy,” he said. “Building owners need to stop thinking of their energy systems as cost centers and start thinking of them as revenue-generating assets.”
That mental shift is already underway in some corners of the market. Large REITs like Boston Properties and Brookfield Asset Management have publicly committed to net-zero carbon targets, which requires not just reducing energy consumption but actively managing the timing and source of every electron that flows through their buildings. Corporate tenants — particularly tech companies — are increasingly demanding that their office space be powered by clean energy, and some are willing to pay a premium for it. The convergence of sustainability mandates, rising energy costs, and new revenue opportunities from grid participation creates a window that didn’t exist even three years ago.
For the broader energy industry, the implications of the virtual power plant model extend well beyond any single startup. If distributed energy resources in commercial buildings can be reliably aggregated and dispatched, it changes the calculus for grid planning. Utilities might not need to build as many peaker plants — the expensive, often gas-fired generators that run only during the highest-demand hours. Grid operators might be able to defer billions in transmission upgrades. And the transition to a cleaner grid might accelerate, because flexible demand and distributed storage can absorb the variability of wind and solar generation more cheaply than building new long-duration storage.
None of this is guaranteed. The history of energy technology is full of ideas that worked in pilot programs but failed to scale, that were technically sound but commercially unviable, that were ahead of their time by just enough to run out of funding before the market caught up. Edo is a four-person company with $4 million and a thesis. The thesis is strong. The market is real. The technology is plausible.
But so is the difficulty.
What makes the Edo story worth watching isn’t the company itself — it’s too early to know whether it will succeed or become another cautionary tale. What makes it worth watching is what it represents: a growing conviction among technologists, investors, and even some utility executives that the grid of the future won’t be built by adding more centralized power plants connected by more transmission lines. It will be built by making the infrastructure we already have smarter, more flexible, and more responsive. The power plant of the future might not look like a power plant at all. It might look like the building where you go to work every morning.
And if Edo and companies like it are right, that building might eventually pay you for the privilege.
WebProNews is an iEntry Publication