Here's a mind-bending fact: We're about to spend $720 billion rebuilding America's electrical grid, using the same design principles that created today's mess.
It's like rebuilding the Titanic, just with more lifeboats and the same iceberg ahead.
Our grid is actually efficient at long-distance transmission, we lose about 5-6% moving power across high-voltage lines. But that debate is a distraction. The real catastrophe happens at the last mile, in the distribution network and inside buildings. According to Department of Energy data, the cumulative losses through transformers and building systems mean that commercial buildings waste roughly 30% of delivered energy, five times what we lose in transmission.
We build billion-dollar power plants and string high-voltage lines across states, only to have commercial buildings throw a third of that power away.
Why? Because we design buildings without intelligence or coordination. Lawrence Berkeley National Laboratory found that plug loads in offices now consume nearly 50% of total energy, often running full-tilt in empty rooms. Lighting, HVAC, elevators, and IT systems pull power with zero central coordination. Each system operates independently, making real-time optimization impossible.
This would be concerning on its own. But the timing couldn't be worse.
The Coming Load Explosion
Data center electricity demand is projected to grow from ~4% of U.S. consumption today to 10-12% by 2028, driven by AI and hyperscale expansion. A single EV truck charger can pull down a megawatt; the same load as 800 typical homes running simultaneously. McKinsey estimates that meeting this demand will require 300 GW of new generation capacity, roughly equivalent to building 300 large power plants.
And what are we doing about it? Everyone is planning in isolation. AI firms are racing to deploy compute; logistics companies are building chargers; utilities are scrambling to upgrade transformers.
It's like trying to build a symphony with 50 soloists who never rehearsed together.
The result? A brittle system where electricity passes through 4 to 6 transformers before reaching your outlet. Each transformation step introduces losses. The cumulative distribution losses, from substation to building, total about 10%, burning roughly 60 billion kWh per year. That's the equivalent of six entire power plants worth of electricity humming away into nothing, all because we can't coordinate loads efficiently.
There's a Better Way (And It Already Exists)
Denmark rebuilt their energy system around a radically different principle: integration instead of isolation. Over two decades, they coordinated electricity, district heating, and transport into a unified system. Buildings aren't just endpoints, they're active participants.
When wind generation dips, buildings coast on thermal mass from district heating. When demand spikes across the grid, flexible loads automatically adapt without human intervention. Industrial heat pumps shift operation to match renewable generation patterns. The entire system breathes together.
The result? Denmark's grid achieves 99.997% uptime (per Energinet, their national grid operator), integrates 80% renewable energy, and maintains some of Europe's lowest electricity prices. In the U.S., we average eight hours of outages per year—100 times worse reliability. That's not bad luck. That's bad architecture.
Denmark started this transformation in the 1970s following the oil crisis. It took sustained policy coordination and roughly $40 billion in infrastructure investment. Could the U.S. replicate this? Not exactly, we lack their compact geography and centralized planning authority. But we can adopt the core principle: make buildings intelligent participants, not passive consumers.
The Building-First Solution
Most solutions focus on generation or transmission. But if 30% of energy is wasted in buildings, shouldn't we start there? Instead of building bigger pipes, we need to make the endpoints smarter.
This is why we built Novele.
Imagine if every building didn't just consume energy but actively optimized it; where building loads were predictable and the grid became a proactive machine, not a reactive one.
Novele's Energy Boards are ultra-distributed battery storage and control units that install inside the building envelope, not just at the utility meter. Think of them as smart thermostats for every major load in your building, except instead of just HVAC, they coordinate lighting, plug loads, EV chargers, and more.
Each Energy Board learns usage patterns through machine learning, stores energy locally, and communicates with other systems. They automatically shed or shift loads to avoid drawing power during peak demand, reducing both costs and grid strain.
When thousands of these devices coordinate across a district, the impact becomes massive: load predictability improves, peak demand drops by 20-40%, and infrastructure strain softens. Instead of planning for 10 megawatts of chaotic demand, utilities see a flexible 6-8 megawatt baseline with predictable peaks.
Why hasn't this happened already? Three reasons: split incentives (building owners don't pay for grid upgrades), regulatory barriers (utilities are compensated for capital expenditure, not efficiency), and technology gaps (until recently, distributed coordination at scale was impossible). Those barriers are finally cracking.
A Fork in the Grid
Right now, over 3,000 gigawatts of renewable generation (per Lawrence Berkeley National Laboratory's interconnection queue data) sits waiting an average of four years to connect. We're trying to plug cutting-edge generation into a grid designed for centralized coal plants. It's like connecting Starlink to a rotary phone.
We face a choice:
1. Keep building brittle, overbuilt, siloed infrastructure; spending hundreds of billions on capacity we'll only use 5% of the time.
2. Build a distributed, intelligent, adaptable grid from the building outward. One where every endpoint actively participates in system optimization.
The grid isn't fundamentally a power plant problem. It's an integration and coordination problem. We have the tools. We have the data. We have the computing power to coordinate millions of devices in real-time.
The question is whether we'll use them, or spend another $720 billion rebuilding yesterday's system.
Charles Conwell
CEO, Novele
.jpeg?auto=webp)
.jpeg?auto=webp)
/1x.jpeg?auto=webp)
/1x.jpeg?auto=webp)
/1x.jpeg?auto=webp)