I just stepped offstage at the Goldman Sachs Carbononmics Conference, where I was on the panel “Power System of the Future." Throughout the entire discussion, one idea kept surfacing: We’re not merely replacing old energy with new. But we are adding a new dimension to what electricity is for.
For a century, electricity served two primary masters: industry and, indirectly, transportation. Now we are adding a third, and it is fundamentally reshaping everything: intelligence.
When Energy Becomes Intelligence
The data center boom is not just another source of demand. These loads are different. They are always on, highly sensitive to disturbances, and growing fast. AI does not simply consume electricity. It turns electricity into computation. Every training run, inference, and query have evolved into kilowatts becoming decisions.
That is a fundamentally different challenge than “transition” implies. Transition suggests substitution: coal to gas, gas to renewables. In reality, we are facing strong load growth in new categories whose behavior looks nothing like traditional industrial demand.
We built the grid for relatively dumb loads: motors, heaters, lights. Now we are plugging in systems that can model the grid itself, predict its behavior, and react in real time. The commodity is no longer just electrons. It is electrons plus information about when and how they are used.
The Intelligence Load
Data centers are at the front edge of this new load class, but they are not alone. EVs, heat pumps, and electrified industrial processes are also flexible and software-driven. They can respond to signals, adjust in real time, and participate in balancing the system. That is powerful if we choose to use it.
The question is not only “how much generation do we build”, but “how much intelligence do we attach to the load we already have.” That is where storage, flexible demand, and analytics meet.
Storage, Supply Chains, and VPPs
Storage dominated the panel discussions for a reason. The economics of 4-hour battery systems are now strong enough for them to firm renewables and provide grid services at scale. In ERCOT and other markets, new designs like real-time co-optimization will increasingly allow batteries and flexible load to get paid for reliability value, not just energy.
Policy is reshaping how this buildout happens. The latest federal rules around “foreign entities of concern” tighten both content and ownership requirements for tax-credit-eligible projects. In practice, that is pushing battery supply chains to diversify beyond China and forcing many players to reconsider where they source cells, modules, and critical components, and how their cap tables are structured.
The result is that pre-tax equity costs are rising, while post-tax equity economics can still be competitive when structures qualify. It is less about deploying the absolute cheapest hardware today and more about rebuilding resilient, compliant supply chains for the next decade.
For Rhythm, this reinforces the path we are already on. We are scaled on the customer side, we deliver over 3TWh of clean electricity with more than 100MW of behind-the-meter PV in our portfolio, and we have a proven platform for granular time-of-use and demand response.
We launched our virtual power plant (VPP) capabilities in 2024, built on direct thermostat integrations with partners like Resideo/Honeywell and Amazon, an in-house DERMS, and full integration into our risk management analytics and risk system stack. The roadmap adds EVs, storage and commercial assets, scaling to tens of MWs of controllable capacity over the next few years.
VPPs are not theoretical for us. They are a risk tool and a supply resource. Real-time device control feeds our position engine, and dispatch data feeds back into forecasting, closing the loop between retail behavior and wholesale risk.
Why Affordability Matters More Than Ideology
Customers do not buy renewable power because it is virtuous. They buy it when it is of better value.
Utility-scale solar and onshore wind are now among the lowest-cost new generation options in the U.S. Even before incentives, they sit at the bottom of the cost curve in many regions. The way to translate that into customer benefit is straightforward: combine cheap generation with customer products and automation that shift usage into lower-cost hours.
PowerShift, our granular time-of-use plan, does exactly that. Customers on PowerShift show about a 14% reduction in usage during the most carbon intensive and expensive periods of the year. They do not have to micro-manage their behavior; the product and automation help them move load out of the worst hours. They see lower bills. The system sees lower peaks. And emissions fall as a side effect.
Innovation as Market Design
Technology alone is not enough. Market design has to catch up.
If flexible demand is going to show up when it is needed, markets must pay it to stand shoulder to shoulder with generation. That means:
clear products in real-time and day-ahead markets that value speed and accuracy
rules that allow aggregated behind-the-meter assets to bid and be paid like capacity
performance-based compensation so that households, businesses, retailers and utilities are rewarded for measurable outcomes
In Texas, recent analysis has highlighted the need to better compensate demand response and behind-the-meter flexibility so it can compete on equal footing with new steel in the ground. That is the right direction. When VPPs can earn a stable, transparent revenue stack, capital will treat them as a real alternative to traditional infrastructure.
What Makes This Moment Different
Three forces are converging.
First, load is growing across multiple categories at once: AI, EVs, electrified HVAC, and reshoring of industrial activity. This is not a one-off shock. It is a structural trend.
Second, the cheapest new generation is increasingly clean. Renewables did not just win in policy. They won in levelized cost. That advantage persists even as incentive structures evolve.
Third, software can now orchestrate distributed flexibility at scale. Platforms like our home-grown SYMPHONY platform let us forecast prices and load at the device level, automate product and pricing decisions, and control assets through a VPP that is natively tied into our risk stack and customer experience.
The combination means we can grow electricity consumption, support new forms of economic activity, and still decarbonize and reduce risk, if we focus on the orchestration layer.
That is what “addition, not transition” means. We are not replacing one static system with another. We are adding intelligence, flexibility and coordination to a system that is getting larger and more complex every year. The companies that win will make that complexity feel simple and valuable to customers, while capturing the economic upside of running a smarter grid.
