By Ozge Islegen-Wojdyla, Ph.D., Co-Founder and CTO at TrueMeter

‍

Three papers. One conference. All pointing to the same conclusion for commercial energy customers: the regulatory fixes are years away, but the rate optimization opportunity exists right now. 

‍

Why Voluntary Demand Response Won't Save the Grid — and What Actually Will 

At the 2026 POWER Conference at University of California, Berkeley, Haas School of Business. One paper from the morning session has stayed with me all week.

Fiona Burlig, James Bushnell, and David Rapson ran a randomized controlled trial with residential EV owners through Peninsula Clean Energy in San Mateo County. They tested whether managed EV charging programs could shift electricity load at a meaningful scale. They varied financial incentives across a wide range and also tested steeper time-of-use pricing.

The population was favorable, and the technology worked perfectly when activated.

The headline result: enrollment topped out under 5% even at the highest incentive level: roughly 15 to 20% of a typical participant's monthly electricity bill. The aggregate load-shifting impact was statistically indistinguishable from zero. The paper's conclusion is blunt: voluntary managed charging will not solve the distribution infrastructure problem that widespread EV adoption is creating.

This is a residential finding. But the structural insight underneath it travels farther.

The grid's flexibility strategy rests on customers responding to price signals. This paper provides some evidence yet that voluntary response fails. Even with real money, even in a tech-savvy population, and even with a frictionless program.

The authors point toward improved randomization setup in such studies and rate redesign as the more promising lever: demand charges, service level subscriptions, structures that reflect actual grid costs more accurately.

For commercial operators, that shift is already underway. The residential experiment failed because behavior change is hard. Even $40 a month and a light-touch app could not get more than 5% of EV owners to hand over control of their charging schedule. The grid planned for flexibility and got inertia.

The commercial opportunity exists precisely because rate optimization doesn't require behavior change. It requires knowing how your locations consume power and making sure they're on the rate that prices that consumption most favorably.

That is a data problem. And it's the core of what TrueMeter does. 

‍

1,500 Gigawatts Waiting: How Missing Markets Are Bottlenecking the Clean Energy Transition 

Over 1,500 gigawatts of renewable energy is sitting in US interconnection queues right now. Installed wind and solar is a fraction of that. The bottleneck is market design, not technology.

Nicholas Ryan from Yale presented work at the Berkeley POWER Conference at University of California, Berkeley, Haas School of Business that reframes the queue problem in a genuinely clarifying way. The popular narrative is that grid operators are too slow, that studies take too long. The data say otherwise. Studies are largely delivered on time. Projects are choosing to stay in the queue strategically, using their position as an option to learn costs before committing.

The reason is two missing markets.

When a new solar project enters the CAISO queue, it raises network costs for every other project in the same study area. The entering project pays for its own study, but nothing for the cost it imposes on its neighbors.

Meanwhile, spare grid capacity is allocated through a scoring rule rather than a price, even though receiving that allocation makes a project dramatically more likely to actually interconnect. Enormously valuable capacity is given away rather than auctioned.

Projects enter not because they are ready to build, but because they cannot learn their costs any other way.

This matters for commercial energy customers because a queue bottleneck is a supply bottleneck, and supply bottlenecks show up directly in rates.

Charlotte De Cannière's paper on the Belgian market showed what this looks like from the other side: when transmission constraints are not properly priced into market clearing, you get persistent inefficiency that the market cannot self-correct. The European experience is a preview of what happens when the physics of power flow and the economics of market design stay misaligned for too long.

That gap will close eventually through policy reform. In the meantime, rates are pricing in scarcity that is partly artificial, and it's landing on commercial electricity bills right now.

Knowing which rate structure exposes your locations least to that scarcity premium is the most direct lever available today. That analysis does not require waiting for the regulatory fix. It requires knowing how your locations consume power and what alternatives exist in your markets. That's what TrueMeter does.

‍

How Data Center Growth Is Costing Every Commercial Customer in 13 States $9.3 Billion 

My last piece on Berkeley POWER Conference at University of California, Berkeley, Haas School of Business: The afternoon session was all about data centers and electricity markets that covered two rigorous papers:

Jamal Mamkhezri, Ph.D., Xiaochen Sun, and Yuting Yang from New Mexico State used causal methods to show that data center entry in Virginia raised congestion-related electricity prices by $2.49 per megawatt-hour, a 70% increase over pre-entry levels in affected census tracts. Emilia Chojkiewicz, Aneesha Manocha, Umed Paliwal, Duncan Callaway, and Amol Phadke from Berkeley offered a supply-side response: co-locate data centers at existing gas plants with solar plus storage, with levelized costs ranging from $60 to $126 per megawatt hour and a total potential of 333 gigawatts across all existing US gas capacity.

Both findings are important. But the most clarifying thing said all afternoon came from the discussion.

Outside of ERCOT, capacity markets administratively suppress the LMP signal. They dampen exactly the price information that would tell generators and large loads where to locate efficiently. The data centers went where fiber, water, and tax incentives pointed them. The price signal that should have coordinated that decision with generation siting was too weak to do its job. The 70% congestion cost increase is not purely a consequence of data center growth. It is a consequence of market design.

That distinction matters because the cost does not stay with the data centers. PJM capacity prices rose from $29 per megawatt-day in the 2024/2025 auction to $329 in the 2026/2027 auction.

According to PJM’s independent market monitor, data centers drove 63% of that increase, translating to $9.3 billion in additional costs distributed across every commercial customer in 13 states, embedded in rates that most multi-location operators have never decomposed or optimized against.

The regulatory fix will take years. The rate optimization opportunity exists today. Knowing which rate structure exposes your locations least to costs created by decisions made entirely outside your control is the most direct lever available to commercial operators right now.

That is the data problem TrueMeter is built to solve. 

‍