For a decade, including during the most recent legislative session, state leaders in Texas have mostly ignored the potential to reduce energy demand through energy efficiency and demand response. That negligence had massive consequences in February, when a winter storm led to outages in much of the state and hundreds of Texans died.

There were many reasons for the outages; a key one was a record high demand for electricity driven not only by extreme cold but also by inefficient heating and air conditioning systems in poorly insulated buildings.

By focusing more on strategies that use advanced technologies to reduce demand when supplies get tight, Texas can avoid future disasters and strengthen the grid in ways that are better for consumers and the economy.

The Public Utility Commission (PUC), to its credit, is beginning to ask important questions about the demand side of the equation while simultaneously moving quickly to establish weatherization standards for the supply side (electric generators). (Note the Railroad Commission does not appear to be moving quickly to do anything about gas supply, one of the largest failures behind the February disaster.)

According to its Chairman Peter Lake, the PUC is “taking a blank slate approach for a full overhaul and redesign”’ of the ERCOT market. Commissioners are asking questions, holding workshops, and then asking more questions, driving toward a redesign by the end of the year. This is an incredibly aggressive timeline for a very complex market. The first comments in the process were filed August 16, and the next workshop will be August 26 (the schedule is here).

A few commenters in the market redesign proceeding alluded to a fundamental shift in how Texas’ electric system works. For the past 100+ years there was a simple formula: determine how much demand there is, then build enough large generating plants (coal, gas, or nuclear) to meet that demand and cover a reserve, usually around 15-20%.

This model is fast becoming obsolete because there’s a better way: cheaper, cleaner, and more resilient to extreme weather.

With each passing year, energy consuming devices such as air conditioning units are becoming more technologically advanced. They are getting smarter. This allows them to respond—with their owners’ consent and control—to signals from energy managers in ways that benefit their owners.

In addition, with each passing year, we are adding more and more variable sources of power to the grid: 80% of new capacity added globally last year was renewable.

As a consequence, we are rapidly moving toward the possibility of a grid in which demand can be shaped automatically—with payment for consumers who opt in but without requiring their active participation—around the supply of electricity, rather than the other way around.

It’s an inversion of how things have been done since the days of Thomas Edison, but it’s likely how the system will work for future generations. In a highly regulated environment (as energy markets still are, even in “deregulated” Texas), regulators can hold back the future—or they can speed it up.

I recently read David Eagleman’s fantastic book Livewired: the Inside Story of the Ever-Changing Brain. The way that the brain automatically receives signals and adjusts on the fly is a great analogy for a truly smart grid, one in which energy-using equipment and devices respond to signals from the grid to reduce cost and increase reliability. As Eagleman puts it:

We can also use principles of livewiring to distribute electricity with much greater efficiency than we do now. As we build out the Internet of Things (the connection of everyday devices to the net), we can push and pull resources from our colossal constellations of lights, air conditioners, and computers—using the internet as a giant nervous system that distributes electricity where and when it’s needed … Beyond increasing efficiency, a smart grid could be able to withstand attacks by healing itself. Most countries of the world claim to be working to implement versions of smart grids, but the truth is that there are various levels represented by the word “smart.” Your third grader is smart, and Albert Einstein is smart. We’ll slowly transition from a smart grid to a genius grid as we come to understand and implement the principles of livewiring that Mother Nature has conceived over billions of years.

Our power grids, including ERCOT, are not third-grade smart. They’re infants. For Texas’ electricity system to do the kind of learning that an infant would do, absorbing stimulus and signals and acquiring skills rapidly, we need a policy framework to support demand flexibility. We need markets to value demand-side resources. We need price signals to bring these resources into the electric market.

Let’s look at a few examples of demand flexibility. In February, Texans and people across the country were justifiably outraged at images of downtown buildings lit up as people shivered in the darkness. Fast Company featured a platform from Intel in various stages of deployment around the country (but not yet in Texas):

“The grid, which is sensing the power needs across all of its territory, can turn things off and know exactly what is turning off, and keep other things on,” says Michael Bates, global general manager of energy at Intel. A hospital could stay connected, while other buildings nearby could shut down temporarily. The platform is also designed to work with renewable energy coming from multiple, intermittent sources instead of the traditional model of one source of power from a centralized coal or gas plant. Being able to quickly tap into distributed sources of renewable energy makes the grid more resilient, since it may be possible to get power from one location while others are down. The tech can also help manage the demand for electricity, so utilities can incentivize customers to use power at certain times or use less when there’s reduced supply.

Examples abound. Think about electric vehicles. Most projections say we’ll have many millions on the roads—and in Americans’ garages—within a few years (Texas already has about 30,000). Many of those will be medium and even heavy duty vehicles with batteries in a single vehicle drawing—and storing—enough power to supply dozens of homes. If they all charge in the early evening as people are getting home from work and the sun is beginning to set, we’ll have problems. But most people, on most nights, will plug their car in and not need it until morning. If it can charge overnight, it will still be ready to go when the owner needs it. A neural grid would recharge cars when power is cheap, clean, and abundant—and even discharge electricity back onto the system (with owners’ consent, and with owners getting paid) when power is scarce, expensive, and typically supplied by the dirtiest plants.

In comments to the Texas PUC, a new entrant to the Texas market, OhmConnect, noted: “Reliability of the power grid is no longer only about having enough generation and transmission to meet the needs of load, but rather, is about how to manage the increasingly dynamic relationship between resources and loads as our generation mix changes.” And no less an expert than Pat Wood, former Chair of the Public Utility Commission and the Federal Energy Regulatory Commission, wrote on behalf of Hunt Energy Network (HEN):

The existing market rules and protocols do not permit nor encourage residential participation, leaving these resources inaccessible to ERCOT. HEN recommends the Commission and ERCOT investigate regulatory policy and market solutions to provide greater opportunities to integrate customer aggregation and automated DR programs into management of the grid … There are opportunities to expand load response programs by allowing more loads—including residential customers and aggregated behind-the-meter loads and resources—to participate in demand response at all times, including outside of energy emergencies. Using additional eligible load resources as a source to both decrease capacity needs and respond to frequency deviations has advantages over conventional frequency response services because they have fast response times.

If we had been investing in demand-side resources over the last ten years, the winter outages would have affected far less Texans. Like planting a tree, the best time to make these investments was 10 years ago; the next best time is now.

Many future-focused demand response resources are ready to go to work for Texas. There are one million smart thermostats in the state, according to PUC comments from the Advanced Energy Management Association (AEMA), but less than 10% of them are enrolled in demand response programs. So the owners of those thermostats are missing out on potential incentive savings that such programs could offer, even as the state is losing the benefit of this important resource. And Texas will add another 100,000-150,000 thermostats this year alone, but very few of those will be able to respond to a signal to reduce energy use, pay the occupants for their reduction, or help stabilize the grid.

There are too many such opportunities to describe. Pool pumps could throttle down in summer time when electricity demand peaks. So could refrigerators — again, with owners’ participation. There are probably more than 10 million refrigerators in Texas. If they all had a smart plug that could have set their compressors to idle for a few hours before the sun came up on Feb. 15, iit literally would have saved lives.

Demand flexibility is valuable, but it is not sufficiently valued in Texas. If the PUC truly wants to redesign Texas’ grid to increase reliability while lowering costs for consumers, this must be a part of it.