The ADI team participated in multiple events across the city during the second annual Houston Energy and Climate Week in September 2025—which has now grown into a vibrant global event—hosting many of our clients who traveled from as far as Asia and Europe. Houston’s startup landscape feels vibrant and increasingly central to the energy transition; the ION and Greentown Labs have emerged as major innovation hubs, perfectly complementing the established work of the Rice Alliance.
While the week was filled with official events from various organizations, we came away energized and optimistic, finding the most valuable insights came from the honest, granular conversations on the sidelines—a welcome contrast to the conference “slop” of high-level talking points about “collaboration” and “innovation.”
On the macro front, natural gas was all the rage, while the hype around geothermal and nuclear, despite real progress, seems to overshadow their limited near-term impact relative to wind and solar. Meanwhile, the struggles facing circularity and recycling were clear, but incumbents continue unfazed and, in fact, pointed to branding challenges that diminished the sector’s value proposition. A founder of a battery recycling start-up told me in a brief chat that the industry needs to rename “black mass“—the critical minerals-rich powder or filter cake material produced after shredding and crushing spent lithium-ion batteries—to something less demoralizing and more evocative of how valuable it is.
This frankness was essential, and nowhere were the challenges more acute than in the discussions around power demand. A panel on the Texas grid versus the AI gold rush offered a deep dive into some of the most critical, granular insights we gleaned that week. Moderated expertly by Barbara Burger, four leading practitioners—Michael Skelly (CEO, Grid United), David Berry (CEO, Cloverleaf Infrastructure), Arushi Sharma Frank (Senior Advisor, Emerald AI), and Tom Nudell (CEO, Piq Energy)—discussed how utilities, regulators, and tech companies are scrambling to address this challenge.
The surge of Artificial Intelligence (AI) and its accompanying data center boom is creating a fundamental challenge for the U.S. electric grid, with Texas serving as ground zero. The problem isn’t just growth; it’s the unprecedented scale and unique nature of the demand that is “growing up as we speak.” ADI Analytics has forecasted that data centers will account for over 12% of total U.S. power demand—60-90 GW—by 2030.
Here are five critical themes that utilities, regulators, and tech companies are scrambling to address.
The “Home Depot” load is dead: New load complexity
For decades, utilities treated new commercial loads, or anchor tenants, with a predictable, “Home Depot” analogy. The load was simple, constant, and easily factored into planning. The AI data center, however, is a different beast entirely.
- The size of an “anchor tenant” is suddenly growing by 50X, looking bigger than the entire neighborhood behind it. A single load can be the size of a city in one location.
- The profile of these loads is not the same as the Home Depot. The dynamics of the compute inside these data centers are very complex, making the load shift-able and less predictable. In other work, ADI Analytics has looked at how energy efficiency could significantly alter data center electricity demand. Given that cooling systems now consume nearly as much energy as the IT equipment itself, innovation in efficiency is no longer optional.
The complexity has caught some flat-footed. If a load the size of a city shifts instantly due to a single fault or event in the system, the grid’s stability is threatened. This prompted a NERC Level 2 alert asking planners and operators to improve their large load modeling.
Flexibility is the new generation (and transmission)
The era of hyperscalers (massive data center operators) demanding an all-you-can-eat supply of clean, low-carbon, and completely uninterruptible power is ending. While much of ADI’s voice-of-customer research over the last two years showed data center operators claiming they couldn’t afford any downtime, this expectation is proving unsustainable against real-world grid constraints.
Since new transmission and generation take years to build, flexible load is rapidly emerging as the fastest solution to manage the grid, transforming the energy-hogging data center into a valuable “grid asset.” The electric grid is planned for a small window of maximum stress—somewhere between and hours a year—meaning even a tiny bit of ability to reduce data center load during that peak time can eliminate the need to build substantial new infrastructure. For example, Tyler Norris at Duke University’s Nicholas Institute showed that as little as 2 hours of planned downtime every year would allow the U.S. grid to add GW of new load, well above current data center demand forecasts through 2030.
The practical solution involves bundling different forms of flexibility, most notably load control (AI for the AI computing) which can move or pause workflows to provide the utility with headroom. We see this in action with Google’s flexibility tariff agreements with utilities like the Tennessee Valley Authority (TVA) and Indiana Michigan Power to curtail power during peak grid demand, and by joining demand-response programs traditionally reserved for the manufacturing sector.
Crucially, interruptibility will have to be a feature, not a flaw to be avoided, because flexibility is what will unlock grid capacity and accelerate the interconnection of new projects. Grids are adapting: in Texas, ERCOT (Electric Reliability Council of Texas) is increasing its sophistication by moving toward nodal dispatch for controllable loads, which allows planners to model a controllable load as being able to dispatch down to zero, effectively solving the planning problem by removing the “mattress risk” of unreliable demand.
The Texas tug-of-war: Regional differences and political support
Texas is positioned to “win at the load growth race” due to its great wind/solar resources, phenomenal workforce, and fast interconnection queue. However, the challenges within the state vary significantly, creating an internal “tug-of-war.”
Load profiles are vastly different. In West Texas, the industrial load is flat with a very high load factor, making it harder for new loads to connect as they compete with “everybody’s contingencies moving up together.” In DFW (Dallas-Fort Worth), new loads have a higher chance of getting power by being able to fit into the “peaks and valleys” of residential load.
While SB 6 recently passed, greenlighting big investments in the grid, transmission still lags load growth. The question remains whether the expansion will be “enough and fast enough.”
Texans are generally more optimistic about their economic future than other regions. This has allowed the state legislature to pass large spending bills for the grid, such as the $30, $40, or $50 million expansion, because it’s viewed as a “good thing for the state.” This is far more difficult elsewhere where the view is more of a “zero sum view of things.” Partnerships matter and utilities and tech firms must collaborate to balance growth with grid stability.
Cost allocation and the “social compact” challenge
The massive investment required to upgrade the grid has brought the historical “social compact” between utilities and ratepayers under tension. The core question is: who pays for the new infrastructure?
Historically, economic growth was viewed as a rising tide that lifted all boats. Now, specific companies (many owned by the wealthiest people in the world) want to make massive demands on the power system, and we have to figure out how to pay for it.
The world is moving towards tariffs that are “incremental and causal,” meaning the load pays the cost of the new infrastructure it causes. In some places, data centers are going to pay for almost everything, but in others, upgrades will be “rate based,” effectively making homeowners and business owners pay for it.
While concentrated load growth is problematic, no load growth is also “extremely problematic” for electric ratepayers. Electric rates are the cost divided by megawatts or megawatt-hours, so if that consumption is flat or going down, “stuff costs a lot more to build.”
The mismatch of timing and “too hard to build” problem
The most significant roadblock to a successful energy transition is the mismatch of timing between the data centers’ need for speed and the slow pace of infrastructure development.
We have a mismatch between the time it takes to develop transmission and the speed at which data centers want to grow. This forces us into a world of “second best alternatives.”
While the short-term is challenging, the ten-year outlook is more optimistic. The overall cost curves of solar and storage will drive change. It’s predicted that in five to ten years, nearly everyone will have a battery in their house or garage, solving the problem of just having power on the demand side. (As AI’s energy footprint becomes a more pressing issue, we also examine the water intensity of AI workloads and their impact on local communities.) Efficiency is the next frontier; cooling innovation and workload shifting will define winners.
Conclusion
At ADI, we’re committed to helping companies navigate this evolving landscape. Our work across clean power sourcing, data center energy efficiency, and analysis of low-carbon, high-density power solutions (including nuclear) provides the data and thoughtful design essential for building a more sustainable digital future.
– Uday Turaga
About ADI Analytics
ADI is a prestigious, boutique consulting firm specializing in oil and gas, energy, and chemicals since 2009. We bring deep expertise in a broad range of markets where we support Fortune 500, mid-sized and early-stage companies, and investors with consulting services, research reports, and data and analytics, with the goal of delivering actionable outcomes to help our clients achieve tangible results.
We also host the ADI Forum that brings c-suite executives together for meaningful dialogue and strategic insights across the oil & gas, energy transition, and chemicals value chains. Learn more about the ADI Forum.
Subscribe to our newsletter or contact us to learn more.
