1. Podcasts
  2. Inevitable
  3. Texas Grid Rules with Arushi Sharma Frank
Inevitable

19 Nov 2026 · 1h 10m · 29 chapters

Episode artwork

Texas Grid Rules with Arushi Sharma Frank

In short

How power-grid “rulebooks” and market settlement determine who can monetize new tech (especially batteries and flexible data-center loads), with a focus on Texas rules that enable non-firm access and load flexibility.

Guest backgrounds

Arushi Sharma Frank is founder of Luminary Strategies. She previously authored grid rules for technology participation and payment, including work at Tesla (Austin) where she helped open Texas to large grid batteries and wrote programs enabling home batteries as a virtual power plant. She also worked on power-market regulation, including early Dodd-Frank-era understanding of physical vs financial power.

Key claims

The grid is a regulated marketplace layered on physical delivery; batteries and compute can “arbitrage” value across time and location. Grid operators must be able to measure and audit energy-in/energy-out and match telemetry to metering. Texas is becoming a template for other grids to connect demand faster without overloading wires.

Notable examples

Tesla Megapack rule change in Texas (separate energy-in vs energy-out measurement); ERCOT Aggregated DER Program (ADER) for distributed batteries via QSE/telemetry; June 2026 Texas Speed to Power rule PCLR (Provisional Controllable Load Resources) letting data centers plug in sooner if they can curtail when congested; “telemetry offsetting” to improve data-center bankability (Aurora/Google-funded study).

Written by AI. May contain mistakes.

Chapters

1:17

Arushi's Background and Career Journey

Arushi shares her career path leading to her role at Luminary.

“It's growing fast, and so are the clean and resilient solutions meeting it.”
2:08

Transitioning to Tesla and Energy Innovations

Arushi discusses her transition to Tesla and innovative projects in energy.

“So I started in 2011 in the natural gas utilities industry and pretty much instantly began covering markets and grid design or pipeline grid design, safety, environmental, the whole panoply of issues in Washington, D.C.”
3:38

Understanding the Power Grid as a Marketplace

Insights into how the power grid operates beyond just physical infrastructure.

“I'd love to have you explain that to someone who maybe hasn't thought about things that way.”
6:40

Challenges for New Technologies on the Grid

Discussion on the challenges new technologies face in the power market.

“And on the sell side, we've had very conventional ways of selling power.”
8:12

Early Wins at Tesla and Battery Integration

Arushi recounts her early accomplishments at Tesla related to battery integration.

“One of the reasons it is such a big deal is that it starts to move the economics around for both gas and renewables.”
10:38

Rules for Battery Operations and Grid Trust

Arushi explains the rules governing battery operations and their impact on grid trust.

“Tesla made some of the best batteries in the world and was having a lot of success deploying them in lots of places and was having trouble getting it to be recognized in Texas, as I understand it.”
14:00

Understanding Power Measurement and ERCOT

Learn about how power measurements are taken and integrated into the ERCOT system.

“The rule, well, I wrote them so I can tell you.”
16:46

Scaling Battery Systems for Grid Participation

Discover the differences between large and small battery systems in grid contexts.

“Walk through the differences of large monolithic batteries compared to small distributed battery systems and how that had to work with the grid separately.”
19:38

The Role of Telemetry and Metering in Grid Operations

Explore the significance of telemetry and metering in managing power on the grid.

“20 % galvanizing stakeholders outside of politics, just explaining to them that it's possible for small batteries to aggregate and achieve that.”
23:16

Nodal Participation of Data Centers in Texas

Understand how data centers interact with the Texas grid through nodal participation.

“That's like every grid operator's audit is these two data streams putting them together.”
Show all 29 chapters
26:35

Balancing Non-Firm Power and Data Center Operations

Learn how data centers manage power supply and their non-firm status on the grid.

“I've heard you describe it elsewhere as the equivalent of an 18-wheeler that gets on the highway early by agreeing to pull over in heavy traffic.”
28:00

Understanding Load Flexibility in Data Centers

Learn how data centers can respond to demand and grid signals effectively.

“Should you PPA with a nearby generator who will manage a lot of the grid problems for you and reduce your exposure to non-firm power?”
29:00

Texas's Speed to Power Rules and Their Impact

Explore the new rules in Texas that affect data center operations and grid connections.

“Is that the right way to think about it?”
30:27

Navigating Curtailment Obligations

Discover how data centers prepare for and react to curtailment obligations during peak times.

“The grid tells a 400 megawatt data center, hey, step back.”
32:32

Bankability Strategies for Data Centers

Understand the strategies data centers can use for better bankability and risk management.

“I think what you're seeing happens there is that data centers as a monolith actually have a lot of diversity inside them.”
34:35

The Role of Telemetry in Grid Access

Learn about telemetry offsetting and its significance in managing grid access for data centers.

“So after I passed the speed to power rules that came out in June, one of my startups worked with Aurora Energy Research to do a study that was funded by Google.”
36:35

Location Impact on Data Center Performance

Discover how the geographical location of data centers affects their operational flexibility and risk.

“oh no, in five years, I'm only getting a fraction of the power I thought I was going to get from the grid and any other solutions.”
38:46

Battery Integration and Market Dynamics

Explore how large and distributed batteries interact with the grid and their market implications.

“Does it smooth out the expected power that they will be managing in any way?”
41:15

Deregulation's Effect on Distributed Batteries

Examine how deregulation shapes the operation and success of distributed batteries in Texas.

“Not only that, you know, what you just described, this word firewall, is how I talk about the telemetry offset rule, because it's a shield.”
42:00

The Role of Batteries in Texas's Grid

Explore how deregulation impacts the operation of distributed batteries in Texas.

“That allows the grid to see the electrical similarity of what batteries can do relative to where there are problems on their transmission grid much better.”
44:30

Future Challenges for Texas's Energy Market

Discuss the future hurdles in ensuring effective energy transmission in Texas.

“I'm assuming everything we talked about is going to be vastly out of date three or four years from now, just given how fast everything is moving.”
46:10

Startups and Innovative Solutions in Energy

Learn about the intersection of startups and energy market regulations for future growth.

“At the distribution grid level, when it comes to distribution batteries, to the large grid level, when it comes to larger batteries, aggregation of batteries, and flexible data centers.”
49:50

Post-Election Reflections on ERCOT

Examine the implications of the recent elections on energy policies in Texas.

“help them think about how they should be designing their platform to plug into the rules that exist today and the rules that might exist in the future.”
54:30

The Tornado of Election Outcomes

Discuss the impact of election outcomes on governance and energy policy.

“They just need the platform to go deeper, harder.”
56:06

Understanding Voter Sentiment on Technology and AI

Explore the intersection of voter dissatisfaction and technology's impact on society.

“I wrote a piece while I was in this media circuit over the past couple of weeks called The Desert of Dried Up Hope in Tech and AI.”
58:56

Data Centers and Grid Management in Texas

Learn about the challenges and strategies related to data centers and power grid in Texas.

“I mean, there's the affordability angle with respect to data centers and are my power bills going up, yes or no.”
1:00:47

The Role of Flexibility in Energy Assets

Discuss the importance of flexibility in data centers and energy operations.

“study per grid connection and then socialize the rest of that deliverability or that headroom, if you will, you're creating on the grid to the rest of the power grid.”
1:05:30

Legislative Strategies for Data Center Developers

Get insights on how data center developers can navigate upcoming legislative changes.

“I think the audit of the existing data center universe in ERCOT wraps up under the outgoing governor next month, and then the new administration and the new legislative session start in January.”
1:08:25

Optimizing Energy and Climate Impact

Learn about the potential for optimizing gas usage to reduce climate impacts.

“It's certainly self-serving of me to say so, but I feel like the results over the last few weeks are bad for pure behind the meter natural gas growth, which from a climate decarbonization perspective is a good thing.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On Inevitable, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:00Arushi Sharma Frank: Today on Inevitable, our guest is Arushi Sharma Frank, founder of Luminary Strategies. Arushi has a history of authoring the rules that decide how new technologies connect to the power grid and get paid. Through Luminary, she's worked to help startups leverage those rules, including base power and Emerald AI. While at Tesla, she helped open Texas to large grid batteries and wrote the program that lets home batteries act as a virtual power plant. This June, Texas adopted another rule she helped write that lets data centers plug in sooner if they agree to cut back when the grid is congested. Every grid in the country is now facing the same flood of demand Texas faced first, and her work is becoming a template to help other grids connect it faster without overloading the wires.

0:55Arushi Sharma Frank: We recorded this conversation in October, before Texas voters elected Gina Inouza as their next governor. We got back in the studio with Arushi after the election to hear what the results mean for the grid and for her work on load flexibility. We've added that at the end of the episode, so be sure to listen all the way through. From Override Ventures, I'm Cody Sims, and this is Inevitable. Demand for power is inevitable. It's growing fast, and so are the clean and resilient solutions meeting it. Each week, I talk to the people powering the world and keeping it running to learn what comes next.

1:45Arushi Sharma Frank: Arushi, welcome to the show. Thank you, Cody. You had a lot of my favorite people on your show, and I am glad to join the ranks. Well, I'm excited to dive in with you. I'd love to have you start by just tracing a little bit on your own background. We'll spend a few minutes here and just how ultimately it led you to where you are today at Luminary, and then can't wait to dive in on your work. So I started in 2011 in the natural gas utilities industry and pretty much instantly began covering markets and grid design or pipeline grid design, safety, environmental, the whole panoply of issues in Washington, D.C.

2:24moved into doing similar work but in the power sector for the Electric Power Supply Association a few years later and then transitioned into in-house designing and building programs to make it easier and compliant to run things on power grids all the things nuclear gas solar wind pick your favorite and out of there is how I landed up with a job at Tesla the second that they were moving their headquarters to Austin and building their energy presence in their next level of professional energy business model in 2020, which was also the start of COVID.

3:03Arushi Sharma Frank: Yeah, what a time. Yeah. Most recently in this four years, I've done all sorts of crazy things. At that time, we started Tesla's first retail business for energy, their first wholesale business. I changed a bunch of rules for Grizz so that they could accept Tesla's technology on behalf of Tesla's customers and value it correctly. That whole journey really set me up, not unwittingly, but rather wittingly to help startups achieve some of the same successes and also grow the economy for innovation when it comes to tech and power grids. The thing I maybe was most struck by at the highest of levels when we first chatted is most people, including myself, view the power grid as a bunch of wires and power plants and you view it as a rule book and a ledger.

3:57Arushi Sharma Frank: I'd love to have you explain that to someone who maybe hasn't thought about things that way. Absolutely. Every one of the businesses that interacts with a power grid is effectively engaging in arbitrage. Why? Because electrons move in real time. Money moves essentially in real time or value of money moves in real time. And where the physical location of a thing is is not necessarily where the financial value is. So most people don't realize that power grids in the United States, if someone can trade around the value of power moving or not moving, or being contracted to move and not actually getting there and it goes somewhere else.

4:36There's a whole money market on top of that. There's a financial market on how power moves on top of that. And that's independent of the molecules or the electrons themselves.

4:47Arushi Sharma Frank: Most people's familiarity with that problem probably can be summed up in one word, which is Enron. Yeah. And yet, if you did not have the markets that are sitting on top of the physical delivery system, the physical delivery system would fail consumers. So it's interesting to bring people back from the brink of Enron. I actually was the person when Dodd-Frank passed in 2011. I became one of the first people in the United States who understood power markets and gas markets both and wound up appointed at the Commodity Futures Trading Commission to advise this new agency or rather this agency that was new to regulating energy on the difference between physical and financial power.

5:29It's certainly not an understated space. However, for the average consumer that is simply looking their light switch on, it's not something that they would have access to. But if you come to it from where I did, which is frankly before I built compliance for a single power plant, a physical asset, I had already learned the financial and economic structure that ends up pricing every megawatt hour that flows in and out of the grid.

5:58Arushi Sharma Frank: As I understand it, this rulebook decides who can sell, who can buy, how much gets measured. It's a marketplace, as you said. Yet, because it's a regulated marketplace, there are a lot of rules around what can participate in that marketplace. And I think a lot of the work you've done lately is particularly around batteries and new technologies in general. How do new technologies get a seat at the table in participation? Maybe explain a little bit about some of the realities and challenges that non-incumbents have in trying to play in this market that is the grid. Every grid in the United States has basically willing buyers and sellers.

6:45And on the sell side, we've had very conventional ways of selling power. So we've built some nuclear plants, we've built some gas, we've built a lot more gas since the shale revolution and retired coal. And in the middle of all of that, renewables, and you've had so many guests who have already talked about it, renewables fundamentally finance themselves differently than gas. So if you look at the operating picture, gas is a little bit expensive to build, but it can be pretty expensive to operate.

7:17Arushi Sharma Frank: Renewables are very expensive to build and basically cost nothing to operate because of the nature of their operations. For the first time in my career, my career is basically the poster child for the most rapid transition years in the market for the power grid. And the shape and the pace of it has changed. in the middle of trying to figure out how you equivalently reward or attempt to equivalently reward a megawatt hour on the same footing between different types of fuel resources on the grid has become more and more challenging and more and more of a fight that is unfortunately political more than it should be, which is if you want to fight about something, fight about the physics of the power grid and whether or not the electron can actually move.

8:04And it can move, just pay it. And then, Cody, along the way, you know that battery energy storage has become a big deal in America on every level. One of the reasons it is such a big deal is that it starts to move the economics around for both gas and renewables. So it's neither generation. It's not a simple consumer of power. It's an arbitrage block on the grid. And now we're at the point where the need to arbitrage, the need to move electrons to the place where they have the lowest value to where they have the highest value, it hits every sector.

8:40Arushi Sharma Frank: It moves electricity through time, right? That's the easiest way to think about batteries. Through time and space. And if you think about it, the next layer of what's coming for the market in terms of building out a better picture of putting new technologies on the grid, compute is a reverse arbitrage solution. Because if you can keep putting compute in places where power is less scarce and then move it to places when that scarcity curve shifts, then you've also effectively arbitraged where compute happens relative to where there are problems on the grid, where those resources, where batteries, let's say, aren't installed or where renewables and batteries aren't providing you a clean megawatt hour and you want that megawatt hour.

9:23So the whole system is one big marketplace where we're trying to fight to figure out how do you create fairness and equivalence in pricing a megawatt hour, especially if people have preferences. You and I might have a preference that we flick our light switch on and we have clean power coming to us. But the average American just cares about the cost of the megawatt. And then you have the sector, the data center industry.

9:49Arushi Sharma Frank: They don't seem to care about any of that. Yeah, your market question, the reason it's so important and on the Snow podcast host has asked me that yet, is if you think about the market as a consumer, whether or not the market moves at all based on your preferences is incredibly time sensitive. It's a multivariate. And it's a problem everywhere in the U.S. because five years ago, you might not have cared about the quality of an hour coming to your home, but you do. and you're not that different from a data center. You really just want the least cost option. And that's where it gets ugly when you have politics governing how these markets are made instead of reality.

10:28Arushi Sharma Frank: Let's talk about how you have had a few wins in addressing some of the changes that need to happen and maybe start with your early wins at Tesla. Tesla made some of the best batteries in the world and was having a lot of success deploying them in lots of places and was having trouble getting it to be recognized in Texas, as I understand it. What did you help do? So Tesla's battery system has such a high round-trip efficiency because a lot of the functions that use electricity inside the battery system to make it click are just drawing power at the same time as those same functions are selling the power back out.

11:08And if you can't split them, it can be really hard for a grid operator to read that you only consumed two megawatt hours for your internal functions, and the rest of the 98 megawatt hours that are supposed to come out of your system are actually flowing back to the grid. And so they are a wholesale sale to the power grid, and then someone buys that value. In Texas, we landed an entire company thesis in the state. We moved our headquarters there. We could not connect megapacks to the power grid because the grid operator couldn't measure the difference. It was the only battery system, or the first of its kind, to be so tightly wound and integrated and so energy efficient, essentially as a unit, that the grid operators had no idea how to measure the difference.

11:50And if they can't measure the difference, guess what? The operating rules was that we would charge you, like a retail customer of power, for all 100 megawatt hours going into your battery, which wipes out the economics entirely. I walked into the office, And it was my first assignment in terms of like, what do we do? I was like, wait a minute, what? So that Cody really began my journey in achieving mastery of working with grid rules and the rule books of the grid to change those rules. And so over the next year, through a combination of working with my smartest engineers inside the company and working with third parties and stakeholders in the grid's rulebook creation process, I changed those rules so that we could separately account for energy in and energy out and monetize energy out.

12:46Doing so, mind you, when you change the rulebook for the grid this way, the first project that benefited from this, the Gambit Energy Storage Project, was ultimately operated by a third party. And the reason I did it as an OEM was not because Tesla was going to them self-develop and become a large developer in our cot. It's because everybody wanted to buy our technology. So we're creating essentially a tech mutual rule change that will let Tesla's tech plus anybody else's tech show up for the state. And I'm sure you know this, but that one rule change I effectuated has resulted in gigawatt hours of batteries in Texas.

13:23Arushi Sharma Frank: How does the grid operator trust Tesla to operate the batteries in the best way possible according to power in and power out? Because I assume that trust basis on software and operational management, it's very different than a typical form of power generation, which is we need electricity, flip on the switch, generate the power, send it to the grid. This is more of, as you said, an arbitrage over time and space. And in order to get the credit you need for the power you are moving around with a battery, you have to trust that you're making the right decisions on when to charge and discharge the battery.

14:07The rule, well, I wrote them so I can tell you. I remember them like it was yesterday.

14:12Arushi Sharma Frank: How convenient. The batteries themselves, for people's understanding, energy is a measurement of power over a unit of time. Power literally is power. So it's a unit. Power to energy is from the equipment or the thing that's creating the power to the reader of the energy flow. And the thing that reads energy flow, everybody knows, is a meter. So these rules really involve, first of all, Tesla coming up with a technological box, like a technology box, that takes power measurements that are tied to the battery's consumption of energy, like its end-state consumption of energy, and pushes that measurement into a centralized box that can talk to a utility system's meter.

14:59and the utility meter takes those power measurements and converts them into energy measurements and codeine meters look like your background. They have channels in them, they have little lines and bars that can track power measurements from different sources at a site. So you can have one meter that governs part of a data center and governs a part of the plant that's supporting the data center. So similarly I actually wrote the mechanism for Tesla to make a translated measurement that's incredibly precise and put it in the meter box in a brand new slot in a brand new channel of the meter. And then for ERCOT, the grid operator, to receive that information every 15 minutes and put it into their normal system for measuring power in, power out, and then settle that information.

15:41Not only do we change the fundamental grid code that said this is allowed, we're allowed to measure things this way, we also created the actual way to read power to energy accurately. And there's actually also a form that I wrote that's inside the back of ERCOT's protocol, section 10 of their guide, that is a third-party attestation form. So a third-party engineering team can go in and actually pick up, let's say a MECAPAC project has 52 units, can pick up and sample from all of those units to see whether the battery cell degradation and the operating health of the system is good enough to actually justify using that kind of calculation to accurately give the grid operator a picture.

16:28And we do that validation every year, I think. There's one every year, every three years.

16:33Arushi Sharma Frank: It's essentially some kind of auditing mechanism, it sounds like. Absolutely. Yes. So you had this win with Tesla of these large multi-megawatt, megapacks that you were able to help get to participate in the power grid. And then you turned that to even smaller scale distributed batteries, which has helped the whole residential battery boom become a viable business, including one of my portfolio companies, Base Power, which has benefited from it immensely. Walk through the differences of large monolithic batteries compared to small distributed battery systems and how that had to work with the grid separately.

17:12Powerwalls, which is the Tesla system, and Megapack, which is also the very large Tesla system, they connect at different parts of the power grid. You can go out and buy a Powerwall, an Enphase, a base, a peak battery, and put it on your home or inside your home. To represent all of those devices to the grid operator, to let them trust that those devices are doing what they said they would do. It's kind of interesting, Cody. You first have to build out the customer side, which is the whole offer to the customer that says, I'm going to be your electricity provider. I'm going to give you this opportunity that's tied to a battery.

17:49And I'm going to be your financial strategist. I'm going to be your arbitrage provider. You're back to arbitrage, your first question. I'm going to be the creator of the market that allows you to benefit from using your battery as a hedge for your power bill. So that side of the business, I also worked on. and I helped create the thesis for Tesla Electric, which was Tesla's first energy retailer. And on the other side of it, there was this question of ERCOT, the grid operator, allowing all of these new technological assets to combine into one powerful machine, if you will, and provide the same input outputs with the grid as power plants, the same input outputs as one large megapack.

18:32Doing so basically takes us from the world of treating energy storage or these consumer-sided devices as just basic phone call demand response or three days ahead demand response where the grid operator says there's an emergency, can you take power down? Completely wiping that thesis out and replacing it with, we can be so precise with the grid reliability needs of ERCOT. we can give you instantaneous, sub-second, and five-minute response from our battery aggregation fleet. That to me is a true virtual power plant because you are participating in the fundamental physics and economics of the grid as it operates on a five-minute basis.

19:20and you're providing day ahead certainty to the grid because you're able to then aggregate all of these devices and then pitch in and get paid for being available when the grid has a problem. That's the program. The program is called the Aggregated DER Program or ADER and getting it started is a whole saga many people know but let's just say that it involved 20 % politics, 20 % galvanizing stakeholders outside of politics, just explaining to them that it's possible for small batteries to aggregate and achieve that. 40 % working very closely with the grid operator on demonstrations and structuring the demonstration.

20:01And mind you, the demonstration results from Tesla are 69 pages of hard data that we submitted to the commission in a public file with my name on it and said it is incontrovertible that we can participate in your power grid competitively without subsidies, without special favors, the same way gas plants do. and I got it done. I left Tesla in April 2024 and Zach Dell found me a week later.

20:31Arushi Sharma Frank: Amazing. When you described the role of the meter with the large megapacks, it was sort of, hey, you can go audit the individual meter and see the data coming off of it and whatnot. When you're dealing with hundreds or thousands or tens of thousands of distributed batteries, you're now trusting the fleet operator to aggregate the meter data of all of those and manage them accordingly. How does that audit trail work? This is super important because everything you might ask me next about data centers comes back to this. There's meters that measure energy. And then there's telemetry that measures the state of power in a box.

21:10Telemetry is what governs how a grid operator sees that a service is being provided in real time or even at a sub-second basis. They can take that data, which comes to them through some intermediary. All the intermediaries of different states. In California, the intermediary could be a utility, it could be a scheduling coordinator, it could be other kind of market participants. In Texas, that intermediary is called a QSE. In SPP, like the data center hails chills tariffs that everyone keeps talking about, but the utility is going to actually provide telemetry from their substation for these loads that might be conditionally accessed to get grid power.

21:49Back to Texas, telemetry is the state of the system. Metering is energy beating power over a period of time.

21:57Arushi Sharma Frank: And that's settled or looked at over 15 minutes. What is happening with aggregated DER and with power generators that are participating in grid services and getting paid for them is that the grid operator is actually pinging on a very, very quick latency, the fleet orchestrating unit or the power plant orchestrating unit. Everybody has a QSE. So all power plants don't directly talk one-on-one to the grid. They talk to the grid through a QSE. Define QSE. Qualified scheduling entity. Everyone is talking through the same set of intermediaries to the grid operator in real time. So that trust layer is equivalent across all resources that want to participate with a grid.

22:44There isn't anything there that allows some resources to achieve, I would say, an end run around the instructions that they're getting from the grid operator. Back to the question about metering. Good metering, smart metering, high interval metering basically allows a grid operator to look at that data, stretch it out and see that the real time telemetry that's coming from the site intermediary matches what the meter data is showing. But that's the actual audit. That's like every grid operator's audit is these two data streams putting them together. One of them is a bigger data stream over time, and another one is instantaneous.

23:25Arushi Sharma Frank: Super helpful. And you set the stage for this. Home batteries are small, but in aggregate, very, very large, or the theory to be very, very large, hence the idea of them as a virtual power plant. But the new problem that we're now dealing with on the grid, of course, is the biggest power users anyone has ever seen, which are these large AI data center campuses. If I have the money, the land and the chips to build a data center, I can't just plug in. What does the process look like for me to actually get access to power? So are we talking about the suite to power rules that we wrote in ERCOT or in general?

24:05Arushi Sharma Frank: Yeah, let's start with Texas, because I know that's where you've done a ton of work on this. So your guests now know about metering and telemetry. Conventional power grid users just consume from the grid. They're basically signing up for the grid to build a lot of redundancy so that the grid can continue to do a lot of real-time work to make the whole system stable. And in Texas, data centers or any other large load already have the ability to connect into one specific location and have the grid operator treat that location as a resource for the grid, i .e. we could give you a signal through telemetry and you could respond to that signal.

24:47That has called, Cody, in broad terms, nodal participation. And why it's important to your portfolio company base is that they actually have been behind trying to get nodal treatment for their aggregated battery portfolios. Because locational value of arbitrage, whether it is a data center that wants to be flexible or it is a battery that can be flexible and wants to be valued for that. locational specificity and how you value that response and service means they get paid better and get paid differently than if you simply were to give a phone call to a whole region and say, do you have demand today that you can drop?

25:25That's just a completely different way of managing the system.

25:28Arushi Sharma Frank: How precise of locational specificity is typically needed? A substation. And so behind a very large substation, you know, we have different sizes of substations in ERCOT and every other grid operator. The typical large substation at ERCOT has a variety of residential meters behind it, one or two medium-sized loads. Yeah, I was going to say, you're talking a zip code or a partial zip code's worth of power, I guess. Yeah, a power zip code, exactly. And data centers could be a whole zip code on their own. When they get a new substation built, their whole zip code is the data center. If you're going to tell the grid operator today in Texas that at my substation, I'm going to be flexible relative to you, so when you give me a signal, I can respond to it.

26:15I'm basically behaving as a connect and manage resource, meaning I connect it to the grid, and I'm managing how I consume power or sell power relative to the state of the whole system. In the United States, Texas is the first place to fully, fully complete rules that let data centers behave this way, the same way nodal-based power batteries could behave, the same way an individual large megapack could behave, the same way any gas plant that connects to the power grid today can behave, which is I will accept grid signals that ramp my unit up and down in exchange for you giving me power faster than 2032.

26:56Arushi Sharma Frank: I've heard you describe it elsewhere as the equivalent of an 18-wheeler that gets on the highway early by agreeing to pull over in heavy traffic. Exactly. That seemingly would mean these data centers are agreeing to not have five nines of uptime. Is that the correct translation layer for me or is uptime a different argument? Uptime is actually a different argument, but let's pretend it's not for a second so I can answer your question well. If a data center or a battery or a gas plant is non-firm, it's behaving in this grid connection as non-firm, meaning it doesn't get its firm power every single hour of the year, of which, by the way, there are 8 ,760.

27:36Arushi Sharma Frank: It does not mean that the data center needs to take the hit on uptime for the balance of hours where it is drawn down. What it needs to do is find power firming somewhere else. Just can't pull from the grid. Yes. So non-firm to the grid, the more clear rules like the one I have written for Texas are. The more clear rules are that define what is your non-firm exposure to the grid, the more clear it is what you should do to manage that non-firmness at your site. Should you shift compute? Should you shift IT load? Should you add a battery? Should you PPA with a nearby generator who will manage a lot of the grid problems for you and reduce your exposure to non-firm power?

28:21Arushi Sharma Frank: Essentially, where we've gone in the United States, in my view, is deeply unfortunate. Our understanding of the words load flexibility and data center flexibility automatically transfer people's minds to the data center has to curtail compute. As consumers, we've all had demand response events in our own homes. You've had thermostats and stuff, and it's like, yeah, you need to be more uncomfortable. So it's not that. It sounds like what it is, is they're agreeing to, when they get pinged, saying, hey, you need to tone it down right now, that's when they need to either shift to on-site generation, backup, or find some virtual battery resource that they can use to pull power.

29:00Arushi Sharma Frank: Is that the right way to think about it? Not only that, you've hit all of that. You can come work for me whenever you want. Just come on here and have a side jump. Also, the number of hours in which you are perceived as non-firm to the power grid, the number of hours your data center would have to draw down by a little bit to solve a wires problem. on the system can be reduced in real time by you smartly putting paired resources near you or in a location near you to manage the problem itself. So the source of the problem becomes smaller. What you described is all the remedial action the data center needs to take once it has received a curtailment signal.

29:40What I am describing is that the rules that I've helped write in Texas motivate the grid operator to get better information to prevent the problem in the first place.

29:50Arushi Sharma Frank: These rules passed earlier this year, I believe. Maybe let's name the rule and describe a little bit about when it went into effect. The main Speed to Power rules are in a planning guide revision request 145-145. The name of the rule is Provisional Controllable Load Resources, or PCLR. And it passed in June 2026 as part of the larger package of the brand new way that large loads slash data centers will get studied to connect to the grid overall, which is known as the new batch process intercut. Just to make it concrete, it's August, it's hot, it's Texas. The grid tells a 400 megawatt data center, hey, step back.

30:36Arushi Sharma Frank: What happens over the next 10 minutes? Yeah, it's kind of fun. So the grid already gives instructions to everybody who is connected to the grid every five minutes. Here are your choices for what happens to a data center on a hot summer day as of right now. If you are a data center that does not care to participate in one of these voluntary programs, ERCOT is currently writing rules that will put you into a non-voluntary version of turning down. In the next 10 minutes, what happens is probably not the window in which you know that you'll have a curtailment obligation. You'll know that because in the morning it's just too hot and the grid's already peaking at 9 a.m.

31:14And so you're getting prepared to accept some sort of curtailment on a non-voluntary basis. The problem is you don't know what that is at all because it is a completely administrative, out-of-market, phone call type, last-ditch signal from the grid operator. Which might be delivered through your transmission utility. because you didn't choose to connect to the grid in a sophisticated way where there's this active two-way telemetry and you know the state of health of your system

31:43Arushi Sharma Frank: versus the grid in real time. You might be ready to knock off everything slash put your diesels on if you're in that situation. If you've chosen one of these, bridge to grid, be flexible until firm options. Then when the signal comes in within the first 10 minutes that something bad is going to happen, and that bad thing might not be that bad. It might be you need to knock off 10 megawatts. You'll either have already prearranged because you chose to participate in the solution. You've prearranged to hedge that risk, and those 10 megawatts are going to get shaved off. Or you're going to call your friend who is the contracted battery or gas wind and say, in the next 10 minutes through our common QSE, can you provide some relief to ERCOT so that my load can keep consuming?

32:28And either of those two things can happen, which is really cool. I think what you're seeing happens there is that data centers as a monolith actually have a lot of diversity inside them. Size, shape, end-use customer inference, training, neither of the above. Maybe they're just like boring old electronic processing units. They don't need to have any AI component to them. Our industry is so vast. I can't imagine a world where in a couple of years from now, we don't have like a forest of solutions that have been designed to enable this non-firm access to the grid. And that includes conventional co-location.

33:08There are going to be some data centers out there that are off-grid gas. They want a bridge to the grid. And their bridge to the grid is 10 megawatts of firm power from the grid. And the rest of it, they have to rely on natively from their private gas system. At the end of the day, the more resources you have that participate in any sort of bridge to the grid fashion, the more opportunity the grid has to manage the system in a normal operations mindset instead of always thinking about the grid as, oh, no, the next 10 minutes are going to be a bleep show and I need to go call everyone and their brother to turn off my system.

33:46I don't think anybody wants that. And if there is anybody who's at the center of the reckoning of convincing data centers around what load flexibility really means and what you're not going to have to do and why it's so much better for you to participate there, it's absolutely me. I get all the phone calls.

34:07Arushi Sharma Frank: let's add one more dynamic to that equation which is bankability if i am trying to finance one of these large load projects how do i underwrite the fact that i can't fully determine what it means that a data center might have to step back from access to grid power and some ppa they've struck and find alternate sources of power at some undetermined frequency over the course of a year Oh, yeah, 100%. So after I passed the speed to power rules that came out in June, one of my startups worked with Aurora Energy Research to do a study that was funded by Google. So he also worked with Google directly to come up with a bankability strategy for all of this to work.

34:53That strategy is called telemetry offsetting. So now I'm really glad we've explained to your audience what telemetry is. I think that telemetry offsetting is a holy grail solution for the grid access crisis for N. Here's how it works. If you are a data center that is telling the grid operator that I have 10 megawatts of firm power from you, which is not remotely enough, I really need at least 50 more. The location you selected is great for 10 hours of the year. That location is giving you exactly 10 megawatts, and the rest of it is kind of fuzzy. The whole point of telemetry offsetting is to allow a paired battery or generator, ideally both, so we have smaller gas units instead of ginormous one-to-one facilities, to show up in the telemetry engine that the root operator sees to provide a dynamic support to all of the unknown flexibility that could be a curtailment risk of the data center and solve the problem inside the engine of the grid operator because the grid operator sees that, oh, I can always call on these resources before I have to call on the flexible load.

36:00That offsetting is actually part of a new rule that I'm helping write that we're advocating to get passed called NPRR 1348. NPRR is Notal Protocol Remission Request. That's the rulebook of the grid we keep talking about. 1, 3, 4, 8 is the number. And what that will do if we get it is when, first of all, the long tail of that rule when we get it will help base power. Because base wants to provide the same dynamic offsetting value to data centers as single-side batteries would. So that there's less problem on the grid in general and lows confirm up their power needs.

36:33Arushi Sharma Frank: It can create bankability. Back to just finishing the point is that for data centers to participate in a bankability construct where they say, oh no, in five years, I'm only getting a fraction of the power I thought I was going to get from the grid and any other solutions. For them to get these other solutions instead of off-grid, on-grid, they could participate in this construct where there is some unknown ballooning risk that risk can get tapered down by simply them participating as a PCLR. That's flexible. Basically, without this, it sounds like they're just taking on totally unhedged power risk.

37:09Totally unhedged, yes. In a way, the Aurora report actually tells you and shows you by looking at three of the largest campuses in Texas, the names of which are blinded, but you know who they are, biggest developments, that simply adding a battery next to a load that has registered as a voluntary PCLR reduces curtailment exposure by 50 % in some locations. If you then are able to do this telemetry offsetting, that risk goes down by another 80%. So actually, I would revise the statement or nuance your statement to say that the reason that some data centers would have unhedged exposure to curtailment is because the location they chose is not great for being flexible.

37:55If you put your data center in a site where there is a lot of redundant power flow, i.e., you put your semi-truck on a highway that is built for arbitrage, your truck has more lanes to park in. but also you have more options to have somebody else park for you if you decide to go put your data center on a one lane farm road in arkansas it's one direction there's not a lot of wiggle room there's no place to put your truck unless you stick it on a cow farm then you're not in the position to take advantage of this opportunity and that's what our report very very clearly lays out if you look at the report like you can see like there's a campus in the panhandle versus a campus in West Texas versus one in a highly populated dense transmission area.

38:38And their natural curtailment exposure is wildly different based on where they simply chose to locate.

38:45Arushi Sharma Frank: Looking the opposite direction from bankability, do these changes help large battery projects become more bankable? Does it smooth out the expected power that they will be managing in any way? So large batteries in Texas are surviving on the basis of accessing ancillary services or grid services prices. Funnily enough, my work on passing the Megapack rule change, which is what exploded in the battery market in Texas, also led to the general appetite and enthusiasm to put batteries as nodal arbitrage for grid services on the system. And they've kind of saturated their market. The takeaway I have on large batteries is they're a victim of their own success to some extent.

39:31Arushi Sharma Frank: Because they've smoothed the load out, they no longer can trade power at extreme margin. And now that we have NPR 1348, again, this is telemetry offset in the queue, I am talking to long duration battery providers every week because the proposition for putting batteries on the grid now is not that they are participating in smoothing out load on the system on the grid. It is that their duration, if it is bankable, it is solving wires congestion as a non-wires asset. Your battery participates in a market that's just doing energy-related stuff. They are participating in smoothing the grid as a whole.

40:14If your battery is participating at a location to solve wires congestion so that a nearby load can continue to consume, it's behaving like it is a wire. They call that non-wires alternative, and that's a different treatment and a different revenue model entirely for batteries. So yes, if we get 1348 through, or any version of it, we have created a brand new revenue path for energy storage to be located at data center sites. And to do so, not because they're smoothing the whole system peak, but because they are dynamically managing wires congestion risk and they're parking instead of the semi-truck parking on the side of the road.

40:56Arushi Sharma Frank: Which is also sort of the role that UPSs are playing in data centers, right? Which are basically these large batteries that act as a firewall connection to the data center. It is the power entry point into the data center becomes through this battery system as opposed to batteries backing up at the end of the chain. Not only that, you know, what you just described, this word firewall, is how I talk about the telemetry offset rule, because it's a shield. It's a shield for the data center to do more of what it wants to do, letting the battery take exposure. So the value to the load of the battery shifts dramatically from what we thought batteries were in Texas to what they will be in the future.

41:37In that sense, they're very aligned.

41:39Arushi Sharma Frank: On that note, how do the economics of distributed batteries play out, are they also similarly a victim of their own success? Or does it matter less because each individual battery is such a small contributor to the grid? The current aggregated DER program is going through modifications so that batteries can be notably priced and base power has led that work. That allows the grid to see the electrical similarity of what batteries can do relative to where there are problems on their transmission grid much better. So no, distributed batteries actually haven't been a victim of their own success. They are a victim of deregulation as much as they are benefited by deregulation.

42:24Arushi Sharma Frank: How so? This thing that I said earlier about how batteries can be participating in the market to smooth the whole system versus participating as a non-wire solution means that they prosper in markets where both applications are equally possible. In Texas, transmission ownership, and therefore the things that are not transmission, are owned by utilities, but a deregulated market means that the market revenues are owned by independent competitive players. So today, distributed batteries get to operate as market assets, earning market revenues that are not tied to their value as non-transmission solutions.

43:04By taking batteries nodal, we get to do what in a different area, let's say Portland General Electric is a good example. In Portland General Electric, a calibrated battery is being paid today as a non-transmission solution. And the utility has the governing control because they're not a deregulated market to say, we want to operate your battery under our telemetry network as our non-wires solution, which is freeing up capacity to the data center. Today in Texas, you can't do that because the utilities cannot own and operate batteries and they cannot pay you to operate your battery as a wire. So that's it.

43:39So DERs, they benefit from all of the open access of deregulated markets so they can provide market services. But they don't benefit from being able to contract with utilities wires company to provide the wire services. Not all batteries for me, this change from aggregated DESR is very exciting because it lets us achieve some of that value unlock of what batteries do as a wires mitigation solution. And you are compensated for that, or you're priced in ERCOT to be able to reflect that value for the first time.

44:14Arushi Sharma Frank: That's the idea that not only do batteries move electricity through time, they also move it through space. Through space. If you can view it as a transmission solution, it is a space mover as well. It strikes me that we've spent however long, 45 minutes talking about all of these changes that have gone into place largely in Texas over the last, I don't know, three years, four years. It's a substantial amount of change. I'm assuming everything we talked about is going to be vastly out of date three or four years from now, just given how fast everything is moving. What are the next big hurdles to jump?

44:54First thing I want to say for the audience is that if you care about non-firm transmission access for data centers, meaning you want data centers to be great assets, you want them to be good, flexible, you want them to participate. You don't want a future where we have 75 gigs of behind meter natural gas. You need to care about how much Texas has done in five years and understand the extent to which this work is not irrelevant in five years. It is a foundation for the rest of the country moving in the right direction. And when I say the right direction, I mean finally aligning things like, can we build transmission?

45:34If we don't build transmission, can we pay data centers, batteries, smaller gas plants, modular SMRs, everything, to behave as a hedge for us not building more transmission? Can you pay it? Right now, whether you look at the national regulator, FERC, the Federal Energy Regulatory Commission, or you look at the Bipreciative Bipartisan Act, Bajo, in Congress, or you look at the individual speech-to-power proceedings that are happening in different markets around the country, all of them center around the whole ecosystem, struggling to figure out how to value and then engineer solutions for non-firm access to power bits.

46:17at every level. At the distribution grid level, when it comes to distribution batteries, to the large grid level, when it comes to larger batteries, aggregation of batteries, and flexible data centers. So no, in five years, Cody, I think the work I've done in Texas is something I'll be teaching a lot of grid operators in other markets from the perspective of, well, I'm doing it now already, actually, but from the perspective of here are the building blocks for how you create a meaningful price for all of these solutions that help the power grid and help you achieve commercial success. And it's only going to get wider instead of narrower in terms of how applicable the story of Texas is to the rest of the country.

46:59Arushi Sharma Frank: Your hypothesis is that Texas is a blueprint for how load flexibility will play out around the rest of the country. Are there any competing systems that you think, if I found an Arushi living in one of these other jurisdictions, they would say, no, this is the way. One problem we have that prevents a lot of this from manifesting is that too many projects that utilities want to build for transmission or distribution upgrades are connected to a vision of data centers and other loads simply accepting firm transmission service or firm power service. If that's the only model you live in and you have a view that this is the best way to run the system, then really what the issue is, it's not that someone is going to have a competing business model, but a competing outcome wins because the data centers don't get built on time.

47:58Everybody builds upgrades anyway. It's something else shows up and just takes up power. Maybe it doesn't, and then you have a ton of straining cost.

48:05Arushi Sharma Frank: That's just the utility playbook of let's just go build more CapEx. 100 it is the capex playbook totally there's another thing that i think is an excellent counter thesis to connecting to the grid flexibly which is connecting to the grid non-flexibly but doing so at more locations where there's existing room at substations there's a choice the choice is connect to the grid with a little bit of room for wiggle and privately manage your exposure to degree. Don't go for a new hood connection, but go to every warehouse in the United States that's underutilizing their existing power connection. Or go to places that have retired power plant projects that never got built and use that same substation and convert it to a load connection.

48:53I would be very happy and very excited, and I would share enthusiasm with Michael Thomas at CleanView if that version of the future also played out because it is equally good as what I'm working at on inverting the problem of building 75 gigs of gas off-grid. They're both great. Love it. I want lots of competition for distributing our compute load out so that we can more efficiently use the interconnects we already have. It's utilities like the CapEx. Somebody already paid for it to exist. Somebody built the highway a little bit bigger as a result. Use the whole highway.

49:30Arushi Sharma Frank: Arishi, last question for you. You've already shared a bunch about the work you do with grid operators to help invent the rules of the future that can enable these various new technologies that are coming online to participate in the most efficient way as possible. You also do a lot of work with startups to help them have a voice in these proceedings and also help them think about how they should be designing their platform to plug into the rules that exist today and the rules that might exist in the future. Do you want to share a little bit about the work you do today. We have a lot of startups who listen.

50:06Arushi Sharma Frank: Just to give you a minute to talk your book for a minute. You already know that my first startup, BasePower, benefits from work I did to change market rules and create a greater moat for social grid prosperity in Texas. Then I realized that I can do this for a lot of startups. A lot of startups come in with a wonderful idea that can be valued by VC, but it can't be valued as a social solution because no one's done the work of creating a social opportunity for you. So that's what I do. I work with startups where I know that I can simultaneously create the actual programs where their success will exist, and I can tie the ultimate mission of their work into a broader infrastructure prosperity thesis.

50:49So after Base Power, the next startup I onboarded was Agenda Infrastructure, which is the development mastermind behind PCLR in Texas. If you look at the ERCOT record, Agentix work is everywhere. Their slide decks are everywhere. Agentix co-founder Sam Brandon is actually my colleague for when I was developing the battery storage project, Gambit Energy at Tesla, and he worked at Best Power. Both of my first founders have incredible drive to create a product for people. And they also need support and they need internal and external cohesion around how do you make the market wider and deeper for the abundance outcome.

51:29So startup number two, Agentec, already has PCLR registrations simultaneously while they developed the thesis for how we are going to make non-firm power actually bankable for data centers. Then the third one that came along along a similar vein was Emerald AI. I've got their sweatshirt on today. Varun's going to love that.

51:49Arushi Sharma Frank: Varun was on the show fairly recently. He was on the show a few months ago. Yeah, I need to get him to wear a luminary sweatshirt when he's on another podcast now. I met Baroon in a similar timeframe when I was busy writing speed to power rules that would help my startups. The compute flexibility thesis and sort of the integrated flexibility thesis that Emerald AI and then NVIDIA behind it bring to the table. You can see how it kind of rounds up the picture, because if I solve bankability for non-firm access to the grid, then suddenly how you firm that access how you resolve contingencies is the entirety of the business model all of these companies they're going to use large batteries to resolve some of it they're going to use native low flexibility at the computer to resolve the rest of it then in the middle there's going to be ups doing the it ot etc everything fits honestly in my startup wheel as one new way of making electron arbitrage or reality for these businesses to succeed.

52:47And while they do that, every time I go into a new market and write a new rule, write a new program, I also create a way for more people to participate in a bigger commercial revenue stream that just literally did not exist. Operating Luminary this way allows me to be a category creator on behalf of my startups. It's not just startups. In the past few years, I've also written and enabled the virtual power print program in the city of New Orleans. I'm also helping create microgrid law in Ukraine. I have an infrastructure advisory project in Australia where I'm helping them figure out how to monetize distributed capacity for resilience, so for hospitals.

53:25This work, it's like startups being a part of it because I'm such a power center for changing and creating commercial value in these very large complicated grid code type areas. It allows startups to thrive in the thing they really have to be head down on, which is how do they make their first recurring revenue dollar to tie their universe into the picture that they're already really a part of. It's like when you start one new regenerative farm, you're doing that work by yourself and you're like, how does my work here fit into the big picture for how America grew all of its food? My job is to be the leakage for them.

54:02We all need that kind of food. One farmer in one county is not going to feel great about how they are fixing the whole food picture. It's just too hard. And I think it's a completely untapped space. And I think if I can galvanize how America works in the next 10 years, honestly, Cody, is to get more people like me who have been seasoned operators and rule makers to come in and help more startups faster because they are brilliant people with their mission and their heart in the right place. They just need the platform to go deeper, harder.

54:34Arushi Sharma Frank: Hey, it's Cody. I'm jumping in to interrupt the flow of this conversation a little bit to highlight what I mentioned in the introduction, which is that Arushi and I recorded this conversation in October, which was before the November midterm elections, which had substantial implications for the work that is happening, particularly in ERCOT in Texas, but honestly all across the country. And so I wanted to have Arushi come back after the election and reflect on the outcome and what her takeaways are with respect to the work she does. The remainder of this conversation is from a conversation she and I had in mid-November, a couple weeks post-election.

55:21Arushi Sharma Frank: We are now starting a second part of this conversation that Arushi was gracious enough to come back and join me for after the midterm elections. So Arushi, we're now about a week and a half post midterms. And it seems like you've been on a nonstop media circuit since that happened. The work that you do is obviously a big storyline in particularly what happened in Texas, but honestly, what happened in many parts of the country. What's going on? So Cody, the election outcomes are an incredible tornado. It's not a wave. It's a full tornado. And I think that some of the upsets come with glaring realities that involve a lot of skilled people needing to deliver on the promises of people who have gotten elected on campaign statements.

56:17There's a saying, Cody, about how people who are incredible at campaigning are not so great at governing and people who are incredible at governing can't run to campaign themselves to save their lives and they need lots of people to help them through that portion. You're smiling for the for listeners. Cody is smiling. I wrote a piece while I was in this media circuit over the past couple of weeks called The Desert of Dried Up Hope in Tech and AI. It's on my sub stack. And in it, I basically distinguished the two elements of at least the

56:49Arushi Sharma Frank: wedge we believe that the data center AI issue was in that outcome. And it is definitely a wedge because you can talk to anybody. You can talk to an individual voter. You can look at exit polls. You can talk to campaign offices. The affordability, Iran, general dissatisfaction with the state of government is the dominating theme. And the wedge is things like lack of trust, lack of trust in what's coming for America with things like technology generally and AI specifically. So in my piece, I broke them down as very different because that's what I see on the ground. It's what I'm seeing in the data as well.

57:25And I've done a lot of this data analysis. I've been on shows like Bloomberg Tech in the past couple of weeks, talking to the tech industry or the investment industry behind it in the terms of the language that I speak every day around how you're building a thesis for data centers and tech broadly as assets for people in their lives. Like you have to be there because if you're not there, they disassociate you from their individual likelihood of success, like their individual path to prosperity, their individual desire to feel safe. And I mean safe from the perspective of my health care is safe.

58:00My kids' tuition is safe. The capacity for me to heat my home in the winter and buy medicine is safe. So there's a unsafety in the disassociation that's captured in these election outcomes. I did like deeper analysis on the races too. In Ohio, you have Ramaswamy and now the current sitting elected who is blue. You've got the blue and the red parties, those candidates campaigning equally against data centers as the code word. But underneath that, there's deep-seated mistrust around data centers and AI individually. The data center piece is very resource intensive. It's like, this is a resource we don't understand, we don't want, is taking our things.

58:40And the AI piece is bigger than that. And the AI piece impacts something like personal autonomy, potential for job loss, current impacts of the economy, and also how we're messaging the technology. Like we're telling people that the technology is going to take over, which is not what humans like to hear.

58:56Arushi Sharma Frank: I actually agree a lot. I mean, there's the affordability angle with respect to data centers and are my power bills going up, yes or no. And I think that answer is different in different parts of the country. But I think the bigger issue for a lot of people is just mistrust of the tech industry, people feeling like there's this oligopoly that's making decisions on their lives And it feels like a lot of what happened a week and a half ago is a referendum on that more than on any of the individual details. But going from that theoretical, sociological analysis of what happened down to the actual work that you do, particularly focused on ERCOT and Texas, the governor elect ran on pausing data centers until they pay for themselves.

59:39Arushi Sharma Frank: everything we talked about in our conversation two weeks ago was about the work you're doing to create flexibility and have data centers give back to the grid, etc. Do you think that the fast lane and everything that you've been working on answer what voters asked for? Or do you think a lot of that work gets swept up in the pause that Inosa is now advocating for? So I think that everything that has gotten less attention in this space is going to get 10x attention now because the first phase of our work in ERCOT was simply letting large loads, all kinds of loads, connect and manage. Phase two is what we've been executing.

1:00:21I've had a crazy couple of weeks, Cody, like simultaneously been executing on the next phase of these reforms, which is part of the grand plan to bring more to the grid and to the people who use it. So phase two of the grand plan is called load sponsor generation. It means that if loads come onto the system, they can sponsor additional assets on the system anywhere where they would relieve problems and receive some credit in their own local study per grid connection and then socialize the rest of that deliverability or that headroom, if you will, you're creating on the grid to the rest of the power grid.

1:00:57and in the process mind you bring down rates because you bring down the rates of building you lines by adding more private capital to paying for the cost of connections so phase two we started executing have another article about this actually ERCOT had last week this survey where there were 21 potential enhancement ideas for bring your own and load flexibility and in that there's really two themes that are dead center at the heart of where we are going to be going even more so after this midterm outcome which is increase firm power deliverability with private investment so make it easier for more people to enjoy cheaper firm power and two make it easier for loads that do come into this framework to manage operationally with all the firm power they're bringing, like cover their load curtailment risk with more new steel in the ground.

1:01:56Pablo Vegas, who's ERCOT CEO, and Alan Nye, who is the CEO of Encore in Texas, the largest utility system. They were with me yesterday at a Bloomberg event in Houston. And during their session, Alan said something that I think is critical, and Pablo did too. So Alan said that In winter storm Yuri, during that storm, my industrial customers were very gracious because we were making a lot of phone calls to tell people to start getting off the system because you want to solve as much of the problem before you get into a true emergency condition that causes the utility to have to start rolling feeders that have small businesses and homes on them.

1:02:35And he said on stage, I will take every ounce of flexibility I can get, which is a powerful statement from a WIRES utility that has over 25 % of the queue of data centers in Texas. And on the other side, Pablo, similarly, was talking about how he would like data centers to volunteer their way into pulling off the system before you get into a level two or a level three condition. And the earlier they pull off, the better. Now, if you think about that for a second with what I've been building in Texas, the PCLR framework and then adding new generation and new batteries to the system to even manage those assets to keep them less likely to get curtailed.

1:03:17All of that is putting more regular operating headroom and dispatchability slash flexibility into the day-to-day operations of ERCOT. So that means like anytime there's a problem on the system, the grid operator can solve the problem with more tools and they don't have to go towards a emergency condition. They don't have to go to high alert. We don't have to experience price spikes that are tied to scarcity on the system. All of this is so pro-cyclical. It is so important to put more of these incentives inside the marketplace, because if we don't, people like Pablo and Alan have to get on stage and basically hint that you guys are going first if we have a problem because residents are not going first.

1:04:00Arushi Sharma Frank: Yeah. If it's a flexible asset, it can and should be grid positive. If a data center is not a flexible asset, it is, in theory, a liability in one of two ways, right? If it's all behind the meter power, it's a liability in that it's built up a bunch of power that the public doesn't benefit from. And if it's a grid connected, non-flexible asset, then it is impacting citizens' access to power and potentially cost of power. From a simplistic perspective, That's how I think about it. Yeah. And also, there's a liability for the data center industry. Like, look, if you're a large facility of any kind and you don't know how often, when, because you're not in control of your fate in any way because you did not integrate into the ERCOT system and you're kind of like waiting for the emergency phone call, you have no understanding of retail risk.

1:04:48I said this a year and a half ago when I had, frankly, more of the data center industry quietly. Very few people called me to say, we don't like your ideas. but they certainly go around telling other leadership and I hear about it. We don't like the idea that she's pushing that data center should be part of normal grid operations in this way. We don't want to be controlled. We don't want fiat control over us. And I said, well, you're going to have that anyway if you don't do this, because if you're not in the normal market, and by the way, in the normal market, you can hedge your flexibility exposure with batteries and generators.

1:05:19You can mitigate that risk. You're going to be in the world of just the twilight zone where you'll never know when ERCOT has to call you for for how long.

1:05:29Arushi Sharma Frank: Getting tactical for people in the queue who are developers right now, I think the audit of the existing data center universe in ERCOT wraps up under the outgoing governor next month, and then the new administration and the new legislative session start in January. So if you're in the queue right now, What should you be doing over the next 90 days? For starters, the likelihood is that some of the items in the survey, there's a lot of these different sections. Those items will become meaningful pieces of legislative proposals. And if you can be a part of shaping that conversation and make sure there's consistency, you should.

1:06:14because it would be very bad for us to have created one set of expectations that are based on the outgoing administration's desires for lifting up the industry and the incoming one. So you do need effective, clear transition and participating in that effectiveness means being vocal. I think that in the past, I've seen the data center community in Texas as part of the large load community more broadly participate in this policymaking process in a wedge basis, because there's a lot of existential fear around exposing yourself in a legislative meeting or going to a hearing or whatever it is. But I think that in a transition phase, every legislative session is an opportunity to create and maintain clarity.

1:07:01And that includes like for communities themselves because to the extent that you do have ongoing negotiations that you believe are private right now with community partners, the fact that you're having them, it needs to be exposed in the policymaking forum. Otherwise, what comes out from the top is not what's already working for the bottom of the chain. That's my take always with these transitions is there needs to be more synergy between what's really happening on the ground and what's being asked of you from policymakers. This might be a reason for you to buy me a beer, but I think, Cody, a lot of your companies are going to have a second wave of interest from their potential customers.

1:07:41I don't think it's specific to one kind of startup. I think that broadly startups in the space who have been in the business of trying to make the transition into a tech load is the largest load on our power grids economy, they have too many potential customers right now. And not any one of those customers is the most exciting or the best customer yet. But I think we're going to sort that out because it's all going to come down to utilities, grid operators, data center providers, builders and operators. And every one of those sectors has to lift itself up to be able to meet the mandates coming out of these governor's offices, at minimum, not to mention Congress.

1:08:23Arushi Sharma Frank: Personally, I wholly agree with you. It's certainly self-serving of me to say so, but I feel like the results over the last few weeks are bad for pure behind the meter natural gas growth, which from a climate decarbonization perspective is a good thing. and are good for technologies that enable flexibility, that enable smart utilization of resources, that enable storage of power and reuse of power over time and space and all the related things. And that's generally the kinds of things we've been investing in. So, yes, I feel positive about that from my perspective. And hopefully it's good for the world as well.

1:09:02Yeah, there's two ways to look at this. I think all of the companies that are on your platform in this space and this work that we're doing in Texas, two things are going to happen. One is that we optimize every gas molecule for a better outcome. We're still going to use gas, but we optimize its utilization. And we optimize avoided climate impacts. And I haven't even done a quant analysis yet, but if some of this low-sponsored generation and PCLR work goes through in Texas, I don't know, I think me and my team and my allies will be responsible for abating half of future emissions from off-grid gas.

1:09:36Arushi Sharma Frank: Wow, amazing. That's real impact. That's amazing. Yeah. And I think your companies and our partners in this space, everybody just needs to pay a lot more attention because this is no longer going to be noisy. Every investment dollar needs to be high signal for what we're doing for communities. Awesome. Thanks, Arushi. Thank you. Take care, Cody. Inevitable is an Override Ventures podcast. We back founders who override the default across energy, infrastructure and industrials. If this episode was useful, send it to a friend who funds or builds technologies like these. Find past episodes and our newsletter at inevitable.fm.

1:10:15Arushi Sharma Frank: See you next week.

From the publisher

Arushi Sharma Frank is the founder of Luminary Strategies, where she helps startups navigate and reshape the rules governing how new technologies connect to the power grid and get paid. Previously at Tesla, she helped open the Texas grid to large-scale battery storage and created the framework that allows residential batteries to participate in the grid as virtual power plants. Today, her work is helping companies like Base Power, Agentic Infrastructure, and Emerald AI unlock new opportunities in energy storage, grid flexibility, and data center infrastructure.

In this episode of Inevitable, Frank explains why understanding the power grid requires looking beyond physical infrastructure to the markets, rules, and financial systems that determine how electricity moves and gets valued. She walks through how she helped change ERCOT's rules to accommodate Tesla's Megapack batteries and aggregate residential batteries into grid-scale resources.

Finally, Frank discusses why Texas could become a blueprint for the rest of the country, how the recent election could reshape the future of grid policy, and why the next generation of energy startups needs experienced operators and rulemakers to help create markets where their technologies can succeed.

Episode recorded on October 21, 2026 (Published on November 19, 2026).

In this episode, we cover: 

  • (0:00) Arushi’s path from natural gas and power markets to Tesla and Luminary Strategies
  • (4:02) Why the power grid is a rulebook and financial marketplace
  • (6:40) How batteries are changing the economics of the power grid
  • (10:28) How Arushi unlocked the Texas market for Tesla Megapacks
  • (17:12) How residential batteries aggregate into virtual power plants
  • (20:57) Metering vs. telemetry, and how grid operators verify energy flows
  • (24:00) How Texas’s speed-to-power rules (PCLR) let data centers connect faster
  • (30:39) What happens when the grid asks a data center to cut back
  • (34:35) Telemetry offsetting and making flexible data center power bankable
  • (38:45) New revenue models for large and distributed batteries
  • (44:54) Why Texas could be a blueprint for grid flexibility nationwide
  • (50:10) How Luminary helps startups create new markets
  • (54:34) Post-election update: AI mistrust, the next phase of Texas grid reform, and what data center developers should do now


 

Enjoyed this episode? Please leave us a review! Share feedback or suggest future topics and guests at [email protected]. 

Connect with Override:

*Editing and post-production work for this episode was provided by The Podcast Consultant