a16zAmerica's Energy Problem: We Need A New Grid
EVERY SPOKEN WORD
55 min read · 11,072 words- 0:00 – 1:09
The Future of the Energy Grid
- DUDavid Ulevitch
The energy grid and electrical grid of the future, it's not just gonna be the dichotomy of generation, transmission, and storage.
- RMRyan McEntush
This sort of next generation o-o-of sort of what the grid looks like is gonna be in a much more decentralized way.
- EPErin Price-Wright
Why are delivery costs such a big problem?
- DUDavid Ulevitch
The grid is aging now and brittle. The workforce has aged out.
- EPErin Price-Wright
Should we just leapfrog the grid? Like, I need this power now today.
- DUDavid Ulevitch
We have this insatiable thirst for energy.
- EPErin Price-Wright
How do you get software almost insidiously on the grid?
- RMRyan McEntush
The United States needs to get better at mega projects, things that are a billion dollars, things that are, are at scale.
- DUDavid Ulevitch
There is no safety, there is no national defense, there is no national security without a reliable electrical grid. [upbeat music]
- ETErik Torenberg
U.S. energy peaked in 1973 in terms of per capita usage. China's has increased ninefold over that same time period. Uh, we have some reasons to be o-optimistic now that things will- that things have started to change or will change further. Why don't you give some context there? W-what, what's happened and, and why should we be excited about what's coming forward? Maybe Ryan, why don't we
- 1:09 – 3:08
Historical Context and Current State of the Grid
- ETErik Torenberg
start with you?
- RMRyan McEntush
Yeah. I mean, the history of the grid in the United States was build big power plant, industry formed around it. Uh, the grid grew sort of incredibly fast through the 20th century. Then around sort of the '80s and '90s, things started slowing down. Uh, and then through sort of the early 2000s, the grid effectively froze in the United States. A big piece of that was a lot of the sort of energy pro- generation, a lot of the manufacturing, a lot of sort of the heavy, heavy industry moved to Asia. Uh, and so for the last, like, 20 years, uh, effectively the grid has sort of ossified. Uh, we forgot how to build new sort of power plants. We forgot how to build new power projects, uh, new loads, sort of large data centers, large factories, large mega projects.
- ETErik Torenberg
And when you say we forgot, were we not allowed to, or we, we actually just lost the know-how?
- RMRyan McEntush
We were allowed to, but we lost the skill set, and I think you can kinda see it in more extreme examples with nuclear power plants. A lot of that, that sort of transition happened decades before. Uh, but basically the grid itself, the grid operators forgot sort of how to, you know, plan, how, how to move quickly, how to do it cheaply. Uh, and so, you know, now we're at this point in time where we are reshoring, we are bringing back manufacturing, we are bringing back, you know, data centers, and there's sort of this highly concentrated demand and it's sort of now, now, now, uh, at sort of any price, but, like, they, they cannot move fast enough. And so that's what we're seeing today. We talk about sort of data centers and, and the grid being sort of inflexible to this. Uh, it's sort of playing catch up. We need to do a lot of the growth that happened in China and sort of bring this here and do it, uh, incredibly fast.
- ETErik Torenberg
H-how did this forgetting happen? And I ask to, to, to then ask the question of how can this re-relearning happen or this, the learning happen or kind of retraining happen?
- RMRyan McEntush
It's a good question. I think a lot of it's like a workforce issue. Um, I think a lot of it's a policy issue. Um, I think the, the United States, uh, you know, started, h-historically was a, a bunch of regulated utilities, sort of the top-down, big thermal power plant, big transmission lines connecting to substations, then distribution lines going to individual factories, homes, things like that. Um, and I think, uh, I think some of the newer technologies nes- don't necessarily benefit, uh, from scale in the same ways that these large thermal plants typically
- 3:08 – 7:31
Decentralization and New Technologies
- RMRyan McEntush
did. And so, you know, this, this sort of next generation o-o-of sort of what the grid looks like is gonna be in a much more decentralized way. So there's also an element of sort of relearning of what is a grid actually. I-is, is the grid, uh, is the grid these large sort of power systems and large infrastructure projects, or does it look like a lot more decentralized way where we can eliminate a lot of the wires in between, things like delivery costs, which have increased exponentially? Um, and can we do it in a sort of m- more dynamic and flexible way? So things like solar and batteries, uh, they don't need to be massive. You can kind of put them anywhere. You can put them next to load. Uh, and so, so this is also sort of a, an element that, like, grid operators are thinking through. How do we do that while also managing frequency, voltage, things like that without, you know, causing sort of the grid to, to go down? Uh, so there's a lot of challenges.
- EPErin Price-Wright
Well, if you think about what our grid is, I mean, it's a piece of technology that was designed about 100 years ago, and very little technology on the grid has changed in those 100 years. And, you know, you look at, you, uh, w-why are delivery costs such a big problem? You know, grid is at capacity. Getting a new project onto the grid today, y-y-you know, you sign up for interconnection, it could take a decade. Um, there's a backlog of 20-plus years to get a new transformer. The transformer technology we're using today, um, is kind of bananas, like, if you actually look at what makes a transformer. First of all, there's, like, one company that makes these, um, and there's, like, one plant in the US that produces the, the right type of steel that you need in order to make these transformers. And it's 100-year-old tech, and it's like the wait list for transformers is, is, is, is insane. We're getting to the point, um, in demand starting to rise again and the, the calcification of grid technology where it's like, should we just leapfrog the grid? Like, do we really need to wait in line and, and, you know, wait for this to catch up? I think there's like two, two versions of how you do that. One is, um, you know, how do you get power gen and power storage essentially as close to demand as possible? Um, and that's a problem for new tech to really help solve because we're talking about instead of these, like, mega projects that we're used to building, like... Well, not used to building anymore. Um-
- RMRyan McEntush
Mm.
- EPErin Price-Wright
... like massive nuclear plants, massive new natural gas plants, uh, et cetera. We're talking about much smaller, um, and m-more distributed sources of power, bypassing interconnection altogether in many cases. Uh, we're seeing that as a pretty big trend with data centers, um, as data centers are just building power directly on site and co-locating power with the data center. Microsoft's like, "I can't afford to wait 10 years to get an interconnection with the grid. Like, I need this power now today." So how do you get, um, you know, how do you get power more tightly coupled with the load that it's actually going to serve? Um, and that, that's a really interesting problem for tech also from a software perspective because, you know, if you get generation storage, uh, and usage all co-located very closely together, like, that's a very good problem for AI to solve, like reinforcement learning. Stick that in there, and suddenly you get massively efficient systems that you, you know, you couldn't get at grid scale.
- RMRyan McEntush
I think an interesting point to add on to that is that there, there's very vis- little visibility into the grid itself. Uh, so, like, they understand sort of power plants are operating or not, but, uh, e- especially at sort of the distribution level, uh, the things, the power lines you might see outside your home, um, there's very little understanding of what's actually going on there. And so, uh, there's, like, a reluctance, especially when you have things like net metering, where I'm sending from my home a battery back to the grid, uh, things get incredibly complicated. And so the grid operators don't have a very good understanding of, like, when, when can they allow new projects to go online, how much power, when to actually, you know, uh, uh, cut people off. And so there's a lot of these, these policies, interconnection is sort of the, the general term, but, like, states like Texas have, like, a much more lenient policy of, like, you can build wherever you want, um, but, you know, if, if we need to cut it off from the grid-
- EPErin Price-Wright
We're gonna do so. And so it's sort of this connect and manage approach. Uh, whereas other states, like they will do these incredibly long feasibility studies in like a variety of sort of scenarios. The entire grid is at peak capacity and, but they wanna make sure this specific project can stay online, you know, twenty-four/seven. So that ends up creating these massive delays. And so there's a lot of policy approaches here as well, and there's a bunch of these technologies called grid enhancing technologies, which are sep- effectively like, you know, an average power line might be used, used at like fifty percent capacity, but it needs to be designed for the peak capacity for, you know, the summer when everyone has their AC on. Um, and so there's a lot of, you know, sensors or other technology that could be placed there to have a much more dynamic view of what our infrastructure actually was looking like. And so when we have these new technologies, then we can much more efficiently use the infrastructure we have.
- ETErik Torenberg
D-do you-- what, what are your reactions to this conversation thus far? Is that-- so where are some areas you're, you're particularly excited about or reasons to be optimistic?
- 7:31 – 10:45
Policy and Workforce Issues
- DUDavid Ulevitch
Well, I think the reason we're having this conversation is we're sort of touching on a bunch of these topical, um, themes, which is that we're at a moment in time where exactly like Ryan said, the, the grid is aging now and brittle, the workforce has aged out. Um, I mean, we'll talk about that more I think in, in a minute. But, you know, we had to go out and train, hire and train entire specialized crews, specialized people that work with cement and concrete, specialized people that work with steel to go build the large Vogtle reactors, um, in Georgia. Um, we put them on Vogtle reactors three and four. We turned them on, huge win. Um, and then, and then those people like went back to like building highways or like bridges or something else. And like instead of just going and putting them on Vogtle like five, six, seven, eight, nine, ten and building just like this massive crescendo of nuclear power, we just like put these people back into the general workforce. Um, and so we just are, we are not learning our lesson there on the workforce. Um, and at the same time, we have this insatiable thirst for energy, whether it's EVs, um, whether it's da-data center compute for, for AI, um, or just generally a shift, uh, toward more and more consumption of, of electricity, um, or even just like the reshoring and manufacturing, all these things that are just very, very, um, electron, uh, heavy. Um, and I think at the, at, at, at the same time, a piece that we hardly ever talk about, which is sort of resiliency and not having people be as dependent on the interconnectedness of the grid. I mean, people that you deploy solar are, you know, we talk about like distributed compute for those of us that are in the tech world, and like how important it is to like have distributed compute and have networks be able to suffer a-and survive through, um, uh, segmentation of things. But like the grid is very interdependent. Um, even in the U.S., there's really only a few major regions that, that can segment themselves off. Um, but when you deploy solar or you deploy batteries or you deploy an SMR reactor or your own gen battery, you know, your own power generation on site for your own data center, you don't have to worry about how brittle the grid is because you're, you're fairly resilient, uh, from it. Uh, and I think that's a component is that the energy grid and electrical grid of the future, it's not just gonna be the dichotomy of generation, transmission, and storage, but as Erin brought up, like you might do all three of those things in the same place and not have to worry about how, how robust the grid is or how capable the, uh, the grid operators are. Um, and I think that's a dimension that was never important to people before, but it's important today. Uh, and you can certainly imagine if you're the military, you certainly care about having reliable, uh, access to power, um, at, at all your f-you know, forward operating bases and, and even at home at your, your home military base. It's like you just, you cannot lose your ability to, to have electricity. And so I think all these things are just coming together at once, and it's a really exciting moment in time, and I think it's buoyed by the fact that we're also at, um, this sort of technology inflection point where AI can help some of these things, um, not just be, uh, a consumption driver, but even be an enabler in facilitating-
- EPErin Price-Wright
Mm-hmm
- DUDavid Ulevitch
...um, more efficient use of electricity, better monitoring of the grid, um, be-better, better ways to even go through the regulatory and permitting process, um, which is, which is onerous for, for many cases. So...
- EPErin Price-Wright
Yeah, I think, uh, i-b-building on that, I
- 10:45 – 15:07
The Role of Solar and Batteries
- EPErin Price-Wright
think, I mean, Texas famously historically had massive grid failures several years ago when, you know, big heatwave came through, the grid ke-couldn't keep up with all the air conditioners that were going on, and people saw massive power outages. Everyone was really mad. People were like, "Oh, ERCOT doesn't work. Dereg doesn't work." And what has Texas done in the couple of years since that happened? They have absolutely flooded the grid with solar capacity. Texas has doubled their solar capacity in the last approximately three years, and with that, they've just deployed thousands of batteries. I mean, we-- o-one of our portfolio companies, Base Power, is, uh, y- one of the players here, but there's many bout-- battery power companies, uh, deploying all across Texas, uh, to provide storage for that solar power. And if you look at the performance of the Texas grid versus the performance of the N-New York and surrounding area grid during this ex- heatwave, you know, I, I must have seen t-ten news articles this morning about how well Texas grid has done. Their, the elasticity and ability to react to the change, to, to very quick changes in demand without having to change kind of base load power. Like, you can't build a new natural gas plant or a new nuclear reactor overnight, but, you know, solar is just so insanely cheap. Like, it's basically having a, a giant, massive, huge nuclear reactor in the sky that will go forever. Um, and, i-i-i-you know, Texas isn't a green state. This isn't, you know, this isn't a political issue, but it's like, why aren't we deploying the world's cheapest form of power literally everywhere we possibly can, and then just putting batteries everywhere? Like, there just should be batteries everywhere. It's kind of bananas to me that batteries as a topic has, like, recently gotten caught in the sort of political crosshairs. Like, we really just as a society need to be good at power storage and batteries. Like, this, this shouldn't be a controversial topic. We invented the lithium i-ion battery, and yet today, you know, if you wanna buy a battery, whether it's for a drone or for the grid, or for your car, or for whatever it is, like, you're either buying a battery made in a lights out factory in China, or you're buying a battery produced in Vietnam by a Chinese company, and, like, there's no meaningful effort in the U.S. to change that. This is a really critical problem, not just to managePower load on the grid, but for power for all of the things that we need, uh, to power the, you know, the next generation of innovation in the United States. Like, I think we'd be hard-pressed on the American Dynamism team to think of a company that we've met with an interesting technology in the last two years that doesn't have a battery in it somewhere.
- DUDavid Ulevitch
Sure.
- EPErin Price-Wright
Um, so as a country, we need to be investing in, in, in, uh, in, in battery technology and battery manufacturing.
- DUDavid Ulevitch
By the way, if China decides that, uh, whatever your company is doing that's using batteries doesn't align with, uh, what they like-
- EPErin Price-Wright
Good luck
- DUDavid Ulevitch
... or they wanna punish you, being cut off from being able to buy batteries from China is incredibly punitive, uh, to a company, and we've certainly seen that happen with some of our startups. Um, and then they have to go... You know, then you find out quickly that the, uh, ability to procure and source batteries from places that are not, uh, in China is very, very difficult. Um, and if you extrapolate that out to what would happen to our whole country if we just were unable to buy batteries from China, um, it could be catastrophic, uh, in a very, very, very short period of time.
- RMRyan McEntush
Just add on to a quick point on the, uh, the grid side of batteries. If, if the rest of the country, which is, uh, presumably watching what's going on in ERCOT, which is the grid operator in Texas, if, if Texas can prove that you can deploy these sort of decentralized, distributed energy resources and to sort of flatten these peaks, provide more resiliency and, and ultimately lower, lower price of electricity, uh, then every state should, should go and do this. Um, there's a very complex web of deregulated and regulated entities, uh, when it comes to the grid. Of course, there are a lot of different policy and, and workforce, uh, and sort of political reasons why, you know, not everywhere is sort of this decentralized world. Um, and it'll probably be more complex than just, like, sort of these deregulated energy-only markets that Texas wor- uh, works with. And I think the United States needs to move incredibly fast to, to make this happen and hook up batteries, solar panels, make it easier and cheaper to do it. Um, even co-location for, for large loads is still a very politically fraught issue. Utilities are pushing against this. It's still really hard to hook up solar and batteries to your home. I think it's actually cheaper to put residential solar on your home in Germany than the United States. Um, and that's largely a permitting, largely sort of an installation issue. Um, that's crazy. That should not be the case.
- 15:07 – 24:24
Diverse Energy Sources and Demand
- DUDavid Ulevitch
Uh, Erin, I believe the quote in your college yearbook was, "Drill, baby, drill."
- EPErin Price-Wright
[laughs]
- DUDavid Ulevitch
And so how do you, uh, think about your sort of love for oil and gas with, you know, with other sources of energy?
- EPErin Price-Wright
Oh, my parents will be sha-shaking their heads to hear this.
- DUDavid Ulevitch
[laughs]
- EPErin Price-Wright
I mean, like, okay, uh, uh, I think broadly speaking, our approach to energy in the US just needs to be yes and. I mean, you look at the fact that, you know, you look at the sort of atrophy of our, of our, uh, power build-out over the last thirty, fifty, whatever y- you know, you name your timeframe, years compared to, let's say, China, um, and if... That we want to accomplish the goals that we've set out as a society to accomplish over the next decade, like, we need more power, and it's just a matter, it's a matter of yes and. And I think, you know, solar and battery is extremely important, but, like, there is a place for oil and gas. Like, I cut my teeth at Palantir working in oil and gas. My husband worked in oil and gas. Like, the first check I wrote at a16z is in an oil and gas company. So this isn't me coming with a, you know, a particular agenda around carbon. This is me re- like, coming and realizing that, like, we basically need every tool in our toolkit, and, uh, we should be using technology to deploy whatever makes most sense, um, wherever it makes most sense, um, at scale.
- RMRyan McEntush
If we're talking, like, energy mix of, of sort of where we're at today and where do we, where do we think we're, we're headed, if I were to make, like, a personal, personal bet, it's, like, solar batteries are just the co- ability to, to be incredibly cheap and deploy incredibly fast. I think it-
- EPErin Price-Wright
Spin up and spin down.
- RMRyan McEntush
Yeah. And I, and I think that that's just going to sort of... It already is a way, but I think that will continue. But I think to be very clear is that you need all different types of energy. Um, but you're gonna need, you're gonna need true sort of base load dispatchable power. It's gonna be gas. It's gonna be nuclear. It's gonna be geothermal. It's gonna be a lot of hydro as well. As you attach more of these sort of, uh, renewable resources or, or sort of these non-reliable resources, while very incredibly cheap and works most of the time, sort of this long tail risk, you know, once you get to, like, 50 to 75% of the grid, uh, is gonna become very, very expensive. You need a lot more battery backup, things like that. And so I, I think it's gonna be very complex, and it's gonna be different for many different regions. But certainly it's not all, uh, all of any given resource.
- EPErin Price-Wright
Yeah, I mean, when you look at, like, the l- the, the, the changing nature of load over the next decade, some of that is gonna come from data centers. Um, some fraction, it's not, you know, I would say it's, it probably overstated how much data centers contribute to the growing load in the United States over the next decade. Um, data centers generally are base load. Like, if you're training a model, you're y- largely using kind of a dedicated amount of power for, for the f- l- for the long term. Maybe there's some fluctuations if you're doing more inference, but I would generally say, like, data centers represent base load. Uh, but then you also have things like electric vehicles. You have things like heat pumps and air conditioners. Um, you have, uh, industrial autonomy, uh, which may or may not be sort of running twenty-four/seven. So you're gonna have some increase in the base level of power we as a society need, but continuing to increase the size of the peaks and troughs of how we use energy on a day-to-day basis, and we should be thinking about, uh, designing our grid and designing our energy mix and power sources around what those loads look like and not over-solving for either base load or variable power.
- RMRyan McEntush
I, I think just to put this in more tangible terms, the sort of like the, the peak summer load in, in places like California might be half of what it is in, in, in, in winter or something like that, and it depends on what climate you're in. And so there, like, the concept of, of base load is like, do you build, uh, all the plants you'd need for, you know, like the 100 gigawatts of power you're gonna need when it's wintertime, or in the summer, you know, half the year you're only gonna need 50. So, like, what, what would be base load? So y- you, what you need to do in, in sort of modern civilization is every time you turn on the switch, like, the power is working. And so how you actually match supply with that very, very fluctuating load, both daily and seasonally, is very complex. And so, like, today y- you might have to build a natural gas what's called, like, a peaker plant that might only operate, like, a week a year.And so that's an incredibly expensive asset that is going to only be delivering very expensive power, but is only needed when all the other resources are tapped, and it's like that last, you know, couple megawatts of power. The alternative now you could do is what's called, like, sort of demand response or with batteries on the grid is say, "Okay, well, instead of, you know, doing this t- $10,000 or sort of a megawatt hour plant, I can just make it so everyone's thermostat in this area turns down a couple degrees." And so that, that in aggregate means that I don't need to build that, you know, that, that large asset, uh, or pay that expensive premium. And so there's a lot of things like that that are, are, are sort of newly available.
- EPErin Price-Wright
I do think, though, okay, pushing back on that, like, I think that the American consumer will fully organ reject that level of that dictation over how they use their power.
- RMRyan McEntush
I agree, I agree with the thermostat, but I think-
- DUDavid Ulevitch
I think a more like, a more likely outcome is that you can do it on the compute side and just say, "Look, these three racks of the data center are just gonna go offline during the peak summer heat when you're running your AC," uh, and maybe at night.
- EPErin Price-Wright
Like this is not a critical job.
- DUDavid Ulevitch
Right. It's a non-critical job. It's not a mission job. It's a back office job, you know, and you're just gonna run it at night instead of during the day, and you're gonna pay less electricity for that-
- RMRyan McEntush
Mm.
- DUDavid Ulevitch
... for that benefit. Um, and, and so I do think that as you, as we build out... I don't know if I agree with Aaron that AI is not gonna suck up all the compute. I mean, I think that Constellation just turned up a new nuclear reactor or is reactivating a reactor, and then I think Meta immediately sucked up all of the power that they're gonna generate or, or 9/10 of it or something, um, from the, from the new Constellation reactor that they're gonna, that, that, that Meta signed the, the, uh, contract extension for. And so I d- I do think that compute is gonna just... I think we actually probably are underestimating the amount of compute that we're gonna soak up, um, o- over the n- with electricity over the next, you know, 10, 20, 30, 40, 50 years. Um, the amount of data we're gonna start storing, I mean, the amount of... Like, just look at video. The amount of video we create per minute, um, has just ballooned way beyond anyone's expectations. I'm sure the same will be true for, for AI compute. Um-
- EPErin Price-Wright
And I think once you start getting into, like, ro- ro- robotics and autonomy, um, a- if you think about compute expansively, I totally agree.
- DUDavid Ulevitch
Yeah, but, and so, like, but those things are gonna be much more responsive than, like, do I wanna go-
- RMRyan McEntush
Mm
- DUDavid Ulevitch
... have my room be, you know, 74 degrees instead of 71 degrees?
- RMRyan McEntush
Yeah.
- DUDavid Ulevitch
Well, let me tell you. Anyone that's done business in, in Tokyo in the summer knows, they, they... A- as a nation, by the way, Japan has done this. It is absolutely terrible.
- EPErin Price-Wright
It's horrible.
- DUDavid Ulevitch
Um-
- EPErin Price-Wright
We're not gonna do that in America, please.
- RMRyan McEntush
We're not, we're not gonna do-
- DUDavid Ulevitch
We're not gonna do the thermostat.
- RMRyan McEntush
We're not gonna do the thermostat idea. But I will say, I will say, like-
- 24:24 – 25:55
Challenges in Grid Monitoring and Communication
- DUDavid Ulevitch
response, uh... Like this went to the point that Ryan brought up, that monitoring the grid is really important, and being able to send signaling on the grid is really important. And you have to remember, like, you know, we're, we're all used to the internet, which has, you know, like, bidirectional communication and messaging and has-You know, d-d-data layer and control layers, and there's like a full control plane and, and things like that for the Internet. The, the electric, the electric grid doesn't really have that, and so to be able to send messaging and things is very, very difficult. Y- and now a lot of people just do it out of band using the Internet. Um, to actually send messaging and do monitoring of the grid itself wi-without an overlay network is very hard. Um, and that's one of the challenges that people are now, I think, starting to address.
- EPErin Price-Wright
Yeah. It's kind of... I mean, it's, it's, it's, it's wild how much of a mystery what's happening on the grid is at any given time. Like, we really have very little visibility, and it's very hard-
- DUDavid Ulevitch
Mm-hmm
- EPErin Price-Wright
...for, I think, centralized utilities to deploy meaningful software to understand that. So when, you know, when we think about as VCs, like, what types of things do we look at and what do we get excited about, I think companies that kind of are going at this monitoring from the opposite direction, like how do you get software almost insidiously on the grid? Like, how do you start learning more about demand, um, demand and generation as close to the source as possible, and then try to feed that information back from each other. Like the idea-
- DUDavid Ulevitch
Mm-hmm
- EPErin Price-Wright
...that you're gonna go sell a software tool to a PG&E or similar and have like a, a reasonably speedy top-down-
- DUDavid Ulevitch
Mm-hmm
- EPErin Price-Wright
...implementation where you actually get good signal and metrics and can actually do interesting things with that data, to me, I find like a little bit unbelievable.
- 25:55 – 27:40
Load Forecasting and Energy Markets
- RMRyan McEntush
So something, something very interesting that I learned is, is a lot of the load forecasting, which is basically like the tasking of wind plants, uh, uh, need to go online. So like there's usually a twenty-four-hour ahead sort of market, day-ahead market that'll basically say, "You need to run your, your plant at this time," and then this sort of supply and demand match to a price, and there's like a merit order and it's, it's complex, but, but that's sort of how it's done. Um, but, uh, uh, like most of this, this predict load forecast is done by just looking at the weather. Like, they look at, like, basically, like the-- one of the best indexes, they look at where, where the homes are, like, what, what, how many people are there, and then what the temperature is gonna be. And sort of that often is sort of the lo- largest factor that goes into this modeling. But then if we have all these sort of connected resources, if we have solar, we have like EV chargers, all of this stuff is spinning off data, telemetry, and things like that, we're gonna get a much better look of how like load is actually being forecasted, which is gonna real time, which is gonna be help a lot of u-understanding, like, you know, where do we actually need to build? Like, what is the actual price of power? And then you can start making these markets, I think, a lot more efficient.
- EPErin Price-Wright
Well, when you look at energy desks, like, uh, for, you know, the big hedge funds or energy-
- RMRyan McEntush
Right
- EPErin Price-Wright
...energy trading companies, their weather guy is usually the highest paid person on the desk-
- RMRyan McEntush
Yeah
- EPErin Price-Wright
...for so outside the portfolio manager.
- RMRyan McEntush
Yep.
- EPErin Price-Wright
Like those, like those w- climate and weather PhDs that are working on a trading desk, they are just absolutely-
- RMRyan McEntush
Mm
- EPErin Price-Wright
...raking it in, uh, because they're kind of like God right now-
- RMRyan McEntush
Yeah
- EPErin Price-Wright
...because there's very little other data.
- RMRyan McEntush
That's why you see when the stuff that goes on in Texas, like heat waves and things like that, if they even get it wrong by a couple of degrees where it's like it's, it's either, you know, it's like they think it's gonna be sort of hot, but it gets actually really hot, that's when you get these crises, like crises that, that, that end up causing a lot of strain on the grid, and then you have to turn on the, all, all these very expensive plants, then, you know, you get the headlines.
- EPErin Price-Wright
They're usually also the ones that are worse for the, the w- the mo- the worst for the environment as well.
- RMRyan McEntush
Yep.
- DUDavid Ulevitch
Do you-- L-let us know if you have any reactions to this or otherwise give us
- 27:40 – 31:08
The State of Nuclear Energy
- DUDavid Ulevitch
the, the, the state of nuclear. W-w-where are we right now? W-what are the, what are the m-major bottlenecks? What, what, what are we excited about? The biggest thing that's shifted in the last two or three or three or four years in nuclear is that, um, everybody now acknowledges that nuclear energy is clean energy. Um, I think that's been one, a major shift in public sentiment and perception. Um, nonetheless, there's still major headwinds politically with nuclear that need to be overcome. Um, Taiwan, for instance, turned off their last nuclear reactor.
- EPErin Price-Wright
Insane.
- DUDavid Ulevitch
This is an, this is an, this is an... It's unbelievable. This is an island country that is, you know, seven days away from a total energy blo-blo-blackout if they get, um, an oil and gas blockade from China so that they can't bring in ships to, to deliver oil, um, and fuel. Uh, so they're, they're like at any given time, they're like seven days away from a total blackout, and they turned off their last nuclear reactor so- Why'd they do it? Why, why did they do that? [laughs] Because they caved to political-
- EPErin Price-Wright
Political will
- DUDavid Ulevitch
...um, I would say minority, uh, g- like, very loud, vocal minority groups. Like active environmentalist activist reason? Yeah, active, yeah. Cyan. And they, they sort of this party ran on a commitment to turn off the reactor before they realized how stupid it was. [laughs] Um, it's just like colossally stupid. By the way, turning off a reactor, like a real full-scale reactor, um, it's not like an SMR where you can just like flip it on like a few days later or a month later. Um, you know, but like with these large reactors, it can take years to turn them back on. Um, and, uh, so it's just, it just, that was just terrible. But broadly, I think the tailwinds for nuclear are just getting stronger, where people recognize that, um, it is clean energy. Uh, I think there's still messaging work to be done. We should stop calling the spent fuel nuclear waste because it's really not waste. Almost all of it can be recycled and reused. It, it is-- People do need to recognize that, like, there are, those tailwinds are shifting, so that's happening. I think that people understand that it's base load power, right? So it's not dependent on it only working during daytime. It's not like hydro where you have to be around an appropriately configured water source. Um, and then I think that, you know, one of the largest inhibitors to creating new power plants in this country, um, it's not, it's not like the, you know, we, we, we can't do it. We c- we can. There's a huge regula- regulatory and permitting, uh, I would say morass that has to be g- you know, swam through that is incredibly expensive, requires an army of consultants, uh, many tens of millions of dollars, many, many thousands of pages of applications and documentation and, and process review. And again, this has to do with like building, building the power plant, getting the fuel, transporting the fuel, storing the fuel. Each step along the way, um, is extremely, extremely laden with regulation, uh, and policy. Um, and some of that's for good reason, but finding ways to better navigate that to make it more efficient is really a step in, in area that a lot of companies and people are focused on now. And actually, I think the government now is also focused on how do we streamline, um, the approval process for a new reactor, or how do we start approving new reactor designs. Um, then I think the last thing I guess I would say is thatYou know, not r-right now, if you're gonna do a lot of ener- if you're gonna put a lot of energy and work into building a nuclear power plant, you wanna build a really big one. So like, we largely only see really big power plants in this country, like the AP1000s that we turned on in Georgia. Um, and those again came in, I think they were 10 years late and like, you know, multiple billions over budget. We are now starting to
- 31:08 – 35:56
Advancements in Small Modular Reactors (SMRs)
- DUDavid Ulevitch
see movement, um, from the government, from in the, in the DOD, in the Department of Energy, um, and from the national labs to really try to create a more fast-track process for these small modular reactors or even microreactors, um, that use a much safer form of fuel, use much less nuclear fuel, use a different kind of nuclear fuel that's not nearly, um, a-as risk-prone as, uh, the kind of nuclear fuel you, the or nuclear material you'd use in like a weapon. Um, it's not nearly as enriched to the same degree. It's not even the same, um, material. And so that, that process is now gaining a lot of steam. We have an investment in a company called Radiant Nuclear. They are building a factory that creates, um, what effectively is an SMR. They would, they would probably call it a microreactor. Um, it's a one-megawatt reactor. It can be put on the back of an 18-wheeler and shipped around. You can move it to where you need power if there's been a natural disaster, like a hurricane, and you need to bring in power overnight. You could bring in a truck, uh, you know, a few trucks with, you know, four or eight of these reactors and power up a whole city after a disaster. And so that kind of flexibility and power is really compelling. And the, you know, the la- last thing I think I should say about on the SMR and microreactor side is that, you know, the Unit- the United States military spends an incredible amount of money just around the lo- dealing with the logistics of moving fuel around to forward operating bases. Um, anytime we do a military exercise, anytime we are engaged in a conflict, the movement of fuel factors in as a primary concern, uh, and consideration of what we deal with. And so just think about it, and, and, you know, we've read reports that they spend well over two hundred dollars a gallon at times, sometimes up to four hundred dollars a gallon for diesel effectively to get diesel into the right place at the right time. And so you can just imagine that having, uh, a nuclear reactor you can put on the back of a C-130 and fly, [chuckles] fly around the world to wherever you need power, drop it in the middle of the desert, turn it on, you have power for five years, is just an incredibly compelling value prop. Um, there is no question that nuclear needs to be part of the equation. Um, not only is that the base load power, but on the SMR and microreactor side, it gives us just incredible flexibility and grid resilience. Uh, you know, there should not be a single military base in this country that's not nuclear-backed, um, from a power standpoint, because if the grid goes down, whether it's from a cyberattack or just instability or, or demand issues or cascading failures, you wanna be able to fail over to nuclear power and not worry about the runway lights turning off.
- EPErin Price-Wright
Yeah. And especially as you, we start to look at the kind of, uh, electrification of our weapons systems, our military vehicles, our, you know, d-drones, et cetera, like those all have, ha-have, uh, need to get charged up somewhere, and w-with how better to charge them than a nuclear reactor.
- RMRyan McEntush
Other thing I'll add to your nuclear comment is like the, I think the advantage of nuclear, and I think that Radian has done very well, uh, uh, of really sort of leaning in on, is the, the power density factor. Like if, if you want a, a, a reactor that is reliable, uh, and, and power dense, you want it to operate at very high temperatures. You want as highly enriched fuel as you possibly can, where it's like makes sense commercially. So you want HALEU fuel. Um, and you wanna serve customers that will pay the premium for that. They'll, they'll be able, they'll, they'll, they'll buy sort of this reactor that they know is going to work. And if you're doing that, you want to, you want to have as, you know, you wanna have these economies of scale on the manufacturing side. You want it to be done out the door and don't need to like, you know, assemble it on site. You don't wanna have to like, have constant maintenance. And I think the, you know, the, the other sort of reactors that we see, you know, maybe on the civilian side, um, if you're gonna start having to do, uh, you know, if you build a reactor in, in a factory or you build modular components in a factory, but you still need to do construction work on site, you're, you're still a construction company. Um, like even if the technology is there, um, and I would argue a lot of the existing AP1000 technology is quite good, and other countries can do it quite cheaply. China is using a very similar design. The UAE just built one for incredibly cheap. Um, they have very similar nuclear, like, regulation, um, like in terms of frameworks. And obviously their, their regulatory bodies might move faster and things like that, but, but they're not, they're not like, you know, completely ignorant o-o-of, of some of the, the concerns. Um, and so, you know, I think in, you know, high level, and maybe this is a much more broad question, the United States needs to get better at mega projects, things that are a billion dollars, things that are, are at scale. Um, and I would argue like, you know, the NRC is, is a big component of why it's expensive, but I think it's also, you know, the same reason that it takes a billion dollars to build a bike lane in San Francisco is why we are not able to build small modular power-
- EPErin Price-Wright
Or why we don't, why we don't have a high-speed rail in, in California.
- RMRyan McEntush
Yeah. Yep.
- EPErin Price-Wright
And there, there is a, I mean, w-we, we debate a lot-
- DUDavid Ulevitch
Well, we, we might not have a high-speed rail in California-
- EPErin Price-Wright
[laughs]
- DUDavid Ulevitch
... 'cause nobody wants it and, and nobody wants it in-
- EPErin Price-Wright
I want it.
- RMRyan McEntush
That's a hundred dollars a flight.
- EPErin Price-Wright
I fly to LA all the time.
- DUDavid Ulevitch
Well, nobody... Sorry, nobody wants it where they're building it.
- EPErin Price-Wright
Sure. Yeah. Um-
- DUDavid Ulevitch
Bakersfield is not a, a prime destination.
- EPErin Price-Wright
[laughs]
- RMRyan McEntush
It could be.
- EPErin Price-Wright
I want a train from San Francisco to LA.
- RMRyan McEntush
[laughs]
- EPErin Price-Wright
That takes an hour and a half. Um-
- RMRyan McEntush
Me too.
- 35:56 – 38:20
The Role of Technology in Mega Projects
- EPErin Price-Wright
Yeah, and I mean, we debate a lot internally, like where does it make sense for VCs and VC capital plug it, to plug in? And like, arguably, like we're not gonna move the needle on, you know, these multi-billion dollar mega projects in the U.S. Like, it doesn't... You, you know, we're not gonna build... We're, we're probably not the best people to figure out how to, you know, capitalize and build, you know, a multi-billion dollar project in, in, in California to generate the power for the grid. But I do think that there is a role for technology at kind of like every single layer and every single phase of how mega projects like get built. It's like, how do you use AI to navigate kind of site selection? How do you, how do you use tools to like, move through the various permitting processes faster? Like, how do you use AI to help you do extremely complex and interdependent project management better and more effectively? So that's something that, you know, you, you have a project with 4,000 people working on it, and everyone engaging with different suppliers and timelines that are dependent on each other. Like, how do you get all those things to align better, um, so that you don't get these 10-year delays so that, you know, projects actually happen on time and, and on budget, and, you know, as a result, attract private capital, um, backers. Like, I, I think that there's a role of technology here, you know, what that looks like.TBD
- RMRyan McEntush
We've seen a lot of companies that maybe, you know, five years ago were primarily trying to sell to utilities and grid operators, which is incredibly painful, incredibly difficult, perhaps rightfully so. Like, they have poles in the ground that are fifty years old. Why would they trust a two-year-old company to sell them software? Are they gonna be around in twenty years? And this is a fair question to ask, especially for something as critical as the grid. Um, but now sort of they're developing the software, and there's such demand of understanding how grid operators might think, uh, and potentially get there faster or, or, or, you know, ha-have different conclusions. And so now you can go to data centers or, or people who wanna build solar farms, or, or people who wanna build massive, like, battery farms. And you can sell a very similar software-
- EPErin Price-Wright
Or even individual people who wanna-
- RMRyan McEntush
Yeah
- EPErin Price-Wright
... make sure that their power isn't gonna go out and they're gonna be, um, you know, c-caught without energy during an important moment in their lives.
- RMRyan McEntush
Yeah. And so there's, there's now-- Like, every-everyone cares now. There's a lot more money who cares about what do, what is the grid actually going to think and where can I build? Where is there excess capacity? Uh, maybe, maybe I, I'm connected to the grid, but, you know, I also need some battery and solar backup or, or like a, a radiant microreactor or something like that to, you know, to be used in certain situations. It's, it's a lot more complex, this sort of microgrid setup, but it's, it's the way we're headed and, and software is gonna be a big piece
- 38:20 – 43:08
Opportunities in Grid Management and Monitoring
- RMRyan McEntush
of that.
- ETErik Torenberg
I wanna hear more about our request for startups or things that we want to e-e-exist that we haven't yet discussed or, I mean, put differently, like, I'm curious where we think there's most bang for the buck in terms of the issues that we've been talking about in terms of if there was, like, a regulatory intervention or some sort of technological u-unlock. What, what, what comes to mind?
- DUDavid Ulevitch
One area where there's probably a venture-scale software company to be built is really around grid management monitoring. And like, you know, I think we see this, um, in the IT landscape, we see it in the OT landscape, but we don't really see it, uh, in the grid where there's just full, you know, very, very large... Like, there is no Splunk for the electrical grid. Um, there's no Palo Alto Networks for the electrical grid yet. Like, there's a whole bunch of things that mirror the IT and OT landscape, whether it's around cyber and monitoring and logging and analytics. Um, you know, there's no, like, Looker for, for the electrical grid yet. There's just none of these companies exist. Um, I'm not sure if it's three separate companies. I'm not sure if it's one company, you know. But there, there is a big company to be built in really managing and monitoring the grid, uh, and helping to orchestrate and even, even deal with some of the things Ryan spoke around, around demand response, coordinating that, creating those marketplaces, tracking all those incentives. Um, you know, so I think when we see a company that we think can really be the breakout company there, we would, we would lean into it.
- EPErin Price-Wright
I also think around sort of project planning and development, um, h-how do you make it faster and easier to build product, projects within the current regulatory framework? You know, how do you, how do you do site selection? How do you navigate permitting? How do you, um, you know, how do you navigate project management? How do you navigate your sort of construction supply chain? Um, you know, we, we, we're starting to see companies kind of pick off pieces of that, but I think, you know, broadly speaking, there's room for tech and software, um, in that kind of project development space as well.
- RMRyan McEntush
Anything that can bring generation capacity or storage capacity closer to load, I think is gonna be very compelling. Um, and a lot of the times it's less maybe the technology, novel technology, but it's sort of a system integration or it's an innovative business model. I think something like radicalizing that I experienced is, and, and I, I implore everybody to go home and check their power bill. They'll now, like, often separate, like, the delivery costs from the actual generation costs. So what we've seen, and we've sort of mentioned it, but, like, the cost to generate electricity, the cost of, like, power has dropped immensely, gas, solar, things like that. Um, but the cost to actually deliver that electricity has increased a ton. And so in net, it's sort of not changed. Uh, and that's, you know, I think that's terrible, and I think we all agree that's bad. Um, and so I think there's a lot of opportunity o-of bringing sort of that generation capacity. In some ways, this is sort of like this, you know, more, more liberalizing force. It's like we all should have our own backup. We all should have our own technology. Uh, a-and, and I think there's a lot of really interesting ways to, to do that and scale it.
- EPErin Price-Wright
Overall, as a grid gets more heterogeneous, um, you know, all of the seams and intersections between things, like, that's the, the... There's just so much more opportunity for technology than, you know, when you had a single utility managing a single source of power centrally kind of distributed out, uh, distributed out broadly.
- DUDavid Ulevitch
All the regulation and permitting and policy, uh, frameworks that we have in this country, you can sort of think of those as part of the, like, infrastructure that we all have to, like, live and work with and interact with. So companies that really facilitate, um, a-and I would say applying AI to navigating the permitting process. So nuclear is a good example. Again, a nuclear, uh, reactor application or a fuel transport, uh, license or, uh, a fuel, uh, manufacturing license, these things have thousands and thousands of pages of regulation and documentation that go with them. You make one small change in your application, it has these reverberation effects, so you have to update all your documents elsewhere. If you're the regulator trying to go through all these applications, it's just incredibly onerous, um, border-borderline impossible to imagine that a regulator can even possibly get it right. Um, it's, uh, you, you could argue that it's actually not possible. They just do a best effort. But AI could actually help these things, could help the applicants, um, go through the process of, of filling out their applications and saying, "Hey, this is where you should drill down. This is where you should clarify." They can look at all previous published applications and say, "This is how you need to tailor it. You can probably make it eighty-five percent the same, and then based on your design or your location or whatever, make it, make some modifications." And then the regulator can do the same and say, "Look, here's an application that came in. Highlight all the areas I need to drill down or make sure, or show me the things that are different from every other nuclear, uh, fuel transport application we've ever seen. Um, are they using the rail infrastructure? Are they using, um, the national highway infrastructure to move the fuel?" Like, you know, just AI can just automate all these things that take armies of consultants, you know, months or years to do, can be, can be, you know, d- uh, brought down into being minutes or, or, or hours. Um, so I think there's potentially... You know, I, I'm not sure how big of a company, but I think potentially there's a very large company to
- 43:08 – 49:00
The Future of Energy Infrastructure and Policy
- DUDavid Ulevitch
be built there.
- ETErik Torenberg
If our check sizes were in the billions, not, not just the millions, which we're in Andreessen Horowitz, you never know, um, h-how would our, how would our strategy change?
- DUDavid Ulevitch
Well, Ryan, you're talking about mega projects. What, what kinds of mega projects would you-
- RMRyan McEntush
I think, I think you need more than just billions. It's tens of billions, hundreds of billions. Um, it's such a tough question. I think there's, there's, there's so-- there's, you know, tons of policy, um, around this as well. I, I, I think... I hesitate to say we should, we should look to, to how China's built up their, their grid. But I think, like, uh, I think sort of the elephant in the room is, like, they, they-- You know, in the early 2000s, they were experiencing blackouts. Like, this was a very common thing, and this was, this was horrific. Um, but, uh, you know, now they've-- I think they've, like, 4X'd their grid, like, in, in the last couple of decades. And so the way they've done this is, is by basically deploying generation capacity, building hydro, building massive storage facilities, of course, BYD, CATL, like, tons of battery production. Um, they built HVDC, sort of these, these large sort of high voltage transmission lines. Um, I would do all of that. I mean, I would look, I would look to all of it. And whether or not, you know, it's a, you know, a good investment or not is gonna be a number-- you know, look at a number of factors. But, like, it, it is, it is much more of the infrastructure projects, sort of the glue that connects this stuff together. I think our lens today is looking at these technologies that enable a lot of sort of this, this more flexible grid. Um, but I think there's also gonna be these large infrastructure, this-- the webbing in between it. Uh, and I think software's a big piece of it that we're spending a lot of time on, uh, looking at. But I think, like, you know, how is, how is ERCOT gonna be connected to, to the rest of the, the grid? Or how are we gonna move this power around if, you know, it's really sunny in, in the Southwest, and there's, like, solar's gonna be really cheap. Is there an efficient way to move that to, to New York or something like that? Um, and China's done this effectively. Uh, and I think, you know, if you had hundreds of billions of dollars spent or trillions of dollars, what does the grid look like? Like, it's gonna be a lot more interconnected.
- EPErin Price-Wright
Another answer to your question or a different answer to your question is I think the energy industry is probably medium to long term, like one of the most prime, um, spots to deploy physical autonomy. So when you think about applications of robotics, whether it's humanoids or more kind of task-specific robotics, um, I, I-- y-you know, I think these are the-- We're, we're talking about dangerous jobs often. We're talking about manufacturing jobs, um, to build up, you know, s- whether it's small scale reactors or batteries or whatever. So, uh, I, you know, I don't know what the shape of the company is here and how reliant it would be on some of the robot learning work that's happening. But I do think that, you know, th-there's gonna be, um, a-as we scale our energy capacity, there's gonna be a pretty massive application of, um, industrial bo- robotics to, uh, to the energy sector.
- DUDavid Ulevitch
You know, like, we survived the greatest nuclear disaster in U.S. history just recently when we finished the Vogtle three and four reactors and let all those employees go back to other jobs. So I think if we were writing a billion-dollar check into, into, uh, power, what we would do is we would just give jobs to those people and not let them go back to whatever it was they were doing before they were building nuclear, uh, reactors. And we would just really work to streamline the process to make sure that we go build Vogtle five, six, seven, eight, nine, ten in all these different states around the country, um, and just put these people to work for the next decade plus, uh, building reactors. I don't think it's particularly our opportunity, but I do think it's an opportunity, uh, for somebody to do.
- EPErin Price-Wright
I mean, l-labor broadly, like, I mean, this is a, a little tangential, but when, uh, Microsoft was building their new data center in Georgia, um, last year, at one point, they had on staff at Microsoft or on contract more than a third of the electricians in the state of Georgia, and they basically maxed out. They hired every single electrician that they possibly could.
- DUDavid Ulevitch
Mm-hmm.
- EPErin Price-Wright
Um, so it-- I don't think it applies to just electricians. It's to-- due to your point, like, s-the cement mixers. It's the, um, you know, mechanical engineers. It's the nuclear engineers. Like, how do we actually train the next generation energy workforce that we're gonna need to, y-you know, modernize the grid is a big-
- DUDavid Ulevitch
By the way, these are very high pa-
- EPErin Price-Wright
... a big challenge.
- DUDavid Ulevitch
The-these are very high-paying jobs where you don't have to check your email on your phone at, like, nine PM at night-
- RMRyan McEntush
[laughs]
- DUDavid Ulevitch
... after you go home from work. So they're high-paying, um, you know, it's good exercise jobs, and, uh, they're relatively low stress. Uh, so these are, these are good jobs for people.
- RMRyan McEntush
I, I, I think one more, one more comment on this, a-and it's more of an industrial policy question, is, like, we're talking about specific things, but, like, oftentimes that just moves the bottleneck. Like, we could, we could solve a lot of sort of the grid connection hookup. We could build a lot of, like, transmission lines, but then we need more transformers. And to build more transformers, you need more electric steel. And it's like you can do this, you can do the same sort of e-equation for, for much of the supply chain. Uh, battery is another good example is then, okay, cool, we're building cells, but then we also need active materials, and we need to mine and things like that. And so, you know, it's sort of a whole effort o-of examining sort of our, our infrastructure and our supply chains, and you need to do all of it. Um, a-and I think that's a, that's a complicated question. That's an expensive question.
- DUDavid Ulevitch
You know, people underestimate, um, how critical and important, uh, a resilient, reliable, dispatchable electrical grid is to our national security. You cannot have national defense and national security without reliable electricity. It's just not possible. So all these things we're talking about are about the upside, about capitalizing on AI compute, the tran-- you know, the switch to electric vehicles and our insatiable thirst for electricity. But at a fundamental level, there, there is no, uh, there is no safety, there is no national defense, there is no national security without a reliable electrical grid.
- RMRyan McEntush
People want che-- uh, reliable, cheap, and clean power in that order. A-and, and I think that's, that's, I think, largely how we should think about, uh, our energy policy. Um, a-and I, and I think that's, uh, I think that's sort of the direction we're going. Um, a-and, and I think we need to make sure we, we, we stay aligned with that. [upbeat music]
Episode duration: 49:14
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode yjKmMxUz2xc
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome