Assertion Contradicted
Kann: Software engineering is the only job category truly displaced by AI to date
“And in fact, the only jobs that have been lost really because of AI to date are like software engineers.”
Prediction Not checkable as stated
Kann: Electric vehicles will never discharge to the grid at scale
“I do not think we'll ever see vehicles discharging into the grid at scale.”
Assertion Supported
Kann: Sulfur injection could cool Earth 0.5°C for $2 billion
“Specifically, you might be able to get about a half a degree Celsius of global cooling out of stratospheric sulfur injection for a couple of billion dollars, which is nothing, basically, comparatively.”
Opinion
Kann: Data centers lower local electricity rates but increase macro power prices
“At the local level, data centers seem to on balance, probably decrease electricity prices currently. And at the macro level, they are definitely a factor contributing to an increase in electricity prices.”
Opinion
Kann: V2G economics do not pencil for EV drivers with $7,000 hardware
“GM, for example, their bi-directional charging bundle runs about 7000 dollars before installation. And grid programs so far have paid drivers a lot less than that. It just doesn't pencil.”
Opinion
Kann: Power is the rate limiter to AI growth
“And understanding the multiple factors that drive failure of new transmission, I think, is key to understanding why power is, in my opinion, The rate limiter to AI growth, and why upward pressure on electricity prices will probably sustain for a while.”
Assertion Partly supported
Kann: Almost none of the 50 GW data center pipeline is off-grid
“There's like, 50 gigawatts of behind-the-meter generation and development at data centers, right? And they, and some people have interpreted that to be, oh, 50 gigawatts of off-grid data centers getting built. It's actually close to zero of those that are true…”
Prediction Held up
Kann: 300 MW data centers have lower cost per FLOP than edge
“My guess is, you know, your 300 megawatt data center on a fully loaded levelized cost of flop is just gonna be cheaper.”
Prediction Open · timeframe Feb 2031
Kann rejects Musk's claim that orbital data centers will be cheapest soon
“Because I think that orbital data, I mean, Elon says he thinks that they will be the cheapest way to get new compute in, like, three to five years. I do not think that is possible. Not in that time period.”
Prediction Open · timeframe Feb 2029
Kann predicts US electricity prices will rise over the next three years
“But I could tell you, I think over the next three years, you know, electricity prices are probably going up, not down.”
Assertion Supported
Kann: Solar geoengineering could deliver 0.5°C cooling for two billion dollars
“The rough estimates, just to an order of magnitude, are that, at least what I've seen, it might cost a couple billion dollars to deliver something like a half a degree of cooling, half a degree centigrade of cooling.”
Assertion Supported
AI training data centers run dummy workloads to flatten dangerous power spikes
“What they do in large part, because those spikes are actually problematic, not just to the grid, but to the equipment inside the data center as well. So what they do at least sometimes to manage that is they create dummy workloads. So they keep the power profi…”
Prediction Held up
Kann: Data centers will partner with DER aggregators for grid capacity
“We're just starting to see this wave of announcements that are inexorably coming where there are going to be partnerships between hyperscalers or data center operators and the Third parties who do aggregations of DERs to deliver capacity to get data centers on…”
Assertion Supported
Grid interconnection is faster than building new nuclear in most locations
“Most places you can get interconnected faster than you can build the new nuclear reactor.”
Insight
Kann: Geoengineering is too cheap to support large venture market sizes
“It's actually hard to get to a big market size for it, ultimately, because it is so cheap. Like, I don't remember exactly the number, but we did an episode on this, like, I don't know, a couple years ago, and it was something like a couple billion dollars migh…”
Opinion
Kann: Make Sunsets' rule-flouting balloon releases are the wrong approach for SRM
“There is another startup called Make Sunsets, which a lot of people have heard about, who has decided to just do SRM brazenly in the open, kind of flouting all rules and regulations to the extent that they can. I actually went to a party a couple years ago whe…”
Opinion
Kann: Rising electricity prices are the biggest risk to EV adoption
“To me, you know, the biggest risk to EV adoption right now, above all else, is rising electricity prices.”
Prediction Not checkable as stated
Kann: A US nuclear power renaissance will happen despite realistic timeline hurdles
“I'm now very optimistic about it. I'm realistic about the timeframe of it, I think. Which is, a lot of people are not, but I do think it's going to happen.”
Insight
Kann: Being green and more expensive is insufficient for climate tech demand
“The big lesson of the current state of affairs in climate tech is that in a lot of the subcategories of the market,
Being clean, being green, whatever, is insufficient.
Being ex but clean particularly if it's ex but clean and more expensive.
Is just not eno…”
Opinion
Data center developers are crowding out sites for new U.S. manufacturing electrification
“I can tell you for certain that it is getting harder and harder to find a site to put any of that stuff because there are data center developers hunting around the entire country trying to find anywhere where that is possible. So it's making it, and that's dif…”
Prediction Not checkable as stated
Kann: The US lacks end-user demand for 1,000 new hyperscale data centers
“And I will state categorically, I don't think there will be a thousand new Hyperscale data centers in the United States anytime soon. I don't think there's actually that much demand for it.”
Prediction Open · timeframe Dec 2030
Kann: US data center capacity will more than double by 2030
“I was on the side of, we will see total capacity of data centers as measured in megawatts, or gigawatts, I guess, really double by twenty-thirty, and I was on the side that we will see it more than double.”
Prediction Not checkable as stated
Kann: Carbon removal startup oversupply will force a market reckoning
“I think the overwhelming likelihood is that we're in for a bit of a reckoning in that market. In that market, I mean, both the general voluntary carbon market, and more specifically, the CDR market. Not to say everything fails, but there's just the, there's an…”
Insight
Khan: Clean tech startups need a 5x-10x improvement to beat scaling incumbents
“If you're building something that you think is a better mousetrap, like, it's 20% better, Than whatever the incumbent technology is. In all likelihood, your reaction, my reaction, when we see a company that's doing that is like, 20% better is not better enough…”
