TL;DR: Last month at New York Climate Week I leaned heavily into sessions from financial and consulting organizations. In the process, I got a lot of insights into macroeconomic trends, power (in every sense of the word), and of course, AI. I'm going to cover each of these in the coming weeks, because hitting all at once could easily be the length of a new book. (If you need a brief recap of the strange way that Climate Week operates, you can find that in the newsletter from 26 September.)
Today's topic: power. And we're going to start by connecting the dots between last week's newsletter about AI.
Energy is getting really expensive for residential customers. Why? Because of AI.
On the surface, that doesn't make sense. What does some OpenAI server farm have to do with my electric bill? Because our incentive systems are upside-down.
If I, Sam Altman, want to build a hyperscale datacenter in your town (always your town, never mine), I'm going to submit proposals to several municipalities on a variety of power grids and play them against each other in order to secure the best possible deal for my company. That means that I'm going to negotiate an incredibly low rate per kWh of energy that I use. Oh, but wait: in order to accommodate this increased energy use, the local power company will need to build out the grid: more generation, more robust and redundant systems, more of everything that makes a power company function.
But if I'm not charging the datacenter for it, where am I going to get the money? I'm going to give you three guesses. Fortunately, this is starting to get some traction in mainstream media and not just in academic papers.
But I thought that energy costs for renewables were getting cheaper, and were lower than traditional sources like coal and natural gas? Correct. Even subsidized fossil fuel plants can't compete with wind and solar on price.
Oh, so you're wondering why we don't just add more wind and solar capacity? After all, power generation stands as the foremost contributor to global carbon dioxide (CO2) emissions. Well, at Climate Week, I actually got to the heart of that one. I had been operating under the assumption that adding grid capacity was time consuming and costly. And it is. But that's not the only reason. The real reason is rooted in political power.
I went to a panel discussion that featured representatives from various local governments in the U.S. They spoke openly about denying plans for new renewable power plants because... wait for it... solar and wind farms don't generate jobs like fossil fuels and nuclear power do. So until these lower-maintenance sources of energy generate more hands-on, long-term jobs for the community, they're not interested. (Note: this thinking predates 2025, but is only bolstered by current structures in the U.S.)
So they'd rather reopen nuclear plants that are beyond their lifespan, or keep open highly polluting coal plants that were scheduled to close? What about the health of constituents? In a one-on-one conversation where I voiced some reluctance about old nuclear plants (yes, there are better and safer ways to do "new nuclear," but I was a kid in Pennsylvania during the Three Mile Island accident), I was told to think of it like the sell-by date on a yogurt or a bottle of ibuprofen. Just because it's past the date doesn't mean it's not still ok. It just means that they can't guarantee it.
This was supposed to reassure me.
It did not.
And as for the pollution from coal plants... well, nobody really notices those continued health impacts. After all, they've been living with them for years. Who's to say that the particulate matter from the power plant is really the source of your lung cancer instead of, say, microplastics or red food dye?
And even if you do have a power plant that primarily runs on solar or wind, it's using fossil fuel backup generators, which are still generating PM2.5 airborne particulates that cause a whole host of health issues. You'd think that we'd want all of those emissions kept far from human populations, but instead places like San Jose are looking to build combination datacenter/residential developments.
Isn't anyone thinking about the impacts of all of this? Yes, but it seems like we're not in a position of power. I believe that technology can be done responsibly, but it requires a lot of competing interests to have a seat at the table, and that doesn't seem to be happening. "Smart regulation, accurate risk pricing and allocation, and market-led investment in proven, least-cost solutions can ensure that the AI era advances not just technological progress but also resilience, economic stability, and global responsibility."