@HarrisBerton Depends on the peakiness of the peak. If you need 8 or more hours, yeah that’s a problem. Under 4 and you’re good. 4-8 and you need to augment with other resources.
@HarrisBerton We are seeing the opposite. Significant, consistent kw savings during morning peak when you replace electric resistance. At 5kw per house, you get MW of load reduction pretty quickly.
@JaneAFlegal@sdamico This doesn’t need to be theoretical. We can see this in existing data where load peaks are pointy versus relatively flat. DERs are good for pointy peaks. Base load gen for flat peaks.
@duncancampbell@ArushiSF@TogetherNOLA@grok If you want to go through PJM. But Utilities might be better for DER deals so that you solve distribution constraints too. I’d just PJM capacity credit, all bets are off. Could make local probs worse. If utility in the middle you can be smart about it.
@ArushiSF@duncancampbell@TogetherNOLA@grok There are a lot of details missing that make it hard to know. Bottom line, it’s a good thing. But not an unlock in any meaningful way.
@duncancampbell@ArushiSF@TogetherNOLA In PJM the utility clears the capacity based on settled loadshape across 5cps. I haven’t seen any indication that this capacity is assignable to a data center and the PJM flex tariff doesn’t allow for it yet.
@duncancampbell@ArushiSF@TogetherNOLA Obvs the big outlier. But notice no utility partner. Means prob either a fixed for floating swap on existing capacity or possibly funding new build but Voltus doesn’t do that and there was no partner mentioned.
@boo@ShanuMathew93 I just don’t get what they could be working on to use that many tokens? Like, I’m churning hard on stuff and at best I’ll hit the weekly max on $100/month plan.
@MattGrossmann I would expect more quant analysis of esoteric data that’s hard to compile, e.g., speeches. LLMs really good at scraping and writing Python.