In a world where your electricity is provided by a grid with power plants distributed all over the grid area (ie not necessarily local to your data center), "location based emissions" is...
In a world where your electricity is provided by a grid with power plants distributed all over the grid area (ie not necessarily local to your data center), "location based emissions" is meaningless. Unless the data center is producing its own fossil based power on site, I don't see how this is relevant.
I get that people don't like how it looks to use renewable energy credits to offset emissions, but the best places to produce wind and solar power aren't always the same as where someone may want to build a high electricity consuming something. Recs let you buy that renewable power, even if you're not using those exact electrons. But in a grid where there are many sources of electricity, you're never getting your electrons from just one source. Yes, recs are a mathematical way to account for emissions, but on the net the emissions from electricity use do go down by the measured value and since we only have one planet, it does impact the whole.
It's also worth mentioning that often times these renewable power plants are built because these data center companies, or other high electricity consumers, are agreeing up front to buy the recs from the producer, thus giving the developer an income stream to base their investment decision on.
The regional and inconsistent nature of renewables does mean that a system of accounting for renewable energy use must exist. The alternative is to only build new stuff in regions where the power is produced most efficiently and suffer the intermittency problems some renewables have. It may not feel great but the current system is robust and has real, meaningful accounting practices behind it. That doesn't mean we can't do better, but I don't think this particular issue is the most impactful place to focus efforts.
I never really tried to understand how carbon emissions are calculated, but I suspect in any case they're estimated with a (very) wide margin, the published numbers are probably the lower end of that margin, and in a few years, we'll begin to understand the reality is actually much higher than the worst estimation.
I don't think your feelings are correct in general, but definitely not in this specific instance. There is a huge effort in the scientific community to model and measure emissions correctly. One...
I don't think your feelings are correct in general, but definitely not in this specific instance. There is a huge effort in the scientific community to model and measure emissions correctly. One of those is maintained by Argonne National Lab called the GREET model. It's free to download if you'd like to play around with it. This is the gold standard in estimating life cycle emissions and many companies use this to calculate their impacts or the impact of their products, etc. There are definitely ways to bend the rules or generate a more rosy outcome, but no one is looking at a range of outcomes and just picking the lowest.
This article basically says: these companies claim they have reduced emissions and they do that buy buying credits from renewable energy producers. If you don't count the credits they are buying, their emissions are MUCH higher! Well no shit.
In a world where your electricity is provided by a grid with power plants distributed all over the grid area (ie not necessarily local to your data center), "location based emissions" is meaningless. Unless the data center is producing its own fossil based power on site, I don't see how this is relevant.
I get that people don't like how it looks to use renewable energy credits to offset emissions, but the best places to produce wind and solar power aren't always the same as where someone may want to build a high electricity consuming something. Recs let you buy that renewable power, even if you're not using those exact electrons. But in a grid where there are many sources of electricity, you're never getting your electrons from just one source. Yes, recs are a mathematical way to account for emissions, but on the net the emissions from electricity use do go down by the measured value and since we only have one planet, it does impact the whole.
It's also worth mentioning that often times these renewable power plants are built because these data center companies, or other high electricity consumers, are agreeing up front to buy the recs from the producer, thus giving the developer an income stream to base their investment decision on.
The regional and inconsistent nature of renewables does mean that a system of accounting for renewable energy use must exist. The alternative is to only build new stuff in regions where the power is produced most efficiently and suffer the intermittency problems some renewables have. It may not feel great but the current system is robust and has real, meaningful accounting practices behind it. That doesn't mean we can't do better, but I don't think this particular issue is the most impactful place to focus efforts.
I made this comment not even a month ago...
I don't think your feelings are correct in general, but definitely not in this specific instance. There is a huge effort in the scientific community to model and measure emissions correctly. One of those is maintained by Argonne National Lab called the GREET model. It's free to download if you'd like to play around with it. This is the gold standard in estimating life cycle emissions and many companies use this to calculate their impacts or the impact of their products, etc. There are definitely ways to bend the rules or generate a more rosy outcome, but no one is looking at a range of outcomes and just picking the lowest.
This article basically says: these companies claim they have reduced emissions and they do that buy buying credits from renewable energy producers. If you don't count the credits they are buying, their emissions are MUCH higher! Well no shit.
Thanks for the reference
Speaking of, can anyone with expertise comment on the technical feasibility and reliability of carbon accounting?