Hacker News Reader

Comment by hetspookjee | Hacker News Reader

Apologies for the cynism, but I wonder what big-tech firm bought the emission rights for this to expand their data hunger even further while "staying green". Here in the Netherlands we have a fierce discussion raging on the placement of a newly proposed "hyperscale" from Facebook/Meta. The datacenter would consume more than the power of Amsterdam as a whole, yet we have a lot of trouble finding the right grounds (and communal common ground) on where to place these enormous turbines.

and off-topic:

I bet that this entire windpark can be offset by disallowing arbitrary Javascript code to be ran in every advertisement, and all the wasted cycles that follow. In the same vein, I wonder how much CO2 a regular citizin saves per year by running an adblocker.

View on HN

Somebody in my university did their PhD thesis on that. Let me see if I can find it.

(But the answer was yes. If I remember it correctly, 10% of all IT infrastructure energy consumption goes for displaying ads on web pages)

Found it.

"The online advertising ecosystem resides in the core of the Internet, and it is the sole source of funding for many online services. Therefore, it is an essential factor in the analysis of the Internet's energy footprint. As a result, in 2016, online advertising consumed 20–282 TWh of energy. In the same year, the total infrastructure consumption ranged from 791 to 1334 TWh. With extrapolated 2016 input factor values without uncertainties, online advertising consumed 106 TWh of energy and the infrastructure 1059 TWh. With the emission factor of 0.5656 kg CO2e/kWh, we calculated the carbon emissions of online advertising, and found it produces 60 Mt CO2e (between 12 and 159 Mt of CO2e when considering uncertainty). The share of fraudulent online advertising traffic was 13.87 Mt of CO2e emissions (between 2.65 and 36.78 Mt of CO2e when considering uncertainty)."

The paper was: "Environmental impact assessment of online advertising" https://www.sciencedirect.com/science/article/pii/S019592551...

It was the one of the five papers included in the authors PhD thesis, "Towards Sustainable Data Centers and ICT Services" https://aaltodoc.aalto.fi/bitstream/handle/123456789/38725/i...

Here’s a paper from a few years back with a low end estimate of 11MTCO2e for a year of digital advertising https://www.sciencedirect.com/science/article/pii/S019592551...

For comparison that one year is about the same as the entire history of ethereum, which gets a lot of criticism. Or about 4 world cups.

And a crazy 10x that on the high end of the estimates. Be interesting to see some new numbers on this.

> I wonder how much CO2 a regular citizin saves per year by running an adblocker.

Compared to hauling about 1 000 kg around daily for the office commute and compared to sending an airplane into the air one or more times per year, the advertisements aren’t too bad.

Really open for any feedback and pointers on the below calculations:

The average citizen of Chad emits 60kg CO2 per year, according to World In Data 2020 [1]. Of which roughly ±3% are current Facebook user [2]. Facebook is pushing hard on the "Open Internet" initiave to capitalize on this market, too [3]. If they succeed in doing so, and say, reaches 80% market penetration in Chad, I'm really curious to hear what percentage would be attributable to rendering javascript advertisements. I tried to look up some numbers but they vary widely and it becomes a heavy guessing game. One could argue to simply divide the amount of active users by the total power consumption:

The energy consumption of Facebook has grown somewhat exponentially over the years, reaching 5.1 TWh in 2019. While there users are growing too (9% from 2018 > 2019, it seems there energy consumption is growing more quickly). Regardless, they counted a rough 2.5B active users at the end of 2019. Giving a rough power consumption per user over 2019 5.1TWh / 2.5b = 2Kwh. To put that 2 Kwh in perspective, according to Forbes a phone uses on average a year 2 Kwh of power.

Looking up the average Co2 emissions per Kwh I find 475gram of CO2 / Kwh [6]. Adding that up gives a rough 1kg of Co2 emissions per Facebook user.

Given the extremely low emissions from Chad (DRC is even lower with 30kg / capita, but strongly doubting that). The market penetration of Facebook alone in Chad might be able to bump the actual emission from 60 > 61 kg, a near 2% increase per capita, just for Facebook. Which is a rough 30% of the advertisement market globally at the moment.

That 1000kg vehicle hauling for a rough 1000km is equivalent to a year of a Chad living, too.

[1] https://ourworldindata.org/co2-emissions

[2] https://www.internetworldstats.com/stats1.htm

[3] https://www.theguardian.com/world/2016/aug/01/facebook-free-...

[4] https://www.statista.com/statistics/580087/energy-use-of-fac...

[5] https://investor.fb.com/investor-news/press-release-details/...

[6] https://www.iea.org/reports/global-energy-co2-status-report-...

I remember doing some math on youtube forcing vp9 on hardware without acceleration was wasting as much power as tiny country.

All because licensing.

An insatiable market of deep-pocketed buyers hungry for renewable rights is exactly what we want right now, no matter what they use it for. Demand will drive us further up the learning curve, bringing down cost & building momentum.

Reply on news.ycombinator.com