The AI Boom is Coming, But at What Cost to the Planet? 🤖 🚀 🌍 ❓
Recently, I came across a fascinating article from Caltech that highlights the importance of sustainable computing in the world of AI.
According to the article, the computing industry is already responsible for a staggering 2.1-3.9% of global carbon emissions, and this number is set to skyrocket with the growth of AI and other energy-hungry applications.
The National Science Foundation (NSF) recently announced a new $12 million multi-institutional research initiative over the next five years to delve deeply into this topic, with an ambitious goal to reduce the carbon impact of Information and communication technology (ICT) by 45% by 2030.
Here are a few key takeaways from an informative interview with Adam Wierman who is a part of the NSF Expeditions team.
1. Adopting generative AI can increase carbon emissions by 10-20 times, potentially doubling the carbon footprint of IT in just 5 years if improvements aren't made.
2. There are 2 types of carbon usage for a data center:
a. Embodied carbon - that goes into the manufacturing of the hardware and the building.
b. Operational carbon - from running and doing the training.
3. Attention needs to paid to the local impact where these data centers are setup:
- Impact on energy and water resources for people who live there.
- Pollution associated with the backup generators.
- Data centers create few local jobs beyond construction, as they require minimal human operation once built.
4. Adam is also on the advisory board of Verrus backed by Sidewalk Infrastructure Partners (SIP).
This startup company is working on optimizing the way power works in the data center, of the way scheduling works, of the type of storage and backup power systems that you build in the center.
5. There is a need to move toward a data-center design which is responsive to the grid, shifting workloads to times when green energy is available and reducing loads during peak demand periods (Peak shaving).
6. To minimize embodied carbon, we need to measure and reduce the carbon footprint of manufacturing computing resources, leveraging the three R's: reduce, reuse, recycle.
🔗 Read the full article here: https://lnkd.in/dC8qJPS6
#SustainableComputing #AI #Artificialintelligence #ClimateAction #GreenTech #DataCenterSustainability #TechForGood #InnovationForSustainability
👨💻 JavaScript Engineer | React Engineer| Nature | Wild Life | Stoic
5dAs data centers evolve under the pressure of AI’s intense computational requirements and multi-cloud approaches, how can enterprises guarantee that their forward-looking strategies won’t simply accommodate current hype, but also anticipate long-term sustainability and resource constraints?