Why do I keep hearing about compute? What is it, and why is it important?
How compute power works, or, the resource hungry machines sucking up the world’s water supply
This week, I want to talk about a word that I’ve encountered a lot in the past few weeks: compute. To me, compute means calculate. But in the world of tech, AI, and crypto, compute refers to the power and ability of computers to process information. Think of it as the brainpower behind everything computers and smartphones do.
When you ask Siri a question, when Netflix suggests a movie, or when Instagram shows you pictures, it’s compute at work.
Compute resources are clusters of CPUs (Central Processing Units) working together to handle large amounts of data and perform complex tasks. Every computing task requires a blend of memory, storage, network, and input/output (I/O). Compute is what brings these things together.
Why compute matters
- Compute power allows companies to quickly create and test new apps and features, leading to faster innovation.
- With powerful compute resources, tech companies can make sure their services run smoothly without interruptions, even if some equipment fails.
- Compute power supports the global spread of apps, allowing people from different parts of the world to use the same services efficiently.
While compute resources are essential, there’s a big environmental downside.
The environmental cost of compute
Let’s look at how much power is used for a ChatGPT query. One query to ChatGPT can use between 0.3 to 3 kilowatt-hours (kWh) of electricity, depending on the complexity and length of the response. That’s roughly equivalent to using a standard 30-watt LED bulb for 10 to 100 hours.
While compute resources are essential, there’s a big environmental downside. Data centres, which are buildings full of powerful computers, use vast amounts of electricity. This contributes to greenhouse gas emissions, especially if the electricity comes from burning coal or gas. The extensive use of compute resources results in a hefty carbon footprint, worsening climate change.
Data centres generate substantial heat, requiring sophisticated cooling systems that consume energy and large amounts of water. This puts a strain on local resources, particularly in areas with existing water scarcity. For example, Google uses an average of 4.3 billion gallons of water annually to power its data centres. The International Energy Agency predicts that global data centre energy use could double by 2026, largely due to AI and crypto mining.
Local impact in Chile
US tech giants like Amazon, Google, and Microsoft are building many of their new water-guzzling data centres in Chile, causing concern among local residents and activists following years of drought. The government encourages these tech investments through free trade agreements.
Activists are pushing for stricter laws and more comprehensive environmental impact studies to protect local water supplies and hold tech giants accountable. Some companies try to offset their impact by planting trees, but experts say these efforts aren’t enough to address the environmental damage.
Making compute more environmentally friendly
Despite these challenges, new technologies are making compute power more sustainable. These technologies include:
- Energy-efficient processors: Newer processors use less electricity, reducing overall energy consumption.
- Better cooling: Innovative cooling systems lower the energy needed to keep computers from overheating. For instance, Google says it “aims to” address its water consumption by replenishing 120% of the water it uses in communities where its data centres are located.
- Renewable energy: Google, Amazon, and Microsoft are investing in solar, wind, geothermal, and hydroelectric power to run their data centres.
Computers and their services have rapidly transformed our world. And platforms like Netflix and TikTok aren’t slowing down anytime soon. So, we all have to find ways to ensure that technologies like AI are used sustainably. This involves pressuring large corporations to adopt more sustainable practices and lobbying local politicians to hold big tech accountable.
Weekly disruptions
Weekly disruptions
Sony Will Use AI to Cut Film Costs, Says CEO Tony Vinciquerra (IndieWire) Sony Pictures Entertainment CEO Tony Vinciquerra announced that the company will use AI to reduce film production costs, aiming to make both theatrical and television content more efficiently. This move highlights Sony’s strong support for AI in filmmaking, despite potential controversies with creatives and ongoing labour negotiations around AI usage in Hollywood.
xAI Raises $6 Billion as Elon Musk Aims to Challenge OpenAI (Time) In a significant development in the AI industry, Elon Musk’s startup xAI secured $6 billion in Series B funding to further its competitive edge against OpenAI. This move underscores Musk’s ongoing commitment to shaping the future of artificial intelligence.
