As power needs of AI push emissions up, tech companies explore experimental clean energy projects such as atomic fusion and futuristic geothermal power plants
As power needs of AI push emissions up and put big tech in a bind, companies put their faith in elusive — some say improbable — technologies.
Washington Post
Context & Ripple Effects
AI’s expanding electricity demand is turning energy supply and emissions exposure into operating constraints for technology companies, rather than a peripheral sustainability issue. This story captures an early response: interest in clean-power options whose commercial readiness remains uncertain.
Technology companies must weigh potentially long-duration fusion and advanced geothermal projects as possible answers to AI-driven power demand and rising emissions.
The companies’ clean-energy commitments face a nearer-term credibility challenge because the technologies highlighted are experimental rather than assured sources of power.
Second-order effects
Demand for reliable, low-carbon electricity increases the strategic value of established nuclear generation and creates a stronger incentive for utilities and plant owners to negotiate directly with data-center customers.
Competitors that cannot secure dependable clean power may face a harder trade-off between expanding AI capacity and containing emissions; later coverage of fossil-fuel use for data centers illustrates the fallback risk.
Third-order effects
Power availability is likely to become a material determinant of where and how quickly AI infrastructure can be built, shifting competition from chips and models toward generation, grid access, and project execution.
If experimental clean-energy supply does not mature on data-center timelines, the sector could see a more durable split between long-term decarbonization ambitions and the energy sources used to sustain near-term AI growth.
The trend: AI infrastructure is increasingly being planned as utility infrastructure, with reliable power supply emerging as a gating input to compute expansion.
Good piece on how the AI boom is increasing fossil fuel consumption, hurting the fight against climate change. The piece nicely undermines Silicon Valley PR about meeting the increased power demands with clean energy. https://www.washingtonpost.com/ ...
This is almost exactly what bitcoiners have been saying in recent years in response to crypto mining stressing grids and prolonging the lives of fossil fuel plants https://www.washingtonpost.com/ ... [image]
Microsoft is betting on an effort to generate power from atomic fusion, which has eluded scientists for a century. Other companies are hoping to use small nuclear reactors or geothermal energy to power the AI revolution. https://www.washingtonpost.com/ ...
This is a major problem, and the only solutions being proposed by industry are moonshot fusion bets and pouring more resources into developing an AGI that'll somehow reverse all this. https://www.washingtonpost.com/ ... [image]
The demand for data centers is hurting efforts to fight climate change by keeping demand for coal power plants and also straining power grids. AI leaders like Sam Altman and Microsoft have talked about building more nuclear power plants as a solution. https://www.washingtonpost.c…
The artificial intelligence industry is driving a nationwide data center building boom. Those data centers will require massive amounts of power — and Big Tech is going all in on experimental clean-energy projects that have long odds of success. https://www.washingtonpost.com/ ..…
AI would consciously decide to eliminate humankind, they said. — It was an existential threat, they said. — They were right, but for the wrong reasons. [embedded post]
New AI data centers are coming online so fast that the electricity demand is straining global power grids and threatening clean energy goals https://www.bloomberg.com/... via @bbgvisualdata [image]
What's alarming is that the growth in electricity demand from data centers is such that it could soon equal or exceed some countries' renewable energy supply. In Ireland, data center's energy usage was already equal to 53% of its renewable energy supply in 2022. [image]
The rush to build and distribute AI products from global data centers is wreaking havoc with power systems “I don't think we can move that much electricity around the globe, forget about generating it,” says Ali Farhadi, CEO of the Allen Institute for AI. https://www.bloomberg.co…
🚨NEW🚨: AI Is Wreaking Havoc on Global Power Systems We analyzed data on 1000s of data centers and found that AI data centers are coming online so fast that electricity demand is straining global power grids and threatening clean energy goals 🎁 https://www.bloomberg.com/... [video…
Because of their huge electricity capacity, data centers are also HUGE, literally, some of these buildings are now the size of 130 soccer fields or TWICE the size of the mall of America (the biggest shopping mall in the US with a rollercoaster inside). [video]
Read The Big Take: New AI data centers are coming online so fast that the electricity demand is straining global power grids and threatening clean energy goals. 🔗⚡️🤖: https://www.bloomberg.com/... [video]
The capital intensive nature of generative AI dwarfs the costs of prior software waves like the internet or smartphones. It's not just paying an arm and a leg to Nvidia for GPUs but also needing to build data centers with huge power requirements. https://www.bloomberg.com/...
Following the Virginia, US, model, South East Asia is rapidly becoming a new capital for data centers, especially Malaysia, where thousands of Megawatts of capacity are under development. [image]