Google leased nuclear reactors for AI processing centers

fog of war google research 2023

 

Google has announced plans to use small nuclear reactors to power its AI data centers. For this, the company cooperates with Kairos Power, which develops innovative reactors using molten salt as a heat carrier. The first reactor is planned to be commissioned by the end of the decade, and by 2035 the company intends to launch more such reactors.

 

The solution is due to the growing demand for powerful and sustainable energy resources, caused by the increased use of AI technologies. Nuclear power, which emits no carbon and provides a constant supply of electricity, allows technology companies such as Google, Amazon and Microsoft to reduce their carbon footprint and ensure the sustainability of their infrastructures.

 

Kairos Power specializes in small modular molten fluoride reactors, which are considered more flexible and efficient compared to traditional nuclear plants. These reactors offer a constant supply of energy unlike renewable sources such as solar or wind, which depend on weather conditions.


Don't miss interesting news

Subscribe to our channels and read announcements of high-tech news, tes

Leave a Reply

Your email address will not be published. Required fields are marked *





Articles & testsArticles

Best smartwatches for sports in 2025 Honor Watch 4 Pro

Let’s compare three interesting sports smartwatches: the Samsung Galaxy Watch 7 Ultra, Garmin Instinct 3, and Amazfit T-Rex 3 Pro.


What could be better than the iPhone 17? Comparing 5 Android smartphones this year Google Pixel Watch 2

If you are no longer a fan of the brand or want to experiment with Android smartphones, we can offer as many as five options.


NewsNews
| 09.48
Virtual Instruments has released transparent monitor  
Virtual Instruments

The Phantom monitor has a diagonal of 24 inches with a classic 16:9 aspect ratio and supports 4K resolution.

| 15.30
AI can’t fake toxicity – new Turing test
angry robot

Researchers from the University of Zurich, the University of Amsterdam, Duke University, and New York University have found that modern artificial intelligence language models can still be easily distinguished from humans.