🇮🇷 Iran Proxy | https://www.wikipedia.org/wiki/AI_datacenter
Jump to content

AI datacenter

Add links
From Wikipedia, the free encyclopedia

An AI datacenter (or artificial intelligence datacenter) is a specialized data center facility designed explicitly to support the high-performance computing (HPC) workloads required for the training and inference of artificial intelligence models.[1] Unlike traditional datacenters that host a variety of general-purpose computing tasks (like web services and databases), AI datacenters are optimized for the unique computational demands of machine learning, particularly deep learning.[2] [3] These facilities are characterized by extreme power density per rack (often exceeding 50–100 kW)[4], advanced liquid cooling systems, and low-latency, high-bandwidth networking fabrics to facilitate parallel processing. [5] [6] [7]

The rise of generative AI since 2022 has triggered a global boom in the construction of AI datacenters, making them a critical and strategically important piece of national infrastructure.[8] Companies like Microsoft, Google, Meta, and Amazon are investing tens of billions of dollars to build facilities containing over 100,000 AI accelerators each.[9] [10] [11] This massive build out is causing a resurgence in nuclear power plants.[12] In 2025 Google spent $95 billion on capex, with much of that for AI datacenters.[13][14] In 2025 U.S. tech companies spent $370 billion on capex with much of that spending for AI datacenters.[15] OpenAI wants to create a process for new datacenter expansion every week.[16][17]

By 2026, AI data centers are projected to consume over 90 TWh of electricity annually.[18] In the U.S. energy use is growing by 33% per year attributed to AI datacenter growth.[19] A startup company was created to harness nuclear energy for the AI datacenter boom.[20][21] The largest AI datacenter in 2025 cost $7 billion , and uses 300 MW of power—as much as 250,000 households.[22] Cornell University study estimates that the AI datacenter build out between 2024 and 2025 will contribute between 24 and 44 metric tons of addition CO2.[23]

The doubling of RAM and NAND prices has been attributed to the AI datacenter boom.[24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34]

The Trump administration is promoting the build out of AI datacenters.[35] U.S. president Trump hints at deregulation to promote AI datacenters.[36][37] There is local opposition.[38] [39]

History

[edit]

Early AI workloads in the 2010s were often run on general‑purpose high‑performance computing (HPC) clusters or small GPU servers in conventional data centers.[40] As deep learning models and datasets grew, cloud providers began to build dedicated infrastructures for AI training, including GPU clusters exposed through services such as Google Cloud TPU[41], Amazon EC2 P-series instances[42], and Microsoft Azure’s ND‑series virtual machines[43].

Around 2022–2024, the rapid adoption of large language models (LLMs) and generative AI led to a surge in demand for specialized AI datacenters with thousands or tens of thousands of accelerators connected through high‑speed fabrics.[44] Several technology companies announced multibillion‑dollar investments in new or expanded AI‑focused campuses, often near abundant power supply or renewable energy sources.[45] [46] [47]

Architecture

[edit]

AI datacenters are designed around clusters of accelerators optimized for parallel numerical computation. A typical facility includes: Compute: Large numbers of GPUs, TPUs, or other AI accelerators are deployed in high‑density racks. These devices are often grouped into “pods” or “nodes” that share local networking and storage and can be scaled out to thousands of accelerators for distributed training.

Networking: AI training jobs require high‑bandwidth, low‑latency communication to exchange gradients and parameters across devices. To support this, AI datacenters commonly use technologies such as InfiniBand, RDMA over Converged Ethernet, or proprietary interconnects. Storage and data pipelines: Training large models requires feeding vast datasets at high throughput. Control and orchestration: Software stacks manage job scheduling, resource allocation, and failure handling. Common AI frameworks include PyTorch and TensorFlow.

Technical architecture

[edit]
  • Compute Density: Traditional server racks typically consume 5–15 kW of power.[48] AI racks, utilizing hardware like Nvidia H100s or Google TPUs, consume 40 kW to over 100 kW per rack.[49]
  • Cooling Systems: Due to the heat generated by high-density chips, AI data centers often abandon traditional air cooling (CRAC units) in favor of liquid cooling technologies, such as direct-to-chip cooling or immersion cooling.[50]
  • Networking: AI training requires thousands of chips to communicate simultaneously. This necessitates specialized non-blocking network architectures rather than standard Ethernet used in web servers.[51]

OpenAI

[edit]

OpenAI's datacenter project is called Stargate.[52] The partnership includes Softbank, MGX and Oracle. In September 2025, it was announced the building of 5 new AI datacenters.[53] The new sites are in Texas, New Mexico, Wisconsin, and Ohio. Capacity is estimated to be 6.5 gigawatts and $400 billion investment over 3 years. OpenAI's revenue for 2025 was estimated to be less than $12 billion.[54]

The impact of these new AI datacenters is sparking concern.[55]

See also

[edit]

References

[edit]
  1. ^ elongated_musk (2025-02-18). "How to build an AI Datacentre — Part 1 (Cooling and Power)". Medium. Archived from the original on 2025-02-24. Retrieved 2025-12-09.
  2. ^ Lee, Danny H. (2025-05-15). "AI vs Traditional: Data Centers". Medium. Retrieved 2025-12-09.
  3. ^ Cloud, Cyfuture. "what is ai data center". cyfuture.cloud. Retrieved 2025-12-09.
  4. ^ "1,000 homes of power in a filing cabinet - rising power density disrupts AI infrastructure". www.goldmansachs.com. Archived from the original on 2025-09-15. Retrieved 2025-12-09.
  5. ^ Tithi, Jesmin Jahan; Wu, Hanjiang; Abuhatzera, Avishaii; Petrini, Fabrizio (2025-09-05), Scaling Intelligence: Designing Data Centers for Next-Gen Language Models, arXiv:2506.15006, retrieved 2025-12-10
  6. ^ Garcia, Heidi. "Optimize AI Fabric Design for Efficient LLM Model Training". www.keysight.com. Retrieved 2025-12-10.
  7. ^ "AI data center". F5, Inc. Retrieved 2025-12-09.
  8. ^ "US data center build hits record as AI demand surges, Bank of America Institute says". Reuters.
  9. ^ "From OpenAI to Google, firms channel billions into AI infrastructure as demand booms". Reuters.
  10. ^ Carvão, Paulo. "Why OpenAI's AI Data Center Buildout Faces a 2026 Reality Check". Forbes.
  11. ^ https://www.fastcompany.com/91450218/ai-data-center-boom-unexpected-winner
  12. ^ Goudarzi, Sara (2025-12-05). "When it all comes crashing down: The aftermath of the AI boom". Bulletin of the Atomic Scientists. Retrieved 2025-12-09.
  13. ^ "Google boosts AI spending again as cloud unit soars | CIO Dive". www.ciodive.com. Retrieved 2025-12-09.
  14. ^ Cooper, Ian (2025-12-08). "AI Data Centers are Booming and These 3 Stocks Are Cashing In". 24/7 Wall St. Retrieved 2025-12-09.
  15. ^ "The AI Data Center Boom Is Warping the US Economy | Medial". medial.app. Retrieved 2025-12-09.
  16. ^ "Abundant Intelligence". Sam Altman. Retrieved 2025-12-09.
  17. ^ "Will OpenAI Really Build 60 Football Fields Worth of AI Infrastructure Per Week?". PCMAG. 2025-09-23. Retrieved 2025-12-09.
  18. ^ "How AI Is Forcing a Rethink of Data Center Power". DataCenterKnowledge. Retrieved 2025-12-09.
  19. ^ "AI Data Center Building Spree Hits $40 Billion in a Single Month". USFunds. Retrieved 2025-12-09.
  20. ^ "Wall Street really likes Rick Perry's nuclear-powered data center company". Texas Standard. Retrieved 2025-12-09.
  21. ^ "Gigawatt Scale Power for Next-Gen AI | Fermi America". 2025-06-24. Retrieved 2025-12-09.
  22. ^ "Trends in AI Supercomputers". arxiv.org. Retrieved 2025-12-09.
  23. ^ Xiao, Tianqi; Nerini, Francesco Fuso; Matthews, H. Damon; Tavoni, Massimo; You, Fengqi (2025-11-10). "Environmental impact and net-zero pathways for sustainable artificial intelligence servers in the USA". Nature Sustainability: 1–13. doi:10.1038/s41893-025-01681-y. ISSN 2398-9629.
  24. ^ Anton Shilov (2025-12-01). "The RAM pricing crisis has only just started, Team Group GM warns — says problem will get worse in 2026 as DRAM and NAND prices double in one month". Tom's Hardware. Retrieved 2025-12-09.
  25. ^ Butler, Sydney (2025-11-10). "RAM prices have doubled, here's my plan to survive the 'RAM-pocalypse'". How-To Geek. Retrieved 2025-12-09.
  26. ^ "Why is RAM so expensive right now?". NZXT. Retrieved 2025-12-09.
  27. ^ www.bacloud.com https://www.bacloud.com/assets/error-pages/403.html. Retrieved 2025-12-09. {{cite web}}: Missing or empty |title= (help)
  28. ^ "SSD & RAM Price Surge Guide: Why Costs Are Rising in 2025". ACEMAGIC. Retrieved 2025-12-09.
  29. ^ Gamers Nexus (2025-11-16). RAM: WTF?. Retrieved 2025-12-09 – via YouTube.
  30. ^ Stephen Warwick (2025-11-04). "Bewildered enthusiasts decry memory price increases of 100% or more — the AI RAM squeeze is finally starting to hit PC builders where it hurts". Tom's Hardware. Retrieved 2025-12-09.
  31. ^ "'This Is Insanity': DDR RAM Prices Soar Due to AI Demand". PCMAG. 2025-10-23. Retrieved 2025-12-09.
  32. ^ "What's going on with RAM?". Windows Central. 2025-12-09. Retrieved 2025-12-09.
  33. ^ Roth, Emma (2025-12-09). "RAM price hikes: the latest on the global memory shortage". The Verge. Retrieved 2025-12-09.
  34. ^ Jason England (2025-12-04). "RAM prices are exploding — here's why and everything you need to know about surviving RAMageddon". Tom's Guide. Retrieved 2025-12-09.
  35. ^ "Trump's push for more AI data centers faces backlash from his own voters". Reuters.
  36. ^ "Donald J. Trump (@realDonaldTrump)". Truth Social. Retrieved 2025-12-09.
  37. ^ Metzger, Bryan. "Trump says he'll sign an executive order restricting states' ability to regulate AI". Business Insider. Retrieved 2025-12-09.
  38. ^ "Trump's push for more AI data centers faces backlash from his own voters". Reuters.
  39. ^ Editor (2025-08-16). "Local Communities Rise Up Against Massive AI Data Centers". Corporate Crime Reporter. Retrieved 2025-12-09. {{cite web}}: |last= has generic name (help)
  40. ^ Coates, Adam; Huval, Brody; Wang, Tao; Wu, David; Catanzaro, Bryan; Andrew, Ng (2013-05-26). "Deep learning with COTS HPC systems". Proceedings of the 30th International Conference on Machine Learning. PMLR: 1337–1345.
  41. ^ Jouppi, Norman P.; Young, Cliff; Patil, Nishant; Patterson, David; Agrawal, Gaurav; Bajwa, Raminder; Bates, Sarah; Bhatia, Suresh; Boden, Nan (2017-04-16), In-Datacenter Performance Analysis of a Tensor Processing Unit, arXiv:1704.04760, retrieved 2025-12-09
  42. ^ "Introducing Amazon EC2 P2 Instances, the largest GPU-Powered virtual machine in the cloud - AWS". Amazon Web Services, Inc. Retrieved 2025-12-09.
  43. ^ "Azure Deep Learning and ND, NC series with NVidia | Microsoft Community Hub". TECHCOMMUNITY.MICROSOFT.COM. Archived from the original on 2025-04-17. Retrieved 2025-12-09.
  44. ^ "NVIDIA Hopper GPUs Expand Reach as Demand for AI Grows". investor.nvidia.com. Retrieved 2025-12-10.
  45. ^ "Meta's new AI supercomputer: 16,000 x GPUs, insane 175PB bulk storage". TweakTown. 2022-01-26. Retrieved 2025-12-10.
  46. ^ Freund, Karl. "Meta Builds World's Largest AI Supercomputer With NVIDIA For AI Research And Production". Forbes. Retrieved 2025-12-10.
  47. ^ "Eagle (Microsoft)". Glenn's Digital Garden. 2024-12-05. Retrieved 2025-12-10.
  48. ^ "How Much Electricity Does A Data Center Use? 2025 Guide". 2025-10-02. Retrieved 2025-12-09.
  49. ^ "1,000 homes of power in a filing cabinet - rising power density disrupts AI infrastructure".
  50. ^ "The Evolution of Data Centers: From Traditional to AI-Powered". hexatronicdatacenter.com. Retrieved 2025-12-09.
  51. ^ Wang, Bill (2025-10-04). "Why AI Training Needs Special Networks". Medium. Retrieved 2025-12-09.
  52. ^ "Announcing The Stargate Project". openai.com. Retrieved 2025-12-10.
  53. ^ "OpenAI, Oracle, and SoftBank expand Stargate with five new AI data center sites". openai.com. 2025-12-09. Retrieved 2025-12-10.
  54. ^ "OpenAI's Data Center Expansion: A Strategic Shift Fueling AI Dominance". Ainvest. Retrieved 2025-12-10.
  55. ^ published, Jowi Morales (2025-11-13). "OpenAI's colossal AI data center targets would consume as much electricity as entire nation of India — 250GW target would require 30 million GPUs annually to ensure continuous operation, emit twice as much carbon dioxide as ExxonMobil". Tom's Hardware. Retrieved 2025-12-10.
[edit]
Retrieved from "https://en.wikipedia.org/w/index.php?title=AI_datacenter&oldid=1326746818"