#secsolutionforum2026 #BolognaFiere #datacenter #droni #videosorveglianza http://dlvr.it/TRdvzv
Когда вычисления станут бесконечными: разбор интервью Илона Маска от 5.02.26
Маск говорит, что самое дешёвое место для размещения ИИ скоро будет в космосе. Звучит как очередной футуристический разгон, но за ним прячатся вполне земные проблемы: вычисления, память, энергия и пределы текущей инфраструктуры. В статье разбираю, какие из этих тезисов подтверждаются данными, а где футуризм уже начинает работать отделом продаж.
https://habr.com/ru/articles/1013460/
#elon_musk #ai_infrastructure #gpu #hbm #datacenter #energy #memory #spacex #inference
𝗢𝗥𝗔 𝗦𝗨 𝗣𝗥𝗘𝗦𝗔 𝗗𝗜𝗥𝗘𝗧𝗧𝗔 𝗥𝗔𝗜 𝟯: "𝗥𝗜𝗩𝗢𝗟𝗨𝗭𝗜𝗢𝗡𝗘 𝗔𝗥𝗧𝗜𝗙𝗜𝗖𝗜𝗔𝗟𝗘", 𝗖𝗢𝗦𝗔 𝗖'𝗘' 𝗗𝗜𝗘𝗧𝗥𝗢 𝗟'𝗔𝗜
Un'analisi a tutto campo sulla #tecnologia che sta cambiando le nostre vite, anche in modi che non ti aspetti.
𝗚𝘂𝗮𝗿𝗱𝗮 𝘀𝘂 𝗥𝗮𝘆𝗣𝗹𝗮𝘆:
https://www.raiplay.it/dirette/rai3
#AI #IA #ambiente #clima #energia #territorio #sviluppo #sostenibilita #datacenter #speculazione
Power failures cause 54% of data center outages.
Not servers. Not software. Not networks. The electrical system.
I analyzed how AWS, Google, Microsoft, and xAI each solve this differently. Four philosophies, four architectures.
Technical deep dive:
https://resistancezero.com/article-13.html
How xAI powered Colossus: 2 GW — 40% of Memphis's entire grid capacity. For one facility.
AWS eliminated floor-level UPS, moved to rack-level. 35% less distribution loss.
Google: 100 million Li-ion cells, 1.09 PUE.
Power architecture defines everything.
Power failures cause 54% of data center outages. So how do AWS, Google, Microsoft, and xAI actually design their power systems?
AWS: rack-level UPS, 35% less loss
Google: 100M Li-ion cells, 1.09 PUE
xAI Colossus: 2 GW -- 40% of Memphis grid
Full technical deep dive:
https://resistancezero.com/article-13.html
#DataCenter #PowerDistribution #AI #CriticalInfrastructure #Engineering
Power failures cause 54% of data center outages.
Not servers. Not software. Not networks. The electrical system.
I analyzed how AWS, Google, Microsoft, and xAI each solve this differently. Four philosophies, four architectures.
Technical deep dive:
https://resistancezero.com/article-13.html
How xAI powered Colossus: 2 GW — 40% of Memphis's entire grid capacity. For one facility.
AWS eliminated floor-level UPS, moved to rack-level. 35% less distribution loss.
Google: 100 million Li-ion cells, 1.09 PUE.
Power architecture defines everything.
Power failures cause 54% of data center outages. So how do AWS, Google, Microsoft, and xAI actually design their power systems?
AWS: rack-level UPS, 35% less loss
Google: 100M Li-ion cells, 1.09 PUE
xAI Colossus: 2 GW -- 40% of Memphis grid
Full technical deep dive:
https://resistancezero.com/article-13.html
#DataCenter #PowerDistribution #AI #CriticalInfrastructure #Engineering