Showing posts sorted by relevance for query exabyte. Sort by date Show all posts
Showing posts sorted by relevance for query exabyte. Sort by date Show all posts

Monday, November 2, 2026

Memory -- DRAM Capacity -- Update -- Exabytes And Zettabytes -- November 2, 2026

Locator: 52073AI.  

Shay, Google's backlog -- link. Link here: https://x.com/StockSavvyShay/status/2117239774461444126.  

Shay: link here --  

Query: see above -- UBS projects dram capacity to grow ~147% by 2030 reaching ~89 eb. What is eb?

Link here. Or here. Link: https://share.google/aimode/m49qCEbRcrhtPVCrp. 

In a word search of this blog, which has been around since 2009, and in which I talk extensively about tech, I have only used the word exabyte or exabytes twice, once with regard to NSA, once in 2025, and once in 2013: https://themilliondollarway.blogspot.com/search?q=exabyte.  


 

***************************
Google

Link here. 

Monday, July 14, 2025

AI: META -- Mark Zuckerberg "ALL IN" -- More To Follow -- We'll Have The Story Later Today -- But This Is Huge

Locator: 48732META.

This is to serve Mark Zuckerberg's "own interests."

Link at Reuters.

  • Meta building several multi-gigawatt AI data centers
  • company has formed Superintelligence Labs to unify AI efforts
  • CEO Zuckerberg has been leading a talent war for top engineers
  • one of Meta's AI superclusters will rival Manhattan in size -- are investors even paying attention
July 14 (Reuters) - Mark Zuckerberg said on Monday that Meta Platforms would spend hundreds of billions of dollars to build several massive AI data centers for superintelligence, intensifying his pursuit of a technology he has chased with a talent war for top engineers.

Saying again

  • multi-gigawatt ...
  • one supercluster will rival Manhattan i size ---
  • that's a lot of blades

********************************
Review

Tech:

Type of AI data center:

  • small (edge / enterprise): ~ 100 TB = 0.1 mega GB
  • medium (cloud region): ~ 1 - 10 PB - 1.0 mega GB
  • hyperscale (e.g., AWS, MSFT, Google, META, Nvidia): 10 - 100+ BP = 10 - 100 mega BG

1 PB (petabyte) = one million BB = 1 mega GB
1EB (exabyte) = 1,000 PB = 1 billion BG

Real-world examples (2024 - 2025):

  • Nvidia's Blackwell-powered AI supercomputer
    • tens of thousands of GPUs (Nvidia)
    • requires 50 - 100+ PB storage for training LLMs
  • OpenAI training GPT-4:
    • estimates hundreds of petabytes of data processed across computer clusters
  • Meta's AI Research SuperCluster:
    • 16,000 GPUS (Nvidia), over 1 EB (billion GP) of storage projected when fully deployed

Bottom line:

  • The average modern AI-focused data center—especially for large models—easily exceeds 1 million GB (1 PB) of storage, with hyperscalers using tens to hundreds of petabytes.

Tuesday, October 8, 2013

Why Apple Is Committed To 100% Renewable Energy

One of the polls at the sidebar asks the question why Apple is committed to renewable energy.

The results of the poll:
  • it is the "right" thing to do: 5%
  • it makes economic sense: 8%
  • it is good public relations: 67%
  • there is another over-riding reason: 19%
Generally, I just provide the results of the poll and do not discuss the results. In this case, I make an exception.

Steve Jobs was famous for many things. He always talked about "skating to where the puck would be." In this case, he knew the electric grid was unreliable and for iCloud he absolutely needed a reliable source of electricity. The US electric grid is not reliable and with the war on coal and the war on conventional energy in general, he knew the electric grid would only get worse.

In addition, Steve Jobs was famous for being a control freak. He wanted to control everything that had to do with his company, including controlling his source of power.

Today, the headline story on the front page of The Wall Street Journal: meltdowns hobble NSA data center.
Chronic electrical surges at the massive new data-storage facility central to the National Security Agency's spying operation have destroyed hundreds of thousands of dollars worth of machinery and delayed the center's opening for a year.
There have been 10 meltdowns in the past 13 months that have prevented the NSA from using computers at its new Utah data-storage center, slated to be the spy agency's largest, according to project documents reviewed by The Wall Street Journal.
One project official described the electrical troubles—so-called arc fault failures—as "a flash of lightning inside a 2-foot box." These failures create fiery explosions, melt metal and cause circuits to fail, the official said.
Exactly how much data the NSA will be able to store there is classified. Engineers on the project believe the capacity is bigger than Google's largest data center. Estimates are in a range difficult to imagine but outside experts believe it will keep exabytes or zettabytes of data. An exabyte is roughly 100,000 times the size of the printed material in the Library of Congress; a zettabyte is 1,000 times larger.
But without a reliable electrical system to run computers and keep them cool, the NSA's global surveillance data systems can't function. The NSA chose Bluffdale, Utah, to house the data center largely because of the abundance of cheap electricity. It continuously uses 65 megawatts, which could power a small city of at least 20,000, at a cost of more than $1 million a month, according to project officials and documents.