- The Renaissance Times
- Posts
- Stargate: The $100 Billion Supercomputer
Stargate: The $100 Billion Supercomputer
Dear Artisan,
Every day, our world gets closer to resembling a sci-fi novel.
This weekend, reports surfaced that Microsoft is building a series of data centers that will be used to run OpenAI’s software. The cost to build these out starts at $10 billion for the first one and is reportedly set to reach $100 billion for the largest center, Stargate.
Let’s dive in.
Hot Off The Press
Microsoft and OpenAI are planning a $100 billion data-center project
NYC’s business chatbot is reportedly doling out ‘dangerously inaccurate’ information
OpenAI created a voice cloning tool that can replicate human emotion
The US House banned staffers from using Microsoft Copilot
number 5
The One Big Thing
A few months ago, rumors started swirling that Sam Altman was trying to raise $7 trillion to fund chip fabrication plants and data centers.
Well, it turns out that’s not quite true, and neither is the Saudi money that was supposed to back it. Instead, OpenAI and Microsoft are planning to team up and build this in-house, according to a report that dropped over the weekend.
Details are thin at this point, and the project itself is just a rumor. However, the way things look, Microsoft will be funding this project and contributing up to $100 billion specifically for data centers over the next 5 years. The first of these is set to be built in Wisconsin by 2026, and the largest data center, Stargate, will be built by 2028.
This makes a lot of sense from both companies’ points of view. Microsoft gets to expand up the stack, owning not only the most powerful foundation models courtesy of OpenAI, but now looking to own all the compute that runs it.
The Stargate supercomputer, as they are calling it, could be the thing that gives Microsoft its biggest competitive edge over every other company out there. The same way that NVIDIA is now releasing their own foundation models, Microsoft is integrating the entire AI stack under its roof.
The race to AGI is on and getting heated.
The Gallery
THE GREAT CIRCLE
— Henry Daubrez 🌸💀 (@henrydaubrez)
11:00 AM • Apr 1, 2024
Tools
Must have tools for every artisan to add to their toolkit:
LM Studio: A tool to run LLMs offline locally for private data processing
Prototyper: A tool to convert text descriptions into live code for rapid UI prototyping
MathHandwriting: An API to convert handwritten mathematical equations into LaTeX code
VoiceTrans: A tool to transform voices into various characters and celebrities
ACE Studio: A software to generate synthetic singing
Deep Tech
The newest and coolest in the research world that you need to know about:
Localizing Paragraph Memorization in Language Models
Transformer-Lite: High-efficiency Deployment of Large Language Models on Mobile Phone GPUs
MambaMixer: Efficient Selective State Space Models with Dual Token and Channel Selection
DiJiang: Efficient Large Language Models through Compact Kernelization
Google announces Gecko: Versatile Text Embeddings Distilled from Large Language Models
Closing Thought
Who will win the great AGI war of 2040, NVIDIA or MSFT?
Work With Us!
The AI Renaissance is coming and we are building the best community of the people making it happen.
Contact us to sponsor your product or brand and reach the exact audience for your needs across our newsletter and podcast network.