How Grok 3 Just Changed The Internet.
The release of xAI’s Grok in 2023 was thought to be a novelty. Most of the standard folks that make a comment were certain that the company could not possibility catch up to “OpenAI”. It is true that the first versions of Grok the model tried to be humorous and succeeded as much as it may have failed for many. It was certainly not a “serious” AI platform. Well on February 17th, 2025 at about 8:30 PM PST that equation changed.
Not only did Grok 3 meet up to the hype generated by Elon Musk and others (including me), it surpassed it in many ways. This all came about through an impossible accelaraton that only few companies could ever think to achieve. Let’s look at the timeline:
- Grok-1: 3 November 2023
- Grok on X for Premium+ users: 7 December 2023
- Grok-1 open-source release: 17 March 2024
- Grok-1.5: 15 May 2024
- Grok-2: 13 August 2024
- Aurora : 4 November – 2024
- Grok-3 : 17th February 2025
WARNING, THERE IS PAID CONTENT AT THE BOTTOM. STOP NOW IF THIS IS NOT FOR YOU. YOUR SUPPORT PAYS FOR MY WORK, THANK YOU.
Grok Growth
All of this growth came about because X.AI is by amassing one of the largest collection of GPUs in the world called Colossus, a supercomputer built by xAI, plays a pivotal role in the development of Grok 3. Here are the overview:
Computational Power for Training
Colossus provides the immense computational resources required to train Grok 3. It was initially equipped with 100,000 Nvidia H100 GPUs and now doubling to over 200,000 H100 equivalents moving to over 1 million by years’s end, offering the processing power necessary to handle the complex algorithms and massive datasets involved in developing advanced AI models. This scale of infrastructure is critical for Grok 3, enabling it to process and learn from vast amounts of data efficiently.
Rapid Development and Scalability
The construction of Colossus was remarkably fast, reportedly taking just a few months. This rapid deployment allowed xAI to accelerate the training process for Grok 3, speeding up development compared to traditional timelines for such large-scale systems. Moreover, Colossus is designed to be scalable. This scalability ensures that as Grok evolves and demands more computational power, Colossus can grow to meet those needs.
Advanced Networking and Data Management
Beyond raw compute, Colossus incorporates cutting-edge networking technologies, such as Nvidia’s Spectrum-X Ethernet platform, to manage data transfers efficiently across its vast cluster. This minimizes latency and supports the high-bandwidth requirements of training large models like Grok 3. Additionally, while specific details are not public, Colossus likely features enormous storage capacity to handle the extensive training data and model parameters essential for Grok’s development.
Grok 3 is still in training and my bet is it will constantly be in training until the newer versions are being released. This means you will access a “new” Grok 3 just about every day. But there is more to this. Grok 3 is far more than an AI platform competing with ChatGPT, it is an entirely new way to understand and access the Internet. Some companies will fade into the long goodbye mode.
In this multipart member only series I will explore this in detail with insights not seen anyplace else. You will be the first to understand this shift and profit by it. If you are a member, thank you. If you are not yet a member, join us by clicking below.
“Technology does not die with dignity; it fades into irrelevance”-Adapted from Ernest Hemingway
Become a member today for full access to this exclusive research.
🔐 Start: Exclusive Member-Only Content.
The rest of this article is for Multiplex members
You’ve reached the members-only part. Log in to keep reading right here, or join to unlock it.
- Full member-only articles: insights, wisdom and executive summaries
- The Multiplex members forum
- Member benefits and discounts
Already a member? You’ll come right back to this article after logging in. Forgot password?
Membership status: Not logged in
🔐 End: Exclusive Member-Only Content.
~—~
~—~
~—~
Support ReadMultiplex at Multiplex.Tips
Support ReadMultiplex
Subscribe or donate by Bitcoin.
Copy address:bc1qkufy0r5nttm6urw9vnm08sxval0h0r3xlf4v4x
Send your receipt to [email protected] to confirm subscription.
Members
Become a member for full member-only articles, the Multiplex members forum, and member benefits and discounts.
Past, present and future technology research + insightsListen
🔊 Every article has a free “Listen to this ►” button. Hear it read aloud in your browser.
The Multiplex
- Advertise with us: Rates and booking
- X: @ReadMultiplex
- Buy a T-shirt: Multiplex.Love
- Program Guide: Multiplex.Today
- Podcast: Multiplex.Life
- Join the Multiplex team: Multiplex.Team
Stay updated
Get an email when we post new articles:
Great post! Makes sense when it comes to the new, up to date information. But what about the forgotten internet – the pages disappearing every day, only some of them saved in the Internet Archive/Wayback Machine. I recently found some of my old bookmarks from about 10 years ago, many of them no longer work.
Brian. Interesting premise as usual. Google is unable to innovate, I have first hand knowledge and can confirm. However a bit disappointed that most of this post is AI generated slop. Please take this as constructive feedback — want to understand what you think and why, not read the unmistakeable and overly mechanistic verbiage of a LLM. Cheers