• 80/20 AI
  • Posts
  • 100+ Countries Get Google’s AI

100+ Countries Get Google’s AI

Advertise here | 12-min Read

Welcome to our Wednesday edition!

In today’s menu:

  • AI Inspirational Quote

  • Google Expands AI Summaries

  • Inside Elon Musk’s 100K GPU Colossus!

  • OpenAI to launch its own AI chip by 2026

  • Top Rated AI Tools

  • Byte Sized news

  • How to Bring any Character Image to Life

AI INSPIRATIONAL QUOTE

When you have a dream
You've got to grab it and never let go.

LATEST NEWS

Google Expands AI Summaries

Google is rolling out its AI Overviews feature to over 100 more countries.

  • Now covering places like Canada, Australia, South Africa, Colombia, and Nigeria—though Iran is still excluded.

  • This update means users can get AI-powered summaries in multiple languages based on where they are, with options like English, Hindi, Indonesian, Japanese, Portuguese, and Spanish.

  • For now, ads in these AI Overviews are limited to mobile users in the U.S. After first launching in the U.S. in May, Google’s gradually expanding this feature worldwide.

Google wants to make searches faster and more useful with AI summaries, even though there have been some accuracy issues early on. Despite these bumps, Google’s pushing ahead, seeing AI as the future of search.

Learn AI With 5 min/day

Learn AI in 5 Minutes a Day

AI Tool Report is one of the fastest-growing and most respected newsletters in the world, with over 550,000 readers from companies like OpenAI, Nvidia, Meta, Microsoft, and more.

Our research team spends hundreds of hours a week summarizing the latest news, and finding you the best opportunities to save time and earn more using AI.

Inside Elon Musk’s 100K GPU Colossus!

Elon Musk has lots of unique and high-cost projects, and the AI supercluster Colossus xAI is one of the latest additions. Recently, new details about this project have surfaced in the media.

The owner of the YouTube channel ServeTheHome recently got permission to visit Colossus xAI. This supercluster consists of 100,000 GPUs and officially went online about two months ago, taking 122 days to construct.

In a video released by ServeTheHome, viewers get a close-up look at various aspects of this AI supercomputer.

However, specific details, like its energy consumption, remain confidential under a non-disclosure agreement. Additionally, xAI requested that certain sensitive parts be blurred, though the final video does show key components like the Supermicro GPU servers.

The Colossus xAI supercomputer is built using Nvidia’s HGX H100 GPUs, with each server containing eight H100 GPUs.

Previously, there were reports of Elon Musk ordering AI servers from Nvidia for this project.

The HGX H100 platform is encased in a Supermicro cooling system that provides replaceable liquid coolant for each GPU. The servers are arranged in racks that hold eight servers each, with a total of 64 servers per row.

Additionally, 1U manifolds are placed between each HGX H100 to supply the necessary coolant, and at the bottom of each rack is a 4U Supermicro unit with an extra liquid pump and rack monitoring system.

Each row groups eight racks together, allowing for up to 512 GPUs in each row. Every server has four backup power supplies. Behind the GPU racks are three-phase power sources, Ethernet switches, and a manifold to supply coolant.

Overall, Colossus xAI has over 1,500 GPU racks. Nvidia CEO Jensen Huang previously mentioned that installing this many servers took only three weeks, calling Musk’s endeavor "superhuman."

Finally, each GPU is equipped with a dedicated 400 GB NIC (Network Interface Controller) and an additional 400 GB NIC per server, providing each HGX H100 server with a total of 3.6 Tbps of Ethernet bandwidth.

OpenAI to launch its own AI chip by 2026

OpenAI is taking big steps in hardware, reportedly partnering with Broadcom and TSMC to create its first custom AI chip. At the same time, it’s expanding its hardware supply to include AMD, adding another option alongside NVIDIA, as reported by Reuters.

  • To tackle this ambitious project, OpenAI has assembled a team of around 20 engineers, including former Google TPU designers, with the goal of launching the custom chip by 2026.

  • At one point, OpenAI considered setting up its own chip factories, but ultimately chose to work with Broadcom for the design and TSMC for manufacturing.

  • Besides the custom chips, OpenAI plans to start using AMD’s MI300X processors in its training systems. This addition is meant to lessen its dependence on NVIDIA GPUs, which have been its primary choice.

  • The backdrop to these moves is OpenAI’s mounting compute costs, which could lead to losses of up to $5 billion this year, despite an anticipated $3.7 billion in revenue.

With the AI industry booming, the demand for specialized chips has surged, pushing tech companies to rethink their hardware strategies. As AI scales up, having the right hardware could become a major edge—not just the models themselves.

Top Rated

bolt new: Prompt, run, edit & deploy full-stack web apps

Prelude Verify: Verify more users for less cost with our spam-free SMS API

Langtail 1.0: The low-code platform for testing AI apps

Shoppable widget: Sell more with user-generated content.

AI Studios: Effortless AI-powered video creation with avatars

8020AI Picks

Space: Tamarack Global closes $72M fund for space investments, backing projects like Impulse Space’s Series B and focusing on aerospace, robotics, and energy tech.
Read more…

Crypto: Indian crypto exchange CoinDCX launched a decentralized custody solution, allowing 15M users to control crypto assets with added security through two-factor authentication, aiming to reduce risks seen in the recent WazirX hack.
Read more…

Gaming: Ghostface joins Mortal Kombat 1 as a guest fighter on November 19 (early access) and November 26 for all players, alongside Conan, T-1000, and returning MK favorites Cyrax, Sektor, and Noob Saibot in the new 'Khaos Reigns' expansion.
Read more…

Sports: Ruben Amorim faces a huge challenge at Manchester United as he inherits a struggling squad left by Erik ten Hag, with United sitting 14th after just three wins in nine games.
Read more…

Don't miss: Google Cloud's AI-driven revenue surge boosts competition with Amazon and Microsoft as demand for cloud AI services spikes.
Read more…

How to Bring any Character Image to Life

image source: Joey /// VP Land

RunwayML’s Act-One lets you animate any image character in real time using a real actor’s performance. Here’s how to get started:

  1. Log in on the RunwayML website.

  2. Select "Try Act-One" and pick Gen-3 Alpha as your model, then find Act-One in the left toolbar.

  3. In the Act-One window’s top section, upload or select a video for your driving performance. Ensure it's forward-facing, even if your character image is angled differently.

  4. In the bottom section, choose a character image—either use a preset or go to the Custom tab to upload your own.

  5. Click Generate and wait a few seconds to see your animated video. (Note: Subscription required)

FEEDBACK

Help us improve the newsletter for you.

How was 8020AI today?

Login or Subscribe to participate in polls.

Login or Subscribe to participate in polls.

If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.

👋 THAT’S A WRAP

SPONSOR US

Get your busines in front of over 55k+ AI professionals

Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Get in touch today.