Technology and Innovation Community

NVIDIA has just intoduced it's next-gen AI superchip - Vera Rubin AI - at it's annual GTC event.

The processor will succeed the Blackwell Ultra chip at the back end of 2026.

The Vera Rubin superchip combines CPU and GPU, Vera being the CPU and Rubin being the GPU.

The chip will provide 2x the overall performance of the Blackwell Ultra chip.

At the back end of 2027, it will introduce also the Vera Rubin Ultra superchip all this available in a large server rack system.

------------------------------
Todor Kostov
Director
------------------------------

Thanks for the update Todor,

For those willing to learn more about hardware, please see below a couple of articles published by CFA UK last year on this topic:

From Zero to Hero - A Data Scientist's Guide to Hardware Part 1 : Primer on CPU and GPU concepts for absolute beginners.

From Zero to Hero - A Data Scientist's Guide to Hardware Part 2  : Follow up focused on GPUs, Blackwell Chip and  a quick guide on how to leverage the power of local GPUs on your laptop when using deep learning libraries like Pytorch. 

Unfortunately, articles quickly become outdated due to the rapid pace of advancement in the chip industry.

------------------------------
Carlos Salas
Portfolio Manager & Freelance Investment Research Consultant
------------------------------

Thanks Carlos.

All great information and timely reminder about the fast pace of developments in that area.

Below just a quick high-level snapshot from Bloomberg on the hardware AI supply chain as we stand.

------------------------------
Todor Kostov
Director
------------------------------

Original Message:
Sent: 20-03-2025 09:19
From: Carlos Salas
Subject: NVIDIA $NVDA GTC 2025 - Vera Rubin superchip

Thanks for the update Todor,

For those willing to learn more about hardware, please see below a couple of articles published by CFA UK last year on this topic:

From Zero to Hero - A Data Scientist's Guide to Hardware Part 1 : Primer on CPU and GPU concepts for absolute beginners.

From Zero to Hero - A Data Scientist's Guide to Hardware Part 2  : Follow up focused on GPUs, Blackwell Chip and  a quick guide on how to leverage the power of local GPUs on your laptop when using deep learning libraries like Pytorch. 

Unfortunately, articles quickly become outdated due to the rapid pace of advancement in the chip industry.

------------------------------
Carlos Salas
Portfolio Manager & Freelance Investment Research Consultant
------------------------------

Furthermore, quite an insightfull interview provided by a16z recently with Jensen Huang and Arthur Mensch, cofounder and CEO of Mistral AI.

Headline topic: "Why Every Nation Needs Its Own AI Strategy"

Key discussion points:

• The Impact of AI on National Infrastructure
• Is AI a General-Purpose Technology?
• Why Nations Need to Engage AI
• AI and Cultural Infrastructure
• The Role of the Digital Workforce
• Specializing AI for National Needs
• Building a Digital Workforce
• Challenges and Risks of AI Adoption
• Is AI the Greatest Equalizer?
• Open-Source Models
• Collaborative AI Development
• Open-Source AI in Niche Markets
• Balancing Security and Openness
• Company Building Insights
• Navigating Competitive Partnerships
• Future Trends in Computing 

------------------------------
Todor Kostov
Director
------------------------------