LEO says 61: NVIDIA wants to RULE the world

Leo is back and he has a lot to say about Nvidia. Let us know what you think and would you really bet against Jen-Hsun?

00:00 Start
00:15 Leo starts a rant
01:20 Jim @ Adored
01:50 Leo rant continues
02:47 Nvidia mobile
03:12 Acquisitions
03:55 bet against Jen-Hsun?
05:55 Graphics cards and Omniverse
07:41 Nvidia Compute and the future ?
09:38 A feel for Nvidias ambition

LEOS NOTES FOR VIDEO:
Notes:
Gamers have a distorted view of Nvidia as that damned company that utterly fails to supply over-priced graphics cards.
We know that Nvidia was ditched by Apple back in the days of Intel Core Duo after the bump gate shocker.
Nvidia claimed design win after design win for Tegra yet I have never seen an Nvidia tablet, with the exception of their own Shield product.
Where are Nvidia powered smart phones?

Covering Nvidia drove Jim to quit doing tech analysis videos for AdoredTV
https://youtu.be/s23GvbQfyLA?t=1550

And yet, the graphics cards I have bought over the past few years for my own PC have been an EVGA GTX 980, reference GTX 1080 and a Founder’s Edition RTX 2080.
Just look at Nvidia’s stock price and market cap!

Nvidia acquired Mellanox for $7 billion in April 2020
https://nvidianews.nvidia.com/news/nvidia-completes-acquisition-of-mellanox-creating-major-force-driving-next-gen-data-centers

In September 2020 Nvidia announced they want to buy ARM for US$40 billion
https://nvidianews.nvidia.com/news/nvidia-to-acquire-arm-for-40-billion-creating-worlds-premier-computing-company-for-the-age-of-ai/

AI and 5G
Nvidia is a computing platform
With just a GeForce, every student can have a supercomputer
Omniverse and Digital Twins
Auto – Orin can replace many electronic control units with a single SoC with Atlan, due in 2024.
https://www.kitguru.net/components/graphic-cards/matthew-wilson/nvidia-atlan-soc-to-combine-grace-next-cpu-and-ampere-next-gpu/

Hyperion 8 is a platform for cars
Nvidia DRIVE includes Grace, Bluefield and Ampere Next

Three chip strategy for data centres – GPU, CPU and DPU
https://www.kitguru.net/components/cpu/joao-silva/nvidia-reveals-arm-based-grace-cpu-aimed-at-ai-and-hpc-applications/

GPUs generate a lot of cross-datacenter traffic, including Deep learning,
There is need for Nvidia's networking processors, the Data Processing Unit (DPU)
The Bluefield family of DPUs comes from Mellanox, and now a core part of Nvidia’s offerings
Bluefield is designed to offload a major part of network functions, including all the processing that goes with them, such as SSL and security analysis.
SuperPOD has been updated with Bluefield 2 and can include the A100 with 80GB
Enterprise AI A10 and A30, SuperPOD with BlueField 2 and EGX
SuperPOD starts at US$7 million and scales to US $60 million for a full system

Nvidia announces announcing Bluefield 3 DPU 400Gbps network processor with 22 billion transistors
Nvidia is already working on Bluefield 4 for 2024, which will be around 64B transistors, and incorporate Nvidia's AI acceleration technology

Nvidia DGX Station 320G can transfer 8TB of data in one second.
Consumes ‘just 1,500W’ which is the maximum safe limit for a US mains power socket

We shouldn’t compare Nvidia to AMD or Intel but instead should think in terms of Amazon, Apple and Tesla.