Nvidia will spend $40 billion to buy Arm Holdings from Softbank 🔥

a quick look at what makes this deal so special —
Arm Holdings makes processors that most commonly go into your smartphones, tablets and other electronics devices

These CPUs are based on the ARM architecture — a type of reduced instruction set, low power, low cost processor line, best suited for simple devices
But why would Nvidia, a high-end graphic card company, pay a bomb of a price for basic processors?
Well, in the first go the deal boosts Nvidia’s position in smartphones, edge devices, IOT and other embedded electronics gadgets.

Nvidia can sell graphic cards to a whole new segment now, as demand for high-graphic processing capabilities (cameras and such) keeps increasing.
But the real deal is about bringing the ARM processors inside gigantic datacenters — a fledgling business for Nvidia

As you know, the cloud industry has exploded over the last two decades

Massive data centers power our daily apps including the one you’re reading this tweet on
NVDA sells GPUs to these datacenters, because GPUs help speed up AI and other data-hungry processing

Like graphics processing, data processing takes a lot of parallel computations

The GPU, which has multiple small cores, is well suited to handle these parallel computations
In fact GPUs can be so much better at this task, that powerful GPUs combined with a small, not-very-expensive CPU is actually preferred in modern data centers over a traditionally dense CPU-only setups.

It's cost effective, fast, and futuristic.
If NVDA can combine its awesome GPUs with simple Arm CPUs and sell the end-to-end setup to datacenter buyers, it can keep a larger share of the datacenter market for itself

Not to forget, Arm processors are insanely cheaper than the current Intel processors used in datacenters
The global datacenter chip market is worth $70 billion — dominated by Intel today

But as data-intensive workloads grow, NVDA can easily displace Intel

Moreover, NVDA could drastically deflate the cost curve of deploying and managing datacenters, while improving performance!
Nvidia has been looking to dominate the modern computing stack — one that favors data-hungry processing, for a long time now.

That's why it quietly purchased high-speed data center networking company Mellanox last year for $7 billion.
With that, Nvidia now own 2 of the 3 key elements of the modern computing stack —

1. Compute (GPUs + ARM Processors)
2. Networking (Mellanox)
3. Storage (___________________)

The stock markets have been anticipating these moves for a while now.

NVDA trades at a $320 billion market cap, Intel on the other hand, lags at a $210 billion market cap
Nvidia is going head on here with Intel.

If Nvidia wins this war, over the next decade or so the company could rapidly scale a $50 billion+ annual business selling hardware just to data centers.
Links to relevant reads 🤙

Nvidia's press release —

nvidianews.nvidia.com/news/nvidia-to…

Nvidia's future — nextplatform.com/2020/09/14/now…

GPUs and datacenters —nextplatform.com/2020/02/15/the…

nextplatform.com/2019/06/17/nvi…

Other —
svi.nl/HuygensGPU

and thanks @tejas_rd for helping break this down
Hop on over to filtercoffee.co for some less serious commentary! ☕

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Filter Coffee ☕

Filter Coffee ☕ Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!