Skip to main content
No. of Recommendations: 0
Hi Tinker

My thoughts are, the GPU is currently the best general purpose chip for a variety of specific problems. This was kind of a lucky break for Nvidia, it just happened that GPUs were the best available option for a number of matrix-based operations.

However, theres plenty of reason to believe that more specific chips (ASICs, FPGAs) can be orders of magnitude better than GPUs for those specific problems.

Its a 'general purpose versus domain-specific' battle. General-purpose by definition will tend to lose those battles.

So the GOOG/AMZN/FB etc of the world don't need to care about NVDA's roadmap. They have specific problems to solve (eg: AI training and inference) and if customers want to solve those problems, they can invest in provisioning custom chips.

The challenge for NVDA is if those datacenter type problems (AI et al) consolidate around a few specific problems that can be solved by specific ASICs (eg: TPUs) which can be commoditised.



cheers
Greg
Print the post  

Announcements

What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.