Hi! 👋 Welcome to The Big Y!
Nvidia’s stock is almost a vertical line going up these days as the GenAI rush has fueled massive GPU consumption. Not to be left out of the gold rush, AMD is making a concerted effort to get into the GPUs for AI game and steal some of Nvidia’s market share, facing an uphill battle.
Nvidia has a massive first-mover advantage. AMD might be able to gain market share as Nvidia has challenges keeping up with production demand, but the cost for startups to switch to AMD might not be worth the savings. AMD chips are at the moment cheaper than Nvidia chips, and some might argue that they could be better if correctly optimized. But Nvidia definitely owns the software piece with CUDA while AMD relies on OpenCL, which just isn’t as performant as CUDA.
A generalized assumption would be that any startup that is building cutting edge GenAI tools or applications and spending millions on compute, would probably stick to Nvidia GPUs and CUDA to be safe. Trying to save a few dollars on GPUs switching to AMD is risky since you’re going against the tried and true formula.
Is it potentially too late for AMD to build the reputation and products needed to be truly competitive in the GenAI rush? Maybe, but if they look forward, nobody is owning the edge inference market at this point and that is where our fancy new GenAI will be running in the future.
Alphabet soup. Google has been the butt of many jokes recently but it can be a little hard to not chuckle when they continue to release, rename, remove and merge products. A funny tweet (not on purpose) from Google Cloud’s CEO last week which accidently showed some of Google’s fumbling to align and keep a clean strategy when it comes to their AI products.
Know someone who might enjoy this newsletter? Share it with them and help spread the word!
Thanks for reading! Have a great week! 😁
🎙 The Big Y Podcast: Listen on Spotify, Apple Podcasts, Stitcher, Substack