As much as I can’t picture Jim Cramer naming a puppy AMD, maybe Micro if he’s small, I think he and the rest of the market don’t truly understand these two players. A computer primarily consists of four parts; CPU, GPU, RAM, and Hard Drive. The rest is fluff to accommodate human interaction, a little wire, and a fan. There is an old saying that he who dies with the most toys wins. In my opinion, advantage AMD.
The CPU or Central Processing Unit is a computer’s primary engine. For years the top two CPU manufacturers were AMD and Intel. Today AMD wins. They have decidedly the best, fastest technology when compared to Intel, dollar for dollar. Not sure there is a huge top potential power difference, but the cost differential along with AMD’s reputation compared to Intel today puts them on top.
Also for many years, the top two video card manufacturers were AMD and Nvidia. Some years the AMD Radeon series was best, and others the Nvidia GeForce series won. Today, Nvidia has the better-performing product, per dollar so prior to AI if you wanted the best video gaming experience, did high-end graphic design, digital video editing, or A/V streaming, you coughed up as high as a few thousand dollars for just the graphics card in your computer. Today that would be an Nvidia GeForce card.
AI was not hampered by our level of technology. It was the algorithm that teaches a computer to think, virtually. Nothing we can do now that the genie is out of the bottle. Fast, cheap, global connectivity, fast, cheap computers, and inexpensive mass storage medium are the breeding grounds for the tech. As the AI ship has sailed, not to be judgmental, let’s get to the money.
The advancement of this technology, beyond the knowledge and scope it will reach inherent in the tech itself, will be dependent on speed and throughput. Eventually systems will be developed specifically for AI and the best machines will have the fastest overall throughput based on how the 4 basic components communicate with each other. Now you can affiliate yourself with other companies and optimize how you connect. So your CPU can read and write data faster and stack it more efficiently on your hard drive. And when your video is taxed your GPU uses RAM for additional memory to avoid lag. Having your RAM optimized to receive and return the specific type of data the AI engine requires increases efficiency.
The easiest way to advance the “state of the art” is to not have to consult with anyone. That is why today, in my opinion, though its stock price and market sentiment don’t agree, AMD is ahead! I guarantee you Lisa Su has engineers in one or many labs working out how to get THEIR CPUs to communicate better with THEIR GPUs to accommodate AI protocols. Six months from now Nvidia won’t know what hit them. I wouldn’t be surprised if Lisa Su didn’t quietly pick up Kensington and Seagate or Western Digital. Then they would own the whole enchilada! Now Jensen Huang could just as easily tender an offer for Intel and pick up a RAM and Hard Drive manufacturer of his own. Then again the anti-trust hammer has been coming down hard and fast lately. They could block Nvidia from acquiring Intel, but AMD already owns one of the top two global CPU and GPU manufacturing firms.
The bottom line, even if Nvidia were to assemble its own 4 part computational harmony, it would still take a while to familiarize itself with the other 3 component manufacturers. AMDs CPUs and GPUs already have an intimate understanding of each other. In car terms, Nvidia makes a better turbocharger than AMD today. But AMD builds turbochargers and the engines they feed fuel to. How long do you think it will take them to increase the flow of fuel into the motor to accommodate the AI track it’s driving on if they have total control of both the engine and turbocharger design?