The development of the AI market and technological progress are obvious to all, and many new technologies are being rapidly realized in the rapid development of AI hardware! The large-scale production of AI models has led to an almost unlimited demand for GPU hardware, and many companies may want to invest in GPU development! However, they are unable to form a competitive relationship with Nvidia! The reason is that even if they develop and produce GPUs, they cannot achieve the cost-effectiveness of Nvidia's products of the same level! They may even be quickly eliminated! The investment will be lost! Large-scale procurement of GPUs is similar to computer upgrades! It requires continuous adaptation to more advanced AI models in the future through larger-scale and higher-end new GPUs! This is a huge process! After a few years, the currently expensive high-end GPUs will be quickly eliminated! This process and speed will be much faster than Moore's Law for CPUs!
The problem is, AMD's GPU research and development is far behind nvidia! Even with short-term large-scale adoption, the pace of being rapidly eliminated is fast! Purchasing should not be based on boosting performance!
Another unicorn is brewing! Currently, there is a company called Arm that has independent R&D and patents for GPUs. The problem is that this company is a partial patent development company! "For example, if this company has the ability to develop and produce a complete set of cars, but the company only focuses on the development of core car engines and energy technologies, and sells patents to car manufacturers for profit!" SoftBank's acquisition of Arm has not been involved in Arm's development in the GPU field! It is estimated that this year SoftBank will reduce its holdings of Arm and use the funds from Arm's shareholding reduction to establish a...