AMD says it still believes in Moore’s Law

Nvidia released the 4nm Ada Lovelace architecture GPU in September of this year and launched the GeForce RTX 40 series graphics card. After the release of the new graphics card, many players do not focus on the performance improvement, but on the price, thinking that the price is too high.

Later, Mr. Jensen Huang, the founder, and CEO of Nvidia responded in a video call with the media, saying that the foundry price of 12-inch wafers is now much more expensive than before. The cost has risen sharply, and it is believed that Moore’s Law is dead. It is impossible to provide double performance at the same cost or halve the cost of the same performance. The idea that the cost of chips falls over time is outdated.

As a competitor, AMD does not fully agree with Nvidia’s view, at least for now. AMD Senior Vice President and Chief Technology Officer Mark Papermaster recently said in an interview with the media, AMD believes that Moore’s Law will continue for six to eight years, and it will not die in a short time.

Mark Papermaster said that in the past, transistor density could double every 18 to 24 months while maintaining the same cost, but chip development and manufacturing costs are getting higher and higher. AMD sees this exactly, using innovative technologies such as small chip design and 3D packaging, and combining special accelerators such as CPU, GPU, and FPGA to overcome the problem of rising costs in design, production, and manufacturing.

Although the ways of choosing are different, both AMD and Intel have chosen the road of small chip design, which represents the development direction for a period of time in the future, whether there is Moore’s Law or not, the pace of innovation on both sides continues.