AMD CEO Lisa Su announced in her speech at CES that the world will eer a new era of computing called Yotta Scale in five years. He explained that this change occurs due to the jump in demand for computing power in artificial ielligence infrastructure, and the Utah scale will be proposed as a new performance measure.
According to Wccftech, the artificial ielligence infrastructure market is eering a new phase in which accelerators and rack-scale systems play a leading role. Lisa Su predicted that the computational capacity required in the Utah scale era will reach more than 10 Utaflops by the end of this decade. This figure is a measure of the processing power of artificial ielligence systems on a global scale. It should be meioned that Utah is one of the SI prefixes such as kilo, mega and giga, which is equal to 10 to the power of 24 or septillion.
Utah’s scale is revolutionizing the AI accelerator market
The CEO of AMD announced that the need for computing power is not only formed by cloud ceers. Artificial ielligence has eered all major areas of processing; From edge computing and personal computing to health, space and other industries. The simultaneous growth of these sectors has brought the level of demand for computing power to a poi that was unimaginable a few years ago.
According to Lisa Su’s estimate, in the era of Utah scale, the required computing capacity will be about 10 thousand times higher than in 2022. This estimate shows that the total available market for accelerators and rack-scale systems is very large and not limited to one company. This trend will iensify the competition among the main players of AI infrastructure.
AMD CEO’s speech at CES showed that the company plans to take advaage of the new wave of demand in the cloud and personal computing sectors. The iroduction of Instinct MI455X artificial ielligence accelerators, EPYC Venice processors and Helios AI solutions showed that the chip gia iends to play a ceral role in the next generation of artificial ielligence infrastructure.
AMD is also investing in Gorgon Poi hybrid processors to strengthen its position in the PC market. These processors iegrate graphics processing and computing power io one package to make it easier to run artificial ielligence workloads on PCs. This approach can blur the border between cloud and local computing.
