South Korea’s web portal giant Naver Corp. has replaced the main chip supplier of its artificial intelligence server for its map service, Naver Place, from Nvidia Corp. to Intel Corp.
Naver has so far used Nvidia’s graphic processing unit (GPU)-based server to run its AI-powered location information provision service but recently replaced it with Intel’s central processing unit (CPU)-based server, people familiar with the matter said on Monday.
Naver’s AI server switch comes as global information technology firms are increasingly disgruntled with Nvidia’s GPU price hikes and a global shortage of its GPUs.
Nvidia is the world's top AI GPU maker The Korean tech company utilizes an AI model to distinguish false information from useful data when users search for information using such terms as “nearby restaurants” on the Naver Map app for the Naver Place service.
This process involves data processing, inference and verification, and Naver has used Nvidia’s GPU chips to power such processes.
SHORT SUPPLY OF AI GPUs
However, it has become increasingly hard to secure GPU accelerators, including Nvidia’s H100, which use multiple chips for data processing and computing, due to growing demand for their use in generative AI devices.
Intel's fourth-generation Sapphire Rapids Xeon scalable processors The price of an AI accelerator using Nvidia’s GPUs has risen to as high as 80 million won ($59,193) from about 40 million won earlier this year, making it costlier for AI server users to upgrade their systems, industry officials said.
To make things worse, the lead time, or the time to get Nvidia’s GPU-based accelerators after placing an order, has grown to 52 weeks, making it even harder for a system upgrade, they said.
Nvidia produces GPUs used in generative AI devices such as ChatGPT, a chatbot developed by Microsoft Corp.-backed OpenAI, which requires semiconductors that can process large data quickly.
The US chip designer has been heavily relying on Taiwan Semiconductor Manufacturing Co. (TSMC) for the chip packaging of its GPUs and accelerators. With TSMC’s packaging process lines almost fully booked, however, Nvidia is turning to other foundry players such as Samsung Electronics Co.
Intel's Sapphire Rapids Xeon scalable processors INTEL IMPROVES CPU CAPABILITIES
Industry officials said GPU servers usually handle AI inference processes 10 times faster than CPU servers.
However, Intel considerably improved its AI inference process time for those clients seeking alternatives to Nvidia’s GPUs, industry watchers said.
Naver has gone through a month of test runs of its Intel CPU-powered AI server before putting it into service, sources said.
The Korean company will likely continue to use Intel’s next-generation CPUs, such as the Sapphire Rapids CPU, for its system upgrades to offer better location and other information search services, they said.
Analysts said the Naver-Intel tie-up may lead to a further diminishing of Nvidia’s influence in the AI processor market.
Microsoft Corp. has reportedly teamed up with fabless chip designer Advanced Micro Devices Inc. (AMD) to help bolster the US chipmaker’s expansion into AI processors.
The two companies are joining forces to compete against Nvidia, which controls an estimated 80% market share in the global AI processor market, media reports said.
