Naver and Intel are starting to build an artificial intelligence (AI) semiconductor ecosystem.With t.. – MK – 매일경제
Language
Change font
A
A
A
A
Share
TOP
Most read
Language
Change font
A
A
A
A
Share
Naver and Intel are starting to build an artificial intelligence (AI) semiconductor ecosystem.
With this alliance, Naver Cloud is planning to lower the cost of ultra-large AI learning and inference through direct design, while Intel is planning to penetrate the AI learning and inference chip market dominated by Nvidia.
According to the information and communication technology (ICT) industry on the 27th, Intel and Naver Cloud will jointly announce this in the U.S. from the 8th to the 9th of next month. The core of this collaboration is the joint construction of a platform ecosystem for “Gowdy,” an AI learning and inference chip developed by Havana Labs, an Intel affiliate. The advantage of Gaudi is that it can be designed differently according to its purpose. It can be used for both AI learning and reasoning.
Currently, parallel processing software is virtually monopolized by Nvidia Cuda (CUDA). As Nvidia dominated 80% of the GPU market, it naturally dominated the AI accelerator development platform. The problem is securing GPUs. Nvidia’s H100, a $40,000 GPU, is known to take more than 50 weeks from order to arrival.
For Naver, it is difficult to update HyperClova X, a super-giant AI. Intel explained earlier, “Cost efficiency, resource utilization, and power consumption are important considerations for corporate competitiveness,” adding, “GPU-based servers consume too much power and idle CPU-based servers.”
Naver is conducting an experiment that can achieve the same performance without GPU. Previously, a test was conducted to improve cloud performance by utilizing Intel’s CPU-based data center chip, ‘2nd Generation 2nd On Scalable Processor’. He also participated in the construction of a Gaudi joint platform this time.
In particular, if ultra-large AI compression solutions are combined with AI chips designed by themselves, it is expected both inside and outside Naver that AI learning and inference efficiency can be increased by more than 10 times. Currently, the experiment is led by Lee Dong-soo, director of Naver Cloud Hyperscale AI.
The domestic tech industry is pursuing global collaboration one after another in the age of AI. NCsoft is a representative place. At Google’s headquarters in Mountain View, California, NCsoft CEO Kim Taek-jin met with Mark Lomair, vice president of Google’s cloud AI and machine learning infrastructure, and sought cooperation in AI, cloud, and productivity. NCsoft will work with Google Cloud to apply AI technology to the entire game development process and jointly develop a platform to build a game ecosystem.
In addition, NCsoft decided to introduce a productivity improvement tool called ‘Geminai for Google workspace’. Starting with this discussion, we will expand global cooperation between the two companies and further solidify cooperation in areas that can create synergy, NCsoft CEO Kim Taek-jin said. “We will work with Google Cloud to build an ecosystem for new gaming experiences for global users.”
2024-07-03 20:14:41
2024-07-02 13:58:30
2024-07-03 20:39:14
2024-07-02 10:41:49
2024-07-02 09:43:35
2024-07-03 09:48:15
2024-07-03 21:23:57
2024-07-02 09:39:02
2024-07-02 22:14:04
2024-07-02 14:20:00
※ This service is provided by machine translation tool, NAVER papago.
Maeil Business Newpaper(MK) provides these translations “as they are” and makes no warranties of any kind, either explicitly or implicitly, regarding accuracy, reliability and marketability, suitability for a particular purpose, etc. of translation. Please be informed that the content provided may not be translated accurately due to limitations in machine translation before using this service.
Copyright (c) 매경닷컴. Maeil Business News Korea & mk.co.kr, All rights reserved.
Prohibition of unauthorized reproduction, redistribution, and use of AI learning
