(LEAD) Naver Cloud consortium to develop sovereign AI model for cybersecurity

This will let Google show Yonhap news articles that match or are related to your search

(ATTN: UPDATES with more details in paras 4, 6-9, photo; CHANGES headline)

SEOUL, Sept. 3 (Yonhap) -- A consortium led by Naver Cloud Corp. has been selected for a government project to develop an artificial intelligence (AI) foundation model specialized in cybersecurity, the Ministry of Science and ICT said Thursday.

The cloud and AI arm of portal operator Naver Corp. beat a rival consortium led by SK Telecom Co., earning high marks for its broad industry-academia-research partnership, technological capabilities and development strategy, among other factors, the ministry said.

The project aims to develop a sovereign cybersecurity AI model tailored to South Korea's security environment amid growing cyber risks stemming from rapidly evolving AI technologies.

The Naver Cloud consortium comprises 32 companies and institutions, including LG CNS Co. LG Uplus Corp., KEPCO KDN Co., Korea Hydro & Nuclear Power Co. and the Financial Security Institute.

Naver Corp.'s headquarters in Pangyo, south of Seoul, is seen in this undated file photo. (Yonhap)

The government plans to provide the consortium with up to 256 Nvidia B200 graphics processing units (GPUs) over the next 10 months to support the development of the model.

On top of the 256 GPUs provided by the government, the consortium plans to deploy 4,000 B200 GPUs from Naver Cloud and 256 H200 GPUs from LG to train two cybersecurity AI models based on Naver's HyperCLOVA X and LG's EXAONE foundation models.

Naver Cloud said the HyperCLOVA X-based model will be optimized for defensive cybersecurity tasks, while the EXAONE-based model will focus on offensive capabilities, allowing the two models to complement each other across cyberattack detection, analysis and response.

The consortium plans to train the models using about 830 terabytes (TB) of data from Naver Cloud, LG CNS and organizations operating critical national infrastructure and industrial systems.

The models will then be tested in seven strategically important sectors -- energy, finance, science and technology, telecommunications, semiconductors, defense and aerospace -- to strengthen AI-based cybersecurity capabilities across critical industries.

The ministry said it plans to sign a contract with the consortium later this month, paving the way for full-scale development.

U.S. chip giant Nvidia Corp. Chief Executive Officer (CEO) Jensen Huang (6th from L) Lee Hae-jin (7th from L), founder and board chair of internet giant Naver Corp. and Naver CEO Choi Soo-yeon (5th from L), pose for a photo at Nvidia's headquarters in Santa Clara, California, on July 25, 2026, in this file photo provided by Nvidia. (PHOTO NOT FOR SALE) (Yonhap)

ejkim@yna.co.kr (END)

7 military officers disciplined over lax management of firearms: data

Los Angeles declares 'BTS Week' ahead of concerts

Constitutional Court rules Seoul's lack of birth registration regulations for foreigners unconstitutional

2 killed in oil products tanker fire off Incheon

(URGENT) S. Korea, U.S., Japan to conduct Freedom Edge trilateral drills Sept. 7-11: S. Korean military

S. Korean response team continues search efforts for 9 missing in Nepal floods

(2nd LD) Lee nominates 6 ministers in major Cabinet reshuffle, including finance minister

White House says U.S. policy on N. Korea unchanged, reaffirms Trump's openness to dialogue with Kim

S. Korea's Protestant church offer condolences to Nepal floods victims

Lee reaffirms commitment to replacing Korean armistice with peace regime

Trump says S. Korea, Japan, others going to Alaska to 'load up' with oil, build pipelines

Lutnick: Trump gov't considering 'targeted,' thoughtful' tariff policy for semiconductors

(3rd LD) 1 dead, 6 missing after tugboat capsizes in waters off Busan

Navy's flagship sub returns after longest voyage across Pacific

Flood-hit hydroelectric power plant construction site in Nepal

Rice paddies in border town

National conference of teachers in N. Korea

(LEAD) Naver Cloud consortium to develop sovereign AI model for cybersecurity

这将使谷歌能够向您展示与您的搜索匹配或相关的韩联社新闻文章。

（注意：第 4、6-9 段更新了更多细节，并附有照片；标题已更改）

首尔，9月3日（韩联社）——韩国科学技术信息通信部周四表示，由Naver Cloud Corp.牵头的财团已被选中参与一项政府项目，该项目旨在开发专门用于网络安全的人工智能（AI）基础模型。

韩国产业通商资源部表示，门户网站运营商Naver公司的云计算和人工智能部门击败了由SK Telecom公司牵头的竞争对手财团，凭借其广泛的产学研合作、技术能力和发展战略等因素获得了高分。

该项目旨在开发一种自主网络安全人工智能模型，以适应韩国的安全环境，应对快速发展的人工智能技术带来的日益增长的网络风险。

Naver Cloud联盟由32家公司和机构组成，其中包括LG CNS公司、LG Uplus公司、KEPCO KDN公司、韩国水力核电公司和金融安全研究院。

这张未注明日期的资料照片显示的是位于首尔以南板桥的Naver公司总部。（韩联社）

政府计划在未来 10 个月内向该联盟提供多达 256 个英伟达 B200 图形处理单元 (GPU)，以支持该模型的开发。

除了政府提供的 256 个 GPU 外，该联盟还计划部署来自 Naver Cloud 的 4,000 个 B200 GPU 和来自 LG 的 256 个 H200 GPU，以训练基于 Naver 的 HyperCLOVA X 和 LG 的 EXAONE 基础模型的两个网络安全 AI 模型。

Naver Cloud表示，基于HyperCLOVA X的模型将针对防御性网络安全任务进行优化，而基于EXAONE的模型将侧重于进攻能力，使这两个模型能够在网络攻击检测、分析和响应方面相互补充。

该联盟计划使用来自 Naver Cloud、LG CNS 和运营关键国家基础设施和工业系统的组织的约 830 TB 数据来训练模型。

然后，这些模型将在七个具有战略意义的重要领域——能源、金融、科学技术、电信、半导体、国防和航空航天——进行测试，以加强关键行业的基于人工智能的网络安全能力。

该部表示计划在本月晚些时候与该财团签署合同，为全面开发铺平道路。

美国芯片巨头英伟达公司首席执行官黄仁勋（左六）、互联网巨头Naver公司创始人兼董事长李海镇（左七）和Naver首席执行官崔秀妍（左五）于2026年7月25日在位于加利福尼亚州圣克拉拉的英伟达总部合影。这张资料照片由英伟达提供。（照片不得出售）（韩联社）

ejkim@yna.co.kr (完)

7名军官因枪支管理不善受到纪律处分：数据

洛杉矶宣布举办“BTS周”演唱会

宪法法院裁定首尔缺乏外国人出生登记规定违宪

仁川附近海域油轮起火，2人死亡

（紧急）韩国、美国、日本将于9月7日至11日举行“自由边缘”三边联合军演：韩国军方

韩国救援队继续搜寻尼泊尔洪灾中失踪的9人

（第二条）李显龙提名6位部长，进行内阁重大改组，其中包括财政部长

白宫表示美国对朝政策不变，重申特朗普对与金正恩对话持开放态度。

韩国基督教新教会向尼泊尔洪灾受害者表示慰问

李重申致力于以和平机制取代朝鲜停战协定。

特朗普称韩国、日本和其他国家将前往阿拉斯加“大量开采”石油，并修建输油管道。

卢特尼克：特朗普政府正在考虑对半导体行业实施“有针对性”、“深思熟虑”的关税政策

（3rd LD）釜山附近海域拖船倾覆，1人死亡，6人失踪

海军旗舰潜艇完成横跨太平洋的最长航程后返回。

尼泊尔受洪水侵袭的水电站建设工地

边境小镇的稻田

朝鲜全国教师大会

（牵头）Naver Cloud联盟将开发用于网络安全的自主人工智能模型