Seoul Economic Daily EnglishCodex · 实盘不做把握 4/10不做原因:风险收益比不佳
北京时间·🕒 发布 19:43:51(首尔 20:43)·⟳ 抓取 19:58:46 +14m55s·⚡ 决策 08-27 20:06:44 +7m58s·60.9 天前
韩国公布主权AI基础模型二阶段详细评分，SK Telecom团队排名第一
本次新闻主要是对已公布二阶段结果的详细评分透明化披露，新增事实不足以支撑对SK Telecom等上市载体建立可量化、可执行的短线正期望。
📄 原文Seoul Economic Daily English↗ 原文
Korea Discloses Full AI Model Scores; SK Telecom Takes Top Spot
South Korea's science ministry disclosed the full detailed scores from the second-stage evaluation of its "sovereign AI…
不做 SK Telecom（017670.KS）
不做SK Telecom（017670.KS）：评分第一是正面声誉信号，但晋级结果已公开、无金额增量且申诉未决，缺少可执行edge。
展开详情(成交明细 / 研究档案 / 原文全文 / 决策记录链)
研究档案
结论摘要
SK Telecom在韩国主权AI项目评分中居首属于声誉和政策资源边际利好，但三家晋级结果已于2026-08-18公开，本次增量只是分项分数和Motif申诉处理期限。新闻没有披露新增合同、补贴金额或商业化收入，且Motif申诉反而增加结果程序不确定性，因此不做。
消息传播链
- 韩国科学技术信息通信部已公布二阶段结果，Upstage、SK Telecom和LG AI Research晋级。
- 披露详细分项评分，SK Telecom以70.6分排名第一，并说明Motif申诉将在15日内处理。市场反应:韩国市场已收盘，未核到实时价格反应。
关键理由
- SK Telecom排名第一强化其韩国主权AI项目领先形象，但不是新的订单或收入确认。
- 8月18日已公开Upstage、SK Telecom和LG AI Research晋级，本次详细评分披露的预期差有限。
- 量级无法估算：现有公司收入基数较大，而本事件未披露新增资助、订单或商业化收费。
主要风险
- Motif申诉将在15日内处理，若复核改变项目排名或入选资格，可能削弱SK Telecom的政策背书。
- 未能通过本会话调用stock_tools取得实时行情、案例库和event_study，执行价与先例校准不足。
反向证据
- 做多前查反向 · 来源 web · 日期 2026-08-27 · 影响 weakens
Motif申诉构成对项目结果确定性的反向因素，削弱追多SK Telecom的边际胜率。
检索:Motif objection SK Telecom sovereign AI model scores reassessment 原文 ↗ - 做多前查反向 · 来源 web · 日期 2026-08-18 · 影响 weakens
核心晋级结果已于8月18日公开，本次评分细节不是全新主催化。
检索:MSIT Phase 2 evaluation results sovereign AI foundation model SK Telecom Upstage LG AI Research 原文 ↗
先例如何影响了这次决策
对本次决策的影响:先例不足进一步压低置信度，最终不下单。
方向:未取得可量化外部先例，不能用历史涨幅支持交易。
涨幅量级:没有同市场、同机制、同相位的高相似先例可校准涨幅。
形态:政府AI项目阶段性名单和评分披露通常需要金额或商业化路径才有强交易弹性，本次缺少该要素。
原文全文
展开原文 · 2,191 字
South Korea's science ministry disclosed the full detailed scores from the second-stage evaluation of its "sovereign AI foundation model" program on the 27th, after the results drew controversy. The move came at the request of Motif Technologies, one of the teams eliminated in the round.
According to the Ministry of Science and ICT, SK Telecom's (017670) elite team finished first with 70.6 points. Upstage followed with 69.9 points, LG AI Research with 69.0 points and Motif with 65.8 points.
The evaluation was scored out of 100 points, combining benchmark tests (25 points from AAII and 15 points from NIA), expert reviews (35 points) and user assessments (15 points from professional users and 10 points from members of the general public).
By category, Motif ranked first in the Artificial Analysis Intelligence Index (AAII) assessment, which applies global standards, with 11.9 points. Upstage came next at 9.4 points, followed by SK Telecom at 8.8 points and LG AI Research at 7.8 points. In the benchmark test run by the National Information Society Agency (NIA), which covered seven areas including Korean-language performance and reliability, SK Telecom scored highest at 13.4 points, ahead of Upstage at 13.3 points, LG AI Research at 12.8 points and Motif at 12.7 points.
In the expert review, conducted by 10 outside specialists from industry, academia and research institutes, LG AI Research earned the highest score at 29.5 points. SK Telecom followed with 29.3 points, Upstage with 29.1 points and Motif with 27.1 points.
Among the user assessments, which measured practical usability, SK Telecom ranked first in the professional user category with 11.6 points, based on reviews by 49 AI startup chief executives. LG AI Research placed second at 11.3 points, Upstage third at 10.8 points and Motif fourth at 8.4 points. In the public evaluation panel, which involved 185 members of the general public, LG AI Research scored highest at 7.6 points, followed by SK Telecom at 7.5 points, Upstage at 7.3 points and Motif at 5.7 points.
The ministry said it will handle Motif's objection to its elimination under established procedures and issue a final decision within 15 days.
According to the Ministry of Science and ICT, SK Telecom's (017670) elite team finished first with 70.6 points. Upstage followed with 69.9 points, LG AI Research with 69.0 points and Motif with 65.8 points.
The evaluation was scored out of 100 points, combining benchmark tests (25 points from AAII and 15 points from NIA), expert reviews (35 points) and user assessments (15 points from professional users and 10 points from members of the general public).
By category, Motif ranked first in the Artificial Analysis Intelligence Index (AAII) assessment, which applies global standards, with 11.9 points. Upstage came next at 9.4 points, followed by SK Telecom at 8.8 points and LG AI Research at 7.8 points. In the benchmark test run by the National Information Society Agency (NIA), which covered seven areas including Korean-language performance and reliability, SK Telecom scored highest at 13.4 points, ahead of Upstage at 13.3 points, LG AI Research at 12.8 points and Motif at 12.7 points.
In the expert review, conducted by 10 outside specialists from industry, academia and research institutes, LG AI Research earned the highest score at 29.5 points. SK Telecom followed with 29.3 points, Upstage with 29.1 points and Motif with 27.1 points.
Among the user assessments, which measured practical usability, SK Telecom ranked first in the professional user category with 11.6 points, based on reviews by 49 AI startup chief executives. LG AI Research placed second at 11.3 points, Upstage third at 10.8 points and Motif fourth at 8.4 points. In the public evaluation panel, which involved 185 members of the general public, LG AI Research scored highest at 7.6 points, followed by SK Telecom at 7.5 points, Upstage at 7.3 points and Motif at 5.7 points.
The ministry said it will handle Motif's objection to its elimination under established procedures and issue a final decision within 15 days.
决策记录链 · 只追加
每次领取产生一条记录,重新决策追加一条并以 supersedes 指向被替代的记录(ADR 0014)。
这条决策早于决策账本(2026-10-08 起逐条追加),没有追加记录。