🐦 推特JukanQwen · 实盘不做把握 8/10不做原因:无增量信息
北京时间·🕒 发布 09-09 13:40:37时区推断·⟳ 抓取 09-09 13:41:05·⚡ 决策 09-14 15:47:42 +5d02h·43.0 天前
SK海力士在2026未来论坛上表示3D DRAM研发可借用逻辑代工工艺，属长期技术路线表态而非商业落地
正文是SK海力士自办论坛上的技术方向演讲：只讲3D DRAM「可以」用代工工艺、需要解决哪些设计与封装难题，没有量产时间表、资本开支、订单、客户或任何金额，属研发路线图而非需求验证或商业落地。跨源查重72小时内无同类已处理记录，但事件本身为9月8-9日，距今已过3个交易日，短线重定价窗口关闭。二阶映射（薄膜沉积、键合、CMP设备）只有长期主题关联，内部KB反而显示混合键合导入节点在推迟，本次表态没有给这些标的带来新的可交易事实。因此不产生LONG或SHORT腿。
📄 原文Jukan@jukan05↗ 原文
SK hynix: "Foundry processes could be used for 3D DRAM" SK hynix said on the 9th that it could use foundry processes in…
不做 SK海力士（000660.KS）
自办论坛上的3D DRAM技术路线表态：只讲可借用代工工艺、尚需解决散热/键合/工艺整合，无量产时间、无资本开支、无订单与客户、无任何金额，属研发方向而非需求验证或商业落地；对超大市值主体缺乏短线增量。事件发生于9月8-9日，距now已3个交易日，重定价窗口关闭。同样不存在做空催化，故不开空。
不做 拓荆科技（688147.SH）
3D DRAM长期利好薄膜沉积与键合设备，但本次表态未带来订单、验证或定点，内部KB反而显示混合键合商业化在推迟（美光至2028、海力士2026Q4才评估无助焊剂键合）；纯主题联想+消息已过期，不构成可执行催化。
不做 北方华创（002371.SZ）
同上，仅有长期设备面联想，本次无客户点名、无采购信号、无时间表，且事件已过3个交易日，无可执行边界。
不做 华海清科（688120.SH）
CMP/减薄与3D堆叠长期相关，但本次论坛无量化信息与验证节点，无法建立事件到收入的因果与时点，不予交易。
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
事实层：SK海力士副总裁在2026未来论坛提出3D DRAM的三大技术方向（代工工艺做外围电路、总线带宽最大化、超短距离传输），并明确指出还需攻克散热、细间距互连、键合与工艺整合，CEO以「弯路」形容行业变化。影响链：3D DRAM是把存储与逻辑边界打通的下一代架构，长期利好先进封装/键合/薄膜沉积设备与材料，但从论坛表态到设备采购、验证、量产收入之间隔着多年研发，且公司自身未给时点。证据层：内部KB（2026-06行业周报）显示HBM4混合键合商业化在推迟——美光推迟至2028年、海力士2026Q4才启动无助焊剂键合评估，说明3D DRAM相关设备需求兑现更远；本次全文无金额、无订单、无指引。价格层：未查行情，因为无论SK海力士还是A股设备股的价格都不会改变「无新增商业事实、且已过期3个交易日」这一结论。结论：AVOID，不做多也不做空（技术路线表态不构成公司级利空，无risk_signals）。
🔎 新鲜度:最早出现 2026-09-08 · 扩散 medium · price-in 风险 high ↗
消息传播链
- 2026-09-08首发SK海力士2026未来论坛（利川SUPEX中心现场演讲）非一手副总裁Kim Kyunghoon、Son Hoyoung现场提出3D DRAM三大技术方向，CEO郭鲁正致开场辞；本轮未取得论坛官方页面链接，故标记为未核验一手源
- 2026-09-09财经媒体扩散韩国媒体英文报道（正文所述「said on the 9th」）非一手论坛次日被媒体整理成稿，属对已发生演讲的转载，未新增金额或时间表
- 分析师账号转发韩媒报道，是本次信号的直接来源；社媒转发不构成新增事实，且时点已是事件后约一天
主要风险
- 若后续出现3D DRAM量产年份或专项资本开支的一手披露，本判断需重估，届时先进封装/键合设备链可能获得真实催化
- 把长期技术路线表态当成短期题材去追高A股封装设备股，容易在没有订单验证的情况下承担情绪退潮回撤
反向证据
- 做多前查反向 · 来源 kb_search · 日期 2026-06-08 · 影响 weakens
内部行业周报显示存储大厂HBM4量产节奏同步、混合键合技术导入节点有望推迟（美光推迟至2028年商业化，海力士2026Q4才启动无助焊剂键合全面评估），说明3D堆叠相关设备需求兑现比论坛叙事更远
检索:SK海力士 3D DRAM 进展 混合键合 先进封装 时间表 受益供应链 - 做多前查反向 · 来源 kb_search · 日期 2026-07-09 · 影响 none
设备国产化专家纪要给出的是2026年新签订单目标与国产化率等既有产业数据，与本次3D DRAM论坛表态无因果关联，未构成A股映射标的的新增催化
检索:SK海力士 3D DRAM foundry 相关A股映射订单/定点 - 做空前查反向 · 来源 web · 影响 none
正文无任何下修、开支失控或产品延期数字，CEO「弯路」措辞属行业环境描述，不构成公司级净利空，不满足做空红线的强空催化条件
检索:SK海力士 3D DRAM 表态是否隐含利空（研发开支上升/路线不确定）
先例如何影响了这次决策
对本次决策的影响:先例缺失未改变结论——本条已因「无新增商业事实+过期」独立成立AVOID，不依赖先例校准幅度
方向:无可用先例；本轮英文与韩文检索均只返回SK集团门户、百科与行情聚合页噪声，未取到任何「存储大厂论坛技术路线表态→股价反应」的可溯源首发案例
涨幅量级:无高相似、锚干净的先例，因此未据此设定任何tp/sl
形态:无法评估见光死/高开低走，样本为0
工具执行结果 · 成功 4 · 降级 0 · 失败 0 · 取消 0
- search_processed_events 成功 stock_tools HTTP · 20 ms
{"result": "search_processed_events:agent query=`3d dram 逻辑代工工艺 存储 | kwak nohjung 未来论坛 3d dram | sk hynix 3d dram foundry process | sk海力士 3d dram 未来论坛`；结构化索引覆盖完整 72 小时；旧记录兼容扫描近 0.0/72 小时、0 条本臂决策(完整)，返回 0 条候选。未命中只表示这些 agent query 表达式没有召回候选，不能据此证明消息首发或不是重复；按当前事件继续正常研判。"} - kb_search 成功 stock_tools HTTP · 5,488 ms
{ - web_search 成功 Bing RSS · 482 ms
Bing RSS 搜索结果（原始查询:SK hynix Future Forum 2026 3D DRAM foundry process Icheon September 9） - web_search 成功 Bing RSS · 271 ms
Bing RSS 搜索结果（原始查询:SK하이닉스 3D D램 양산 일정 파운드리 공정 2026 미래포럼）
原文全文
展开原文 · 2,298 字
SK hynix: "Foundry processes could be used for 3D DRAM"
SK hynix said on the 9th that it could use foundry processes in developing 3D DRAM, a next generation memory chip.
At the 2026 SK hynix Future Forum held the previous day at the SUPEX Center on the Icheon campus, SK hynix vice presidents Kim Kyunghoon and Son Hoyoung presented the use of logic foundry processes for DRAM peripheral circuits, maximizing data bus bandwidth, and ultra short distance transmission as the key technology directions for 3D DRAM.
3D DRAM is a next generation DRAM that increases density by stacking memory cells vertically instead of laying them out on a single plane (2D).
Kim and Son explained that realizing this technology requires solving not only design and heat issues but also packaging challenges such as fine interconnects, bonding, and process integration. They also expected that as the boundaries between front end and back end (packaging) processes, and between foundry and memory technologies, draw closer together, joint optimization that goes beyond existing domains will become important.
"3D technology is changing even the technical boundary between the fab and packaging," Son said. "Along with innovation in advanced packaging technology, it is important to build optimization that crosses existing domains and new systems of collaboration."
Shin Changhwan, a professor at Korea University who spoke on the same topic, said, "3D DRAM goes beyond simply stacking chips vertically. It is a technology that breaks down the boundary between memory and logic. With device scaling approaching its limits and the time and power consumed by moving data between the system and memory increasing, the shift to 3D technology is inevitable." He proposed a 3D integrated design framework that links four areas, materials, devices, packaging, and architecture, through artificial intelligence (AI).
In his opening remarks, SK hynix CEO Kwak Nohjung said, "If the memory market of the past was a straight road where the competition was about who could run faster, today it is like driving on a winding road that changes moment to moment. What matters is not only internal capability but also the speed at which we read the pace and direction of change in the external environment and adapt flexibly."
SK hynix said on the 9th that it could use foundry processes in developing 3D DRAM, a next generation memory chip.
At the 2026 SK hynix Future Forum held the previous day at the SUPEX Center on the Icheon campus, SK hynix vice presidents Kim Kyunghoon and Son Hoyoung presented the use of logic foundry processes for DRAM peripheral circuits, maximizing data bus bandwidth, and ultra short distance transmission as the key technology directions for 3D DRAM.
3D DRAM is a next generation DRAM that increases density by stacking memory cells vertically instead of laying them out on a single plane (2D).
Kim and Son explained that realizing this technology requires solving not only design and heat issues but also packaging challenges such as fine interconnects, bonding, and process integration. They also expected that as the boundaries between front end and back end (packaging) processes, and between foundry and memory technologies, draw closer together, joint optimization that goes beyond existing domains will become important.
"3D technology is changing even the technical boundary between the fab and packaging," Son said. "Along with innovation in advanced packaging technology, it is important to build optimization that crosses existing domains and new systems of collaboration."
Shin Changhwan, a professor at Korea University who spoke on the same topic, said, "3D DRAM goes beyond simply stacking chips vertically. It is a technology that breaks down the boundary between memory and logic. With device scaling approaching its limits and the time and power consumed by moving data between the system and memory increasing, the shift to 3D technology is inevitable." He proposed a 3D integrated design framework that links four areas, materials, devices, packaging, and architecture, through artificial intelligence (AI).
In his opening remarks, SK hynix CEO Kwak Nohjung said, "If the memory market of the past was a straight road where the competition was about who could run faster, today it is like driving on a winding road that changes moment to moment. What matters is not only internal capability but also the speed at which we read the pace and direction of change in the external environment and adapt flexibly."
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。