📰 DigiTimes 实时DigiTimes RealtimeDoubao · 实盘不做把握 8/10不做原因:无增量信息
北京时间·🕒 发布 09-14(台北日)仅日期·时区推断·⟳ 抓取 12:23:06·⚡ 决策 09-15 12:27:33 +4m27s·42.3 天前
DigiTimes 行业评论：CPO 接近量产，下一阶段比拼接口、良率与可维护性，无具体订单或公司级新事实
这是一篇署名行业评论/高管访谈，核心内容是 CPO 走向量产过程中的通用工程挑战（对准、可维护接口、良率、ELS/InP、测试、材料）与 FIT 的能力自述，没有订单、客户定点、量产里程碑、金额或指引等任何可量化、公司级的新增事实；'nears mass production''depends on' 属预期与条件式表述，CPO 又是 2025–2026 年已被充分传播的主题，据此做多任何标的都缺乏预期差与可证伪的交易锚点。
📄 原文DigiTimes Realtime↗ 原文
中文網 繁體 简体 SUBSCRIBE My account Sign out SUBSCRIBE Trending Apple Huawei Nvidia AI Search Scaling AI Glasses Inference E…
不做 鸿腾六零八八精密科技（FIT Hon Teng）（06088.HK）
公司仅在行业评论中陈述 CPO 接口/精密制造的通用能力与开发方向，无订单、客户定点、量产时点或指引等新增事实，且为已被充分传播的 CPO 主题，缺乏预期差与可证伪催化，不做；待具体合同/定点/指引落地再评估。
不做 CPO/光互连主题股（光模块、FAU/精密光学、InP 激光器、CPO 测试等）（N/A）
本文未点名任何供应商获得新业务，也无金额/份额/时点，无法据此从主题映射出有真实敞口的可交易标的；仅凭'CPO 接近量产'叙事追主题属于无 edge 的情绪交易，回避。
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
DigiTimes 文章只是借 FIT 高管之口阐述 CPO 量产的行业瓶颈与供应链要求，既无 FIT 自身的订单/财务增量，也未点名任何供应商拿到新业务；属于已被反复定价的 CPO 主题叙事，不具备事件驱动交易所需的新增事实与边界，回避。
🔎 新鲜度:最早出现 2026-09-14 · 扩散 high · price-in 风险 high ↗
关键理由
- 文章不含任何具体合同、客户、量产时点、份额或收入数字，仅有趋势性、条件式判断，无法形成可量化的基本面变化与正期望
- CPO/NPO 是已长期广泛传播的市场主线，单纯'接近量产'评论不构成预期差，且文中多处强调良率与维护等不确定性而非需求确认
主要风险
- 若后续衍生出 FIT 或某供应商获得具体 CPO 接口/光引擎定点、客户大单或明确量产排产的硬事实，则需重新研判，本文本身不提供该 edge
- CPO 主题股可能因媒体热度短线波动，但这是情绪/资金面驱动而非本事件新增基本面，追高承担见光死与回撤风险
反向证据
- 做多前查反向 · 来源 web · 影响 invalidates
本文及现有材料未提供任何 FIT 在 CPO 接口上的中标、定点或收入数据；做多前提（可验证需求与 materiality）缺失
检索:FIT Hon Teng 6088 CPO 订单 客户定点 量产 2026 / Foxconn Interconnect co-packaged optics order customer 原文 ↗ - 做多前查反向 · 来源 web · 日期 2026-09-14 · 影响 weakens
文章本身反复强调对准良率、可维护性、InP 产能、测试规范与维护标准等规模化障碍，更接近'主题仍存执行不确定性'而非需求确认，削弱短线追多逻辑
检索:CPO 量产 良率 维护 外部激光 InP 瓶颈 推迟 原文 ↗
先例如何影响了这次决策
对本次决策的影响:因无新增可定价事实且无可定位标的，未构建先例面板即判定回避；后续若出现具体定点/订单，再按同市场同相位先例另行校准
方向:无明确公司级方向：纯行业趋势评论缺少可锚定的首发商业事件，历史上此类'XX 接近量产/产业趋势'稿对个股的影响高度依赖后续是否兑现为订单，本身不提供稳定方向或胜率
涨幅量级:不存在与'无增量事实的行业评论'对齐的可比首发事件来校准涨幅；CPO 主题股的历史波动主要由具体客户大单/财报/路线图发布驱动，不能挪用为本事件幅度
形态:经验上已被充分传播的热门科技主题在高管评论/行业稿见报后，相关标的更易呈现高开低走或无反应（缺乏新增基本面），不支持追主题
工具执行结果 · 成功 1 · 降级 0 · 失败 0 · 取消 0
- search_processed_events 成功 stock_tools HTTP · 17 ms
{"result": "search_processed_events:agent query=`co-packaged optics external laser inp passive alignment wafer-level optics supply chain | cpo near mass production interfaces yield maintenance foxconn interconnect technology fit `；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 agen
原文全文
展开原文 · 8,387 字
中文網
繁體
简体
SUBSCRIBE
My account
Sign out
SUBSCRIBE
Trending Apple
Huawei
Nvidia
AI Search
Scaling AI Glasses
Inference Era AI
Advanced Packaging
Trending
SUBSCRIBE
CONNECT WITH US
Sign out
REALTIME NEWS
Corning sees CPO and glass core as new semiconductor battlegrounds
Tomorrow's Headlines
Sep 14, 18:47
China optical transceivers avoid US FCC curbs as NPO and CPO compete in high-speed links
Tomorrow's Headlines
Sep 14, 18:46
China's AIDC token costs hinge on access to advanced GPUs
Tomorrow's Headlines
Sep 14, 18:46
Home
Tech
Semiconductors
CPO nears mass production — next battle about interfaces, yield and maintenance
Ninelu Tu
, Taipei
Sep 14, 2026, 09:50
0
Credit: DIGITIMES
As AI computing bandwidth keeps rising, optical interconnect architectures are moving beyond pluggable optical modules toward near-packaged optics (NPO) and co-packaged optics (CPO). But large-scale deployment will depend on more than optical-engine performance. Precision alignment, serviceability, manufacturing yield, and whether the supply chains for light sources, packaging, testing, and materials can advance in step are becoming equally critical.
Kai-pin Lin, senior director of US business development at Foxconn Interconnect Technology (FIT), said CPO is bringing optical components closer to the chip package, making previously overlooked interface and manufacturing issues central to industry discussions.
Precision fiber alignment is among the most fundamental capabilities and one of the hardest to build quickly, Lin said. Optical connectors must maintain sub-micron alignment over long periods. Producing a small number of laboratory samples is relatively straightforward, but sustaining stable yields across shipments of hundreds of thousands or even millions of units requires engineering capabilities spanning precision tooling, micro-molding, materials control, automated assembly and measurement.
Detachable fiber-to-chip interfaces are another area drawing attention. CPO shortens the distance between optical engines and switch ASICs, but permanently attaching fibers directly to the package can sharply increase replacement costs when assembly yields are poor, fibers become contaminated, or maintenance is required.
Lin said detachable, blind-mate and even field-replaceable fiber interfaces could become important conditions for scaling CPO deployment. FIT is developing such interfaces with upstream optical-material and wafer-level optics suppliers, drawing on its position in connectors and precision mechanical components.
AI racks raise the integration bar
The engineering challenge for next-generation AI racks extends well beyond optical communications.
Lin said suppliers are increasingly encountering co-design issues created by the convergence of high-speed signals, high-voltage power, and liquid cooling. From signal integrity at 224G, safety and arc management under high-voltage DC architectures, to sealing and long-term reliability in liquid-cooling loops, all things traditionally handled by different specialist suppliers.
In AI racks, however, these elements are often concentrated on the same tray, backplane, or even around the same connector interface. Optimizing one part of the system can therefore reduce the design margin available to another.
A second hurdle is the shortening new product introduction (NPI) cycle for AI platforms. Faster specification changes mean prototyping, design revisions and engineering decisions often need to proceed in parallel. Suppliers need not only rapid prototyping capability but also early coordination among signal, power, thermal and manufacturing teams.
Lin said component suppliers that continue to rely on sequential, department-by-department responses to customer requirements will struggle to match the pace of AI platform changes.
A third challenge is the need to combine precision engineering with mass-production capability. Producing a prototype in the laboratory is very different from maintaining yield, cost and supply stability at volume.
FIT therefore aims to use its in-house capabilities in tooling, electroplating, stamping and automated manufacturing, together with Foxconn Group resources in full-rack systems, to bridge the gap between design validation and mass production. The company is also seeking to move beyond a conventional component-supplier role toward joint rack development.
CPO scale-up depends on the wider optical supply chain
CPO and NPO are only part of the broader shift in optical communications. Lin said large-scale adoption of optical interconnects in AI data centers will also require progress in external laser sources (ELS), advanced packaging, wafer-level optics, testing and measurement, field maintenance and materials supply.
For light sources, the reliability and serviceability of external lasers, together with supply capacity for indium phosphide (InP) epitaxy and substrates, will directly affect the pace of system deployment. On the manufacturing side, the industry needs to gradually shift from active alignment to passive alignment and wafer-level processes, requiring foundries, OSAT providers, and optical-component suppliers to establish more complete design and process rules.
Testing remains another weakness. Optical measurement infrastructure is not yet as mature as electrical testing, and future progress by the Optical Internetworking Forum (OIF) and various multi-source agreements (MSAs) on interface and test specifications will directly affect mass-production efficiency.
Maintenance after deployment also remains critical. Fiber end faces are highly sensitive to contamination, and the absence of standardized procedures for on-site cleaning, insertion, removal, replacement and repair could undermine system reliability even when individual components meet performance targets.
The depth of supply for other upstream precision components, including glass substrates, low-loss optical materials, optical fibers and ferrules , will also determine how quickly costs can fall.
Lin said the next phase of optical-communications competition will not be determined by CPO technology alone, but by whether the broader ecosystem can support data-center operations at scale. As transmission speeds move from 800G to 1.6T and beyond, the industry will need what he described as stronger "optical supply-chain discipline" across mass production and operations.
Article
translated by Levi Li and edited by Jack Wu
AI Search
Related stories
Commentary: Huawei builds NPO alliance to shape CPO-era optical interconnect standards
CPO mass production accelerates new four-stage test equipment market
Nvidia CPO roadmap positions TSMC COUPE for next AI infrastructure wave
Huawei targets AI optical networking with InP chip startup investment
Largan, Sunny Optical target FAU in push toward CPO and AI optics
Show more
Categories
Chips + components IC manufacturing IT + CE Server, IPC, cloud computing, IoT
Tags
bandwidth CPO design manufacturing materials optics production
Share this article
Other links
Print
Related topic
Global supply chain: Mobile, telecom, computing
BIZ FOCUS
Sep 14, 08:00
Hauman Unveils Micro LED Testing Breakthrough for AI Optical Interconnects
Monday 14 September 2026
Scaling Photonics: Assembly, Test and Fiber Automation at SEMICON Taiwan
Friday 11 September 2026
DAS Expands in Taiwan and Unveils World First at SEMICON
Friday 11 September 2026
Toward Technologies Redefines Relay Solutions at SEMICON Taiwan 2026
MOST-READ
7 DAYS NEWS
Google's v10 TPU opens door to AMD without sidelining Broadcom, MediaTek
Update: ABF substrate maker Unimicron searched in origin-labeling probe
Qualcomm's HBC play pulls Samsung and SK hynix into a new race beyond HBM
SK keyfoundry expands 8-inch capacity as China demand strains supply
Samsung, SK Hynix step up China NAND upgrades as US tool risks grow
The 1,000-watt problem: a TSMC veteran on the brute force forging Nvidia's AI chips
Wuliangye skips carmaking, turns liquor waste into silicon-carbon anodes
eCloudvalley expands enterprise AI Agent portfolio
CXMT rides AI boom to 80% operating margin
Nvidia, MediaTek deepen 20-year chip ties
Full list
Members only
×
Sorry, the page you are trying to open is available only for our paid subscribers.
Please login to read more
New users, please register first
Notification
×
繁體
简体
SUBSCRIBE
My account
Sign out
SUBSCRIBE
Trending Apple
Huawei
Nvidia
AI Search
Scaling AI Glasses
Inference Era AI
Advanced Packaging
Trending
SUBSCRIBE
CONNECT WITH US
Sign out
REALTIME NEWS
Corning sees CPO and glass core as new semiconductor battlegrounds
Tomorrow's Headlines
Sep 14, 18:47
China optical transceivers avoid US FCC curbs as NPO and CPO compete in high-speed links
Tomorrow's Headlines
Sep 14, 18:46
China's AIDC token costs hinge on access to advanced GPUs
Tomorrow's Headlines
Sep 14, 18:46
Home
Tech
Semiconductors
CPO nears mass production — next battle about interfaces, yield and maintenance
Ninelu Tu
, Taipei
Sep 14, 2026, 09:50
0
Credit: DIGITIMES
As AI computing bandwidth keeps rising, optical interconnect architectures are moving beyond pluggable optical modules toward near-packaged optics (NPO) and co-packaged optics (CPO). But large-scale deployment will depend on more than optical-engine performance. Precision alignment, serviceability, manufacturing yield, and whether the supply chains for light sources, packaging, testing, and materials can advance in step are becoming equally critical.
Kai-pin Lin, senior director of US business development at Foxconn Interconnect Technology (FIT), said CPO is bringing optical components closer to the chip package, making previously overlooked interface and manufacturing issues central to industry discussions.
Precision fiber alignment is among the most fundamental capabilities and one of the hardest to build quickly, Lin said. Optical connectors must maintain sub-micron alignment over long periods. Producing a small number of laboratory samples is relatively straightforward, but sustaining stable yields across shipments of hundreds of thousands or even millions of units requires engineering capabilities spanning precision tooling, micro-molding, materials control, automated assembly and measurement.
Detachable fiber-to-chip interfaces are another area drawing attention. CPO shortens the distance between optical engines and switch ASICs, but permanently attaching fibers directly to the package can sharply increase replacement costs when assembly yields are poor, fibers become contaminated, or maintenance is required.
Lin said detachable, blind-mate and even field-replaceable fiber interfaces could become important conditions for scaling CPO deployment. FIT is developing such interfaces with upstream optical-material and wafer-level optics suppliers, drawing on its position in connectors and precision mechanical components.
AI racks raise the integration bar
The engineering challenge for next-generation AI racks extends well beyond optical communications.
Lin said suppliers are increasingly encountering co-design issues created by the convergence of high-speed signals, high-voltage power, and liquid cooling. From signal integrity at 224G, safety and arc management under high-voltage DC architectures, to sealing and long-term reliability in liquid-cooling loops, all things traditionally handled by different specialist suppliers.
In AI racks, however, these elements are often concentrated on the same tray, backplane, or even around the same connector interface. Optimizing one part of the system can therefore reduce the design margin available to another.
A second hurdle is the shortening new product introduction (NPI) cycle for AI platforms. Faster specification changes mean prototyping, design revisions and engineering decisions often need to proceed in parallel. Suppliers need not only rapid prototyping capability but also early coordination among signal, power, thermal and manufacturing teams.
Lin said component suppliers that continue to rely on sequential, department-by-department responses to customer requirements will struggle to match the pace of AI platform changes.
A third challenge is the need to combine precision engineering with mass-production capability. Producing a prototype in the laboratory is very different from maintaining yield, cost and supply stability at volume.
FIT therefore aims to use its in-house capabilities in tooling, electroplating, stamping and automated manufacturing, together with Foxconn Group resources in full-rack systems, to bridge the gap between design validation and mass production. The company is also seeking to move beyond a conventional component-supplier role toward joint rack development.
CPO scale-up depends on the wider optical supply chain
CPO and NPO are only part of the broader shift in optical communications. Lin said large-scale adoption of optical interconnects in AI data centers will also require progress in external laser sources (ELS), advanced packaging, wafer-level optics, testing and measurement, field maintenance and materials supply.
For light sources, the reliability and serviceability of external lasers, together with supply capacity for indium phosphide (InP) epitaxy and substrates, will directly affect the pace of system deployment. On the manufacturing side, the industry needs to gradually shift from active alignment to passive alignment and wafer-level processes, requiring foundries, OSAT providers, and optical-component suppliers to establish more complete design and process rules.
Testing remains another weakness. Optical measurement infrastructure is not yet as mature as electrical testing, and future progress by the Optical Internetworking Forum (OIF) and various multi-source agreements (MSAs) on interface and test specifications will directly affect mass-production efficiency.
Maintenance after deployment also remains critical. Fiber end faces are highly sensitive to contamination, and the absence of standardized procedures for on-site cleaning, insertion, removal, replacement and repair could undermine system reliability even when individual components meet performance targets.
The depth of supply for other upstream precision components, including glass substrates, low-loss optical materials, optical fibers and ferrules , will also determine how quickly costs can fall.
Lin said the next phase of optical-communications competition will not be determined by CPO technology alone, but by whether the broader ecosystem can support data-center operations at scale. As transmission speeds move from 800G to 1.6T and beyond, the industry will need what he described as stronger "optical supply-chain discipline" across mass production and operations.
Article
translated by Levi Li and edited by Jack Wu
AI Search
Related stories
Commentary: Huawei builds NPO alliance to shape CPO-era optical interconnect standards
CPO mass production accelerates new four-stage test equipment market
Nvidia CPO roadmap positions TSMC COUPE for next AI infrastructure wave
Huawei targets AI optical networking with InP chip startup investment
Largan, Sunny Optical target FAU in push toward CPO and AI optics
Show more
Categories
Chips + components IC manufacturing IT + CE Server, IPC, cloud computing, IoT
Tags
bandwidth CPO design manufacturing materials optics production
Share this article
Other links
Related topic
Global supply chain: Mobile, telecom, computing
BIZ FOCUS
Sep 14, 08:00
Hauman Unveils Micro LED Testing Breakthrough for AI Optical Interconnects
Monday 14 September 2026
Scaling Photonics: Assembly, Test and Fiber Automation at SEMICON Taiwan
Friday 11 September 2026
DAS Expands in Taiwan and Unveils World First at SEMICON
Friday 11 September 2026
Toward Technologies Redefines Relay Solutions at SEMICON Taiwan 2026
MOST-READ
7 DAYS NEWS
Google's v10 TPU opens door to AMD without sidelining Broadcom, MediaTek
Update: ABF substrate maker Unimicron searched in origin-labeling probe
Qualcomm's HBC play pulls Samsung and SK hynix into a new race beyond HBM
SK keyfoundry expands 8-inch capacity as China demand strains supply
Samsung, SK Hynix step up China NAND upgrades as US tool risks grow
The 1,000-watt problem: a TSMC veteran on the brute force forging Nvidia's AI chips
Wuliangye skips carmaking, turns liquor waste into silicon-carbon anodes
eCloudvalley expands enterprise AI Agent portfolio
CXMT rides AI boom to 80% operating margin
Nvidia, MediaTek deepen 20-year chip ties
Full list
Members only
×
Sorry, the page you are trying to open is available only for our paid subscribers.
Please login to read more
New users, please register first
Notification
×
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。