FC-BGA/ABF substrate technology seminar: Larger advanced packaging raises barriers for high-end substrates

Overview

This report summarizes the FC-BGA (ABF) substrate technology seminar hosted by J.P. Morgan on July 14, 2026, featuring Mr. Toshihiko Nishio of SBR Technology. The discussion focused on core substrate material trends including organic core substrates, glass core substrates, HDI core substrates, and ceramic core substrates, as well as advanced packaging routes such as EMIB-T, CoWoS, CoPoS, and CoWoP from the substrate perspective. The report's core focus is how the expansion in AI chip and HBM package size affects industry competition through substrate low CTE, high rigidity, flatness, fine processing, and mass production capability.

Core views

The report believes that as CoWoS/CoPoS package sizes increase, solder reliability, warpage, and flatness issues during the mounting process become more prominent, which may require high-rigidity substrates and glass core substrates. Larger reticles will raise the technological barriers for high-end substrates, and Japanese companies such as Ibiden with high-end ABF substrate capabilities may gain stronger relative competitive advantages. EMIB-T has theoretical advantages in reliability and back-end yield, but still needs to establish mass production capability for external customers, as well as substrate-side and assembly-process capabilities. CoWoP seeks to reduce dependence on traditional ABF substrates and Japanese substrate makers, but currently still faces major challenges in SLP line width, PCB processing accuracy, yield, flip-chip mounting, and chip-side design rules.

Analysis framework

The report uses expert seminar notes and technology route comparisons to assess the impact of changes in advanced packaging technology on electronic component companies and substrate material/equipment suppliers across dimensions such as package structure, differences in material thermal expansion coefficients, reticle size, substrate size, solder reliability, mass production maturity, and supply chain capability.

Methodology notes

• Expert interviews / seminarTechnical expert seminar summary

  Draw investment implications for the supply chain through SBR Technology experts' interpretation of advanced packaging and FC-BGA substrate technology trends.

  This method relies on expert judgment regarding technology routes, mass production barriers, and supply chain capabilities. It is suitable for identifying medium-term technology bottlenecks and potential beneficiaries, but it is not equivalent to company financial forecasts or formal rating changes.

• Supply chain analysisComparison of advanced packaging technology routes

  Compare the structural differences and mass production challenges of CoWoS-S, CoWoS-L, CoPoS, CoWoP, EMIB-T, and different core substrate materials.

  The analysis focuses on interposer usage, RDL routing, silicon bridges, TSVs, low-CTE materials, high rigidity, flatness, microvia processing, and assembly yield.

• Supply constraint analysisAssessment of key materials and mass production capability

  Identify supply chain bottlenecks through high-end glass cloth, drilling tools, glass core substrate production lines, and the mass production capabilities of substrate manufacturers.

  Differences in the capabilities of companies such as Nittobo, Union Tool, Absolics, Ibiden, Shinko Electric Industries, and Unimicron are used to judge future constraints on advanced packaging capacity expansion.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Ibiden

  High-end FC-BGA/ABF substrate supplier that may benefit from the larger size of CoWoS/CoPoS and demand for high-rigidity substrates.

  Strengths

  May have a relative competitive advantage in demand for large-size, high-rigidity, high-flatness substrates; TSMC's cooperation with Ibiden and Innolux to develop CoPoS glass core substrates has drawn attention.

  Weaknesses

  New core materials such as glass core substrates and HDI core substrates are still in development or pre-mass-production stages, and there is uncertainty around mass production timing and customer qualification.

  Comparison

  Compared with most substrate makers still in the development stage, Ibiden is viewed in the report as a potential key supplier in the high-end ABF field.

  Risks

  If CoWoP successfully bypasses traditional ABF substrates, or if competitors catch up and glass core mass production is delayed, its medium-term advantage may weaken.

• Nittobo

  Key material supplier of T-glass for high-end organic core substrates.

  Strengths

  T-glass helps control substrate CTE and improve strength and electrical performance; experts say it is almost the sole supplier in high-end products.

  Weaknesses

  Capacity expansion pace must balance demand from customers such as Nvidia and Broadcom against competitor catch-up and future oversupply risk.

  Comparison

  Other glass cloth manufacturers such as Taiwan Glass are currently technically unable to keep up with high-end demand.

  Risks

  Competitor technology catch-up, end-demand volatility, or future oversupply.

• Union Tool

  Supplier of microvia drilling tools for high-rigidity, low-CTE substrates.

  Strengths

  Experts say it leads in supplying drilling tools capable of stably processing microvias in hard materials.

  Weaknesses

  Demand depends on the expansion pace of high-end substrate materials and advanced packaging.

  Comparison

  Other companies are currently unable to keep up with the relevant processing capabilities.

  Risks

  If packaging routes shift to reduce demand for traditional substrate processing, or if customer capacity expansion is delayed, demand elasticity may decline.

• Absolics

  Early participant in mass production of glass core substrates.

  Strengths

  Experts say it is currently one of the few companies with mass production facilities and that has entered the small-scale production stage.

  Weaknesses

  Overall infrastructure for glass core substrates remains insufficient, and customer qualification and large-scale mass production still require time.

  Comparison

  Compared with manufacturers such as Ibiden, Shinko Electric Industries, and Unimicron that are still in development stages, Absolics is further ahead.

  Risks

  Large-scale mass production before 2030 is difficult, customer qualification cycles are long, and technical yields are uncertain.

• Intel / EMIB-T

  Proponent of EMIB-T technology and a potential alternative route to CoWoS-L.

  Strengths

  It does not use a large interposer, theoretically reducing overall warpage risk and improving back-end process yield; it can optimize power delivery to HBM and LSI through TSVs.

  Weaknesses

  External customer mass production capability, substrate-side capability, and assembly processes have not yet been established.

  Comparison

  Compared with CoWoS-L, EMIB-T has potential advantages in connection reliability; but compared with traditional EMIB, its mass production maturity is insufficient.

  Risks

  IFS external customer capacity may be unreliable, Amkor assembly authorization and outsourcing implementation are uncertain, and the adoption pace among potential customers such as Broadcom and Google is unclear.

• TSMC / CoWoS / CoPoS

  Core participant in advanced packaging capacity and technology routes.

  Strengths

  CoWoS capacity can be expanded, and CoPoS improves productivity through square panels; TSMC's difficulty in controlling CoPoS is relatively manageable.

  Weaknesses

  CoWoS-L faces flatness, CTE mismatch, and solder reliability challenges at 5.5x and 9.5x reticle sizes.

  Comparison

  If CoPoS is delayed, part of the impact can be offset by increasing existing CoWoS capacity; compared with EMIB-T, CoWoS-L is more susceptible to one-time bonding of large modules and warpage.

  Risks

  If 9.5x reticle cannot be achieved, it may constrain Nvidia's Feynman-generation packaging targets.

• Nvidia / CoWoP

  End-demand party that may adopt CoWoP in the future and reduce dependence on Japanese ABF substrate makers.

  Strengths

  CoWoP can reduce intermediate substrates, shorten signal paths, improve thermal design flexibility, and lower costs.

  Weaknesses

  It requires SLP line width to fall to 10um or below, and must solve PCB processing accuracy, yield, flip-chip mounting, chip-side routing rules, and bump pitch adjustments.

  Comparison

  Compared with traditional CoWoS, CoWoP has a more simplified structure; but it conflicts with the standardization of existing CoWoS chips.

  Risks

  Because Nvidia is prioritizing support for the Feynman generation, it may lack the resources to simultaneously develop CoWoP as an independent technology route; feasibility remains in question.

Key data

• CoWoS-S reticle limitapproximately 3.3xExperts believe that as silicon interposer size increases, warpage rises and solder reliability declines, so a 3.3x reticle size is the limit.
• TSMC CoWoS-L roadmapexpanding from 3.3x to 5.5x, with plans to reach 9.5x by 2029The report points out that evaluation of the 5.5x Rubin Ultra has not gone smoothly, and achieving 9.5x is highly difficult.
• Substrate size changes3.3x is 85mm×85mm, 5.5x is 110mm×110mm, and 9.5x is 130mm×140mmLarger package sizes intensify challenges in substrate flatness, warpage, and soldering.
• HBM4 bump pitch trendmay shrink from 65um to 36umFiner pitch requires higher-density routing capability between HBM and LSI.
• EMIB-T potential supported sizeIntel claims it can support up to 12x reticle sizeHowever, EMIB-T has not yet been established as a mass production technology, and external customer capacity and substrate-side capability remain key issues.
• CoPoS panel size310mm square panelCoPoS transfers the CoWoS process, which is usually completed on 300mm wafers, onto square panels to improve area utilization and productivity.
• CoWoP SLP line width requirementmass production requires 10um or below, versus roughly 15-20um currentlyPCB processing accuracy, yield, flip-chip mounting, and adjustments to chip-side routing rules remain the main obstacles.
• Assessment of glass core substrate mass production timingexperts believe large-scale mass production before 2030 will be difficultAbsolics already has mass production facilities and has entered the small-scale production stage, while most other manufacturers remain in development or planning stages.
• J.P. Morgan global equity research rating distributionOverweight 53%, Neutral 36%, Underweight 12%This is a compliance disclosure table as of July 4, 2026, and not a new rating from this report on the relevant companies.

Impact & implications

From an industry perspective, the evolution of AI chip packaging toward larger reticles, larger substrates, and higher HBM interconnect density will make low CTE, high rigidity, high flatness, and fine processing capability key competitive thresholds. High-end ABF substrate makers such as Ibiden may benefit over the medium term from the challenges of large-package packaging; Nittobo and Union Tool hold important supply-chain positions in high-end glass cloth and microvia drilling tools. On the other hand, if EMIB-T and CoWoP achieve successful mass production, they may alter the existing value distribution of ABF substrates, but the report believes these routes still face significant technical and mass production uncertainties at present.

Risks

• CoWoS-L faces flatness, solder reliability, and CTE mismatch issues at 5.5x and 9.5x reticle sizes, which may drag on the high-end AI packaging roadmap.
• EMIB-T has not yet established reliable mass production capability for external customers, and substrate-side and assembly processes still need validation.
• Although glass core substrates have material advantages, mass production facilities, customer qualification, and supply chain infrastructure are insufficient, making large-scale mass production before 2030 difficult.
• If CoWoP succeeds, it may reduce the use of traditional ABF substrates and weaken dependence on Japanese substrate makers, but the route itself also faces high technical uncertainty.
• If key material suppliers such as Nittobo expand capacity too quickly, they may face competitor catch-up and future oversupply risks.
• The report is a summary of an expert seminar, and some judgments rely on expert views; it does not provide company-level financial models, ratings, or target price updates.

What to watch

• TSMC CoWoS-L 5.5x Rubin Ultra evaluation progress and validation of 9.5x reticle solder reliability after 1H27.
• Whether Nvidia's 2029 Feynman-generation packaging roadmap is constrained by CoWoS-L or alternative packaging technologies.
• Whether Intel can establish EMIB-T as a mass production technology serving external customers, and progress on Amkor assembly authorization and outsourcing.
• Progress in cooperation between TSMC, Ibiden, and Innolux on CoPoS glass core substrate development.
• Mass production plans and customer qualification timelines for glass core substrates at Absolics, Ibiden, Shinko Electric Industries, Unimicron, Samsung Electro-Mechanics, and Sumitomo Chemical.
• The pace of Nittobo T-glass capacity expansion, demand pull from Nvidia and Broadcom, and catch-up by competitors such as Taiwan Glass.
• Whether Union Tool can maintain its leading position in microvia processing tools for high-rigidity, low-CTE materials.
• Whether Nvidia's CoWoP route can achieve breakthroughs in SLP line width, PCB yield, flip-chip mounting, and chip-side rule adjustments.