Hybrid bonding adoption on the logic side is accelerating, and Bernstein raises Besi's target price to €310

Overview

This report updates Bernstein's view on the European and Japanese semiconductor advanced packaging supply chain, with a core focus on the adoption pace of hybrid bonding in logic chips and HBM. The report believes the D2W hybrid bonding adoption trend on the logic side is clear, and that TSMC SoIC capacity expansion, Intel's resumption of expansion after 2027, and application demand from AMD, Apple, Google, Nvidia, and others will drive rapid growth in the hybrid bonding equipment market.

Core views

The core view is that adoption on the logic side is more certain than in HBM, which is the main basis for raising TAM and Besi's target price; HBM adoption still faces uncertainty in timing and layer count, potentially starting with Samsung's HBM4E in 2027 or being delayed until HBM5 in 2028; Besi remains a leader on the logic side, but the report conservatively assumes its 2028 logic share at 80% and HBM share at 70%; DISCO indirectly benefits from the migration to hybrid bonding through high-cleanliness dicing and grinding equipment.

Analysis framework

The report uses a combination of technology roadmap assessment, capacity and tool intensity calculations, TAM breakdown, company market share assumptions, and a P/E valuation framework. On the logic side, estimates are derived from tool installations and capacity expansion at TSMC, Intel, and other manufacturers; on the HBM side, estimates are derived from total HBM capacity, hybrid bonding penetration, and the number of tools required per kwpm; company valuation uses future Q5-Q8 EPS and a 40x forward P/E multiple.

Methodology notes

• Market sizingTAM forecast

  Break down the hybrid bonding equipment market by application

  The report divides hybrid bonding equipment demand into logic and HBM, and forecasts 2028 TAM at $887mn and $489mn respectively, for a total of about $1.4bn.

• Capacity-driven modelCapacity-to-tool intensity model

  Derive equipment demand from capacity expansion

  On the logic side, the assumption is that each additional 1kwpm of SoIC capacity requires about 5-6 hybrid bonding tools; on the HBM side, due to lower alignment precision requirements, each 1kwpm requires about 2-3 tools.

• Valuation methodForward P/E valuation

  P/E valuation based on forward EPS

  Besi's target price is based on 40x future Q5-Q8 EPS; the multiple was reduced from the previous 45x to reflect lower visibility for hybrid bonding, but the target price was raised due to upward earnings revisions.

Asset mapping & comparison

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

• BE Semiconductor Industries NV (BESI.NA)

  Core beneficiary of hybrid bonding equipment

  Strengths

  Technology leader and currently close to the sole supplier on the logic side; the report assumes 2028 logic share of 80% and HBM share of 70%, with hybrid bonding revenue reaching about €901mn.

  Weaknesses

  Visibility on the HBM adoption pace is weaker, and lower HBM requirements may attract participation from Korean domestic equipment suppliers.

  Comparison

  The report's forecast for 2028 hybrid bonding equipment shipments is 24% above Besi's base-case scenario and 8% below its high-case scenario.

  Risks

  Lower-than-expected or delayed hybrid bonding adoption, delayed recovery in the mainstream packaging market, loss of hybrid bonding market share, and insufficient acceptance of TCB products.

• Disco Corp (6146.JP)

  Indirect beneficiary of advanced packaging and hybrid bonding migration

  Strengths

  High-cleanliness dicing and grinding equipment is critical for particle control and yield, benefiting from HBM, CoWoS, NAND CBA, BSPDN, D2W hybrid bonding, and SiC investment.

  Weaknesses

  Valuation has come down from its peak but remains not low, and earnings are sensitive to the AI-related capacity cycle.

  Comparison

  DISCO's 1-year forward P/E is about 36.5x, down 32% from the recent peak of 53.9x.

  Risks

  Excess AI-related GPU and HBM capacity, changes in the competitive environment, and adverse FX movements.

• Hybrid bonding equipment TAM

  Thematic supply chain opportunity

  Strengths

  The adoption path on the logic side is clear, improving interconnect density and reducing power consumption, supporting advanced chiplet architectures.

  Weaknesses

  On the HBM side, TCB may still dominate in the short term, and hybrid bonding penetration may cover only a limited number of layers.

  Comparison

  Adoption certainty on the logic side is stronger than on the HBM side; Bernstein is more optimistic on logic and more conservative on HBM.

  Risks

  If HBM4E remains at 12-hi or JEDEC relaxes height standards, adoption of hybrid bonding in HBM could be delayed.

Key data

• 2028 hybrid bonding equipment TAM约$1.4bnComposed of $887mn from logic and $489mn from HBM.
• 2028 logic hybrid bonding shipments237 unitsThe previous forecast was 103 units, reflecting stronger TSMC SoIC adoption.
• 2028 HBM hybrid bonding shipments130 unitsBased on the assumption that part of HBM4E adopts hybrid bonding and HBM5 migrates to hybrid bonding.
• 2028 total hybrid bonding equipment shipments367 units24% above Besi's base-case scenario and 8% below its high-case scenario.
• TSMC SoIC capacity forecast10kwpm in 2025 to 80-85kwpm in 2028The report text mentions both 80kwpm and 85kwpm, both pointing to roughly 8x growth.
• Besi 2028 hybrid bonding revenue约€901mn / 约$1.1bnBased on assumed market shares of 80% in logic and 70% in HBM, and assumed sales volume of about 281 units.
• Besi 2028 total revenue forecast约€1.8bn11% above market consensus.
• Besi target price€310.00Previous €280.00, maintaining an Outperform rating.
• DISCO target price¥85,000Maintaining an Outperform rating.

Impact & implications

If the report's assumptions play out, hybrid bonding will become an important incremental source of demand for advanced packaging equipment, and Besi is expected to gain significant revenue and profit leverage through its technology leadership and high market share; DISCO, meanwhile, will share in the industry's migration benefits through high-cleanliness dicing, grinding, and advanced packaging-related equipment. For investors, adoption progress on the logic side is more worth tracking as the main short- to medium-term validation thread than HBM.

Risks

• The adoption timing of hybrid bonding in HBM is uncertain and may be delayed from 2027 to 2028.
• The TCB solution remains effective in HBM, which may delay hybrid bonding penetration.
• Besi may face competition in the HBM market from Korean suppliers such as SEMES, Hanwha, and Hanmi.
• Vendors such as Shibaura, ASMPT, and TEL may take share from Besi on the logic side.
• If the recovery in the mainstream packaging market is delayed until late 2025 or 2026, it will weigh on Besi's short-term earnings.
• If AI-related GPU and HBM capacity becomes excessive, it may affect DISCO's earnings and valuation multiple.
• FX volatility may pose downside risk to DISCO's revenue.

What to watch

• TSMC SoIC capacity expansion progress at AP7 and AP8, as well as commercialization of related applications at AMD, Apple, Google, Groq, and Nvidia.
• The capacity ramp pace of Intel Clearwater Forest, Diamond Rapids, and subsequent CPU platforms for Foveros Direct 3D.
• Whether Samsung HBM4E adopts 16-hi and whether it begins introducing hybrid bonding in 2027.
• Whether HBM5 mass-production timing, stack layer count, and I/O channel upgrades drive more layers to migrate to hybrid bonding.
• Changes in Besi's share of logic and HBM hybrid bonding equipment orders.
• Changes in DISCO's orders and valuation multiple for high-cleanliness dicing and grinding equipment.