The Structural Issue: JEDEC Thickness Standard Discussions

The core concern for Besi's HBM-related revenue is that JEDEC is actively discussing raising the maximum package height for next-generation HBM, which would allow memory makers to continue using existing thermal compression bonding (TCB) instead of moving to hybrid copper bonding (HCB).

• In early March 2026, reports emerged that industry participants were discussing a target of 825–900 μm for 20-high HBM stacks (HBM4E, HBM5), compared to the current HBM4 cap of 775 μm .
• By late March 2026, JEDEC had begun "formal discussions" to relax these standards .
• By April 1, 2026, TrendForce reported that JEDEC was planning to relax height specs to approximately 900 μm for next-generation HBM, starting with HBM4E .
• The official JEDEC HBM4 standard (JESD270-4), finalized in February 2026, set the package thickness cap at 775 μm — already a relaxation from the previous ~720 μm, which made 16-layer HBM4 feasible without hybrid bonding .

The key implication: if JEDEC raises the limit to ~900 μm, memory makers could use existing TC bonders for at least one to two more generations before hybrid bonding becomes mandatory, pushing out Besi's revenue opportunity by years .

The Competitive Alternative: iHBM and HCB

Both major memory makers are developing packaging innovations that reduce the thermal urgency of hybrid bonding, though they approach the problem differently.

• SK Hynix unveiled iHBM in May 2026 — a thermal solution that embeds cooling elements directly inside the HBM package, built on its proven MR-MUF technology. The company states it reduces thermal resistance by 30% and is planned for next-generation products including HBM5 . MR-MUF itself offers better heat conduction and fewer voids than conventional methods .
• Samsung is pushing Hybrid Copper Bonding (HCB), which it claims improves thermal resistance by >20% over TCB and supports 16+ layers . Samsung also presented comparative data at ECTC 2026 .
• Critically, both Samsung and SK Hynix postponed hybrid bonder adoption for HBM4 mass production as of February 2026, choosing to use TC bonders instead . Samsung plans partial use of hybrid bonders for later HBM generations, while SK Hynix has established a pilot line and is evaluating Besi equipment .

Besi's Q1 2026 Performance: Fundamentally Strong

Besi's Q1 2026 results, reported on April 23, 2026, were objectively strong across the board:

• Bookings of €269.7M beat consensus estimates of ~€250–260M by approximately 4% .
• Besi guided for a 30–40% revenue increase in Q2 2026 .
• Revenue growth was driven by high-end mobile, 2.5D AI/photonics, and data center demand .

Despite these strong numbers, the stock faced downward pressure shortly after — a ~7% drop in late February after Q4 results , then the March and July headline-driven sell-offs.

Management's Guidance: Optimistic, but Contradicted

• During the June 18, 2026 Investor Day, Besi raised its long-term revenue and operating margin targets, citing "increased demand for 2.5D AI-related datacenter and photonics applications as well as new use cases for hybrid bonding in logic, memory and co-packaged optics" .
• Management noted that overall market conditions and order momentum had "improved significantly since Q2-25" .
• The company highlighted "particular strength in hybrid bonding" in their Q1 order commentary .
• However, the July 6 report of client delays directly contradicted the positive narrative from Investor Day just three weeks earlier .

The Long-Term Outlook: Adoption, Not Abandonment

• Hybrid bonding is still the expected technology for 20+ stack HBM (HBM4E, HBM5) — the question is timing, not whether it will happen . Industry estimates show 16-layer HBM requires wafer thickness of ~30 μm vs. ~50 μm for current 12-layer designs, leaving almost no margin under the 775 μm cap .
• Samsung's HCB delivers 33% higher stacking capability vs. TCB, supporting 16+ layers .
• If JEDEC relaxes the standard to ~900 μm, that buys memory makers one to two more generations before hybrid bonding becomes mandatory .
• Besi continues to see strong non-HBM hybrid bonding demand in logic, photonics, and co-packaged optics, which provides a diversification buffer .
• The company's own raised long-term targets reflect management's confidence that hybrid bonding adoption will happen across multiple end markets, even if HBM timing slips .

Bottom Line

Besi's fundamental business is performing well — orders doubled year-over-year, margins are expanding, and management is raising long-term targets. But the stock is being pressured by repeated headlines that the pace of hybrid bonding adoption in HBM is slowing: first via the JEDEC thickness standard relaxation discussions in March, then via direct client delays in July. The long-term bull case (hybrid bonding becomes essential for 20+ layer HBM stacks and proliferates into logic and photonics) remains intact, but near-term adoption timelines are being pushed out, creating significant stock volatility. Investors are effectively weighing a strong present against a delayed future.

The Structural Issue: JEDEC Thickness Standard Discussions

The core concern for Besi's HBM-related revenue is that JEDEC is actively discussing raising the maximum package height for next-generation HBM, which would allow memory makers to continue using existing thermal compression bonding (TCB) instead of moving to hybrid copper bonding (HCB).

• In early March 2026, reports emerged that industry participants were discussing a target of 825–900 μm for 20-high HBM stacks (HBM4E, HBM5), compared to the current HBM4 cap of 775 μm .
• By late March 2026, JEDEC had begun "formal discussions" to relax these standards .
• By April 1, 2026, TrendForce reported that JEDEC was planning to relax height specs to approximately 900 μm for next-generation HBM, starting with HBM4E .
• The official JEDEC HBM4 standard (JESD270-4), finalized in February 2026, set the package thickness cap at 775 μm — already a relaxation from the previous ~720 μm, which made 16-layer HBM4 feasible without hybrid bonding .

The key implication: if JEDEC raises the limit to ~900 μm, memory makers could use existing TC bonders for at least one to two more generations before hybrid bonding becomes mandatory, pushing out Besi's revenue opportunity by years .

The Competitive Alternative: iHBM and HCB

Both major memory makers are developing packaging innovations that reduce the thermal urgency of hybrid bonding, though they approach the problem differently.

• SK Hynix unveiled iHBM in May 2026 — a thermal solution that embeds cooling elements directly inside the HBM package, built on its proven MR-MUF technology. The company states it reduces thermal resistance by 30% and is planned for next-generation products including HBM5 . MR-MUF itself offers better heat conduction and fewer voids than conventional methods .
• Samsung is pushing Hybrid Copper Bonding (HCB), which it claims improves thermal resistance by >20% over TCB and supports 16+ layers . Samsung also presented comparative data at ECTC 2026 .
• Critically, both Samsung and SK Hynix postponed hybrid bonder adoption for HBM4 mass production as of February 2026, choosing to use TC bonders instead . Samsung plans partial use of hybrid bonders for later HBM generations, while SK Hynix has established a pilot line and is evaluating Besi equipment .

Besi's Q1 2026 Performance: Fundamentally Strong

Besi's Q1 2026 results, reported on April 23, 2026, were objectively strong across the board:

• Bookings of €269.7M beat consensus estimates of ~€250–260M by approximately 4% .
• Besi guided for a 30–40% revenue increase in Q2 2026 .
• Revenue growth was driven by high-end mobile, 2.5D AI/photonics, and data center demand .

Despite these strong numbers, the stock faced downward pressure shortly after — a ~7% drop in late February after Q4 results , then the March and July headline-driven sell-offs.

Management's Guidance: Optimistic, but Contradicted

• During the June 18, 2026 Investor Day, Besi raised its long-term revenue and operating margin targets, citing "increased demand for 2.5D AI-related datacenter and photonics applications as well as new use cases for hybrid bonding in logic, memory and co-packaged optics" .
• Management noted that overall market conditions and order momentum had "improved significantly since Q2-25" .
• The company highlighted "particular strength in hybrid bonding" in their Q1 order commentary .
• However, the July 6 report of client delays directly contradicted the positive narrative from Investor Day just three weeks earlier .

The Long-Term Outlook: Adoption, Not Abandonment

• Hybrid bonding is still the expected technology for 20+ stack HBM (HBM4E, HBM5) — the question is timing, not whether it will happen . Industry estimates show 16-layer HBM requires wafer thickness of ~30 μm vs. ~50 μm for current 12-layer designs, leaving almost no margin under the 775 μm cap .
• Samsung's HCB delivers 33% higher stacking capability vs. TCB, supporting 16+ layers .
• If JEDEC relaxes the standard to ~900 μm, that buys memory makers one to two more generations before hybrid bonding becomes mandatory .
• Besi continues to see strong non-HBM hybrid bonding demand in logic, photonics, and co-packaged optics, which provides a diversification buffer .
• The company's own raised long-term targets reflect management's confidence that hybrid bonding adoption will happen across multiple end markets, even if HBM timing slips .

Bottom Line

Besi's fundamental business is performing well — orders doubled year-over-year, margins are expanding, and management is raising long-term targets. But the stock is being pressured by repeated headlines that the pace of hybrid bonding adoption in HBM is slowing: first via the JEDEC thickness standard relaxation discussions in March, then via direct client delays in July. The long-term bull case (hybrid bonding becomes essential for 20+ layer HBM stacks and proliferates into logic and photonics) remains intact, but near-term adoption timelines are being pushed out, creating significant stock volatility. Investors are effectively weighing a strong present against a delayed future.