Semiconductor equipment industry is about to change! Applied Materials and Fanlin are targeting Besi, aiming to acquire the "hybrid bonding" technology that is bottlenecked by AI and storage.

With BE Semiconductor’s proprietary advanced chip packaging technology—namely Hybrid Bonding—an advanced packaging technique—becoming increasingly critical for AI chips and high-performance data center storage systems like HBM amid the global AI superwave, the world’s two major semiconductor equipment manufacturing giants—namely the U.S.-based Applied Materials and Lam Research—are both emphasizing their interest in acquiring this European semiconductor equipment company. Media outlets citing insiders report that BE Semiconductor, headquartered in the Netherlands, is working with Wall Street investment giant Morgan Stanley to discuss the acquisition inquiries it has received.

An insider further added that Lam Research has already issued an acquisition interest; another said that Applied Materials also has a strong acquisition interest. Previously, in April 2025, Applied Materials acquired about 9% of BE Semiconductor. At that time, Applied Materials stated that this stake was viewed as a “strategic long-term investment, demonstrating Applied Materials’ commitment to jointly developing the industry’s most capable hybrid bonding packaging solutions—this technology is becoming increasingly important for the capacity and architecture upgrades of the most fundamental logic and memory chips in AI training and inference systems.”

On Friday, Applied Materials announced a 15% increase in its quarterly dividend, to $0.53 per share, up from $0.46 per share previously. Earlier, on Tuesday Eastern Time, Applied Materials, along with leading storage chip maker Micron Technology and South Korea’s SK Hynix, announced a major collaboration to develop and build cutting-edge solutions and upgrade routes for DRAM, high-bandwidth memory (HBM), and data center NAND storage systems, aiming to comprehensively enhance overall storage chip capacity and the performance of AI training/inference systems.

In February, Applied Materials released its latest earnings report, showing that the company’s nearly complete suite of high-end semiconductor equipment delivered better-than-expected quarterly results and extremely strong future guidance. This highlights that, amid the booming global AI infrastructure build-up and the macro backdrop of a “super cycle” in storage chips, semiconductor equipment manufacturers are entering a super-growth cycle, becoming the biggest beneficiaries of the rapid expansion trend in AI GPU/AI ASIC and DRAM/NAND storage chip capacities.

This American largest-scale semiconductor manufacturing and advanced packaging equipment supplier expects its revenue for the second quarter of fiscal 2026 to be approximately $7.65 billion, with a fluctuation range of about $500 million. In comparison, Wall Street analysts’ average revenue forecast for this quarter—up to April of this year—is $7.03 billion. It’s worth noting that, as production of advanced 3nm and below AI chips, CoWoS/3D advanced packaging capacity, and DRAM/NAND storage chip capacity accelerate, Applied Materials’ revenue outlook has been continuously raised by analysts since the beginning of this year. The second quarter’s non-GAAP earnings per share are projected to be between $2.44 and $2.84, excluding some items, far above the analyst average of $2.29.

Applied Materials’ non-GAAP EPS for the first quarter was $2.38, higher than the Wall Street average of $2.21, and roughly flat compared to the same period last year; the company’s gross margin in the first quarter reached 49%, up from about 48% in the same period last year; its non-GAAP free cash flow in the first quarter was $1.04 billion, representing a significant 91% increase.

The “new wealth code” in the chip industry—hybrid bonding! This technology is the “neck” of AI chips and memory chips.

Advanced packaging has become the new engine driving continuous computing power improvements in the “post-Moore era,” with bonding technology performance directly determining the upper limit of integrated systems. Bonding technology has evolved from wire bonding, flip-chip, thermocompression bonding, to fan-out packaging, ultimately advancing into the hybrid bonding era. Hybrid bonding advanced packaging has become one of the core technologies to break through computing power and bandwidth bottlenecks and reshape the value chain of the AI computing industry. This technology is widely used in AI chips and HBM—such as NVIDIA’s Blackwell series AI GPUs, Broadcom’s AI ASICs, SK Hynix’s HBM3E/HBM4, and soon in high-performance Ethernet switches integrated with CPO technology, playing a key role in advanced packaging.

Hybrid bonding replaces traditional bump bonding with copper-to-copper direct bonding, achieving ultra-fine interconnect pitches below 10μm, bringing order-of-magnitude improvements in interconnect density, bandwidth, energy efficiency, and cost per interconnect. It is a critical breakthrough supporting 3D stacking and heterogeneous integration. The process is divided into wafer-to-wafer bonding—suitable for uniform small chips like memory—and chip-to-wafer bonding—suitable for large chips and heterogeneous integration. BE Semiconductor’s hybrid bonding advanced packaging equipment is widely adopted by leading chip manufacturers such as TSMC, SK Hynix, and Samsung, mainly used in advanced chip packaging stages. Whether for 2.5D CoWoS or more advanced 3D/3.5D packaging, hybrid bonding equipment is essential.

In the AI era, the bottleneck is increasingly not individual transistors but data transfer efficiency between chips, logic, memory, and within packaging. A report from Applied Materials explicitly states that hybrid bonding, through direct copper-to-copper interconnects, increases interconnect density and shortens wiring lengths, thereby improving performance, power consumption, and cost simultaneously; this direct interconnection enables faster data transfer and lower power consumption. For high-performance computing chips like GPUs, AI ASICs, CPUs, and accelerators, this is crucial for whether packaging can continue to enhance bandwidth density and energy efficiency.

For HBM and high-end enterprise storage, hybrid bonding is equally critical, especially for future memory/storage routes with higher layer counts, higher bandwidth, and lower power consumption. Applied Materials’ Kinex materials explicitly mention that their systems are designed for future HBM and logic-memory integration, with demonstrated stacking capabilities of 16 layers or more of DRAM; Besi states that their customer roadmap over the next two years will further incorporate hybrid bonding/TC Next into HBM4/4e.

In massive AI data centers, high-performance networking itself has become part of computing. Besi has listed co-packaged optics (CPO) and ASICs as new use cases; Broadcom previously stated that integrating optical engines with switch chips into a common package on its 102.4Tbps Tomahawk 6 CPO switch chip can significantly improve power efficiency, link stability, and support large-scale AI cluster expansion.

The global semiconductor equipment industry is about to change

The essence of Applied Materials and Lam Research’s bid for BE Semiconductor is not just a typical acquisition but about who can seize the most scarce packaging “bottleneck equipment control” in the AI era. Besi’s “world’s highest precision” hybrid bonding equipment means it is not an ordinary packaging manufacturer but one of the most scarce core assets in advanced packaging in the AI/HPC era.

As tech giants like Microsoft, Google, and Meta lead the global mega-scale AI data center build-out, driving chip manufacturers’ rapid expansion of 3nm and below AI chips, CoWoS/3D advanced packaging capacity, and DRAM/NAND storage chip capacity, the long-term bullish logic of the semiconductor equipment sector becomes increasingly solid.

According to the latest analyst expectations compiled by institutions, Amazon, along with Google’s parent Alphabet, Facebook’s parent Meta Platforms Inc., Oracle, and Microsoft, are projected to have a combined AI-related capital expenditure of about $650 billion by 2026, with some analysts estimating total spending could exceed $700 billion—implying a year-over-year increase of over 70% in AI capital expenditure.

If either Applied Materials or Lam Research ultimately acquires Besi, it would signify a major shift in the semiconductor equipment industry landscape. This would also mark the formal upgrade of advanced packaging from a “back-end process” supporting role to a core battlefield alongside front-end processes. For Applied Materials, this means integrating cleaning, activation, metrology, and platform assembly with Besi’s core bonding into a more complete “full-stack hybrid bonding solution”; for Lam Research, it means further deepening into the most core value capture segments of advanced packaging, building on existing TSV, RDL, high aspect ratio etching/deposition, and hybrid bonding support tools. Whoever wins, the industry’s future will more clearly head toward a single direction: the competition for AI chips and storage chips will not only depend on the iteration of advanced semiconductor equipment like EUV, etching, and thin-film deposition but also on who can turn advanced packaging—the most challenging last mile—into their “super moat.”