Besi Q2 Revenue Hits €250M as AI Packaging Demand Reshapes Customer Base

BE Semiconductor Industries N.V. reported second-quarter revenue of €249.9 million, positioning the Dutch equipment maker at the center of a shift in semiconductor capital expenditure toward AI-driven advanced packaging.

The company guided third-quarter revenue to €275-287 million, a 10-15% sequential increase, as demand for hybrid bonding and photonics equipment accelerates beyond GPU infrastructure into chip-to-chip interconnects and high-bandwidth memory integration.

Customer Mix Transformation

Foundries and subcontractors now represent 63% of orders in the twelve months through Q2-26, up from 45% in 2024. This migration reflects semiconductor manufacturers retooling production lines for AI accelerator assembly, where traditional IDMs cede ground to TSMC-class foundries and OSATs controlling advanced packaging capacity.

First-quarter 2026 bookings reached €269.7 million, a 104.5% surge compared to Q1-25.

Cyclical Headwinds Persist

Not all segments participate equally. Q2-25 revenue of €148.1 million was up 2.8% sequentially but down 2.1% year-over-year, pressured by mobile market softness that continues to weigh on mainstream computing end-markets. Besi's strategic pivot toward hybrid bonding technology aims to offset this cyclicality with AI-specific demand less correlated to consumer device refresh cycles.

Analyst consensus estimates for full-year 2026 revenue have been revised upward from $0.87 billion to $0.94 billion. Earnings per share estimates remain variable across sources.

Why It Matters

Besi operates as a bellwether for AI capital expenditure beyond Nvidia and AMD headlines. The surge in advanced packaging bookings indicates the AI build-out is entering its second phase: after GPU fabrication comes the equally critical work of stacking memory, connecting chiplets, and managing thermal density at the package level.

Hybrid bonding—Besi's core focus—enables chip-to-chip interconnects at micron-scale pitch, essential for scaling AI performance as Moore's Law slows. The 18-percentage-point shift toward foundry customers in one year suggests AI supply chains are consolidating around specialists capable of these processes.

For investors watching semiconductor equipment, Besi's order book signals where the next tranche of AI infrastructure spending will flow: not just to lithography and fab tools, but to the assembly technologies that make AI chips actually function at scale.