Applied Materials announced on May 26 that SCREEN Holdings would join its EPIC Center as a new innovation partner, marking the third major equipment or foundry alliance at what is billed as the largest-ever U.S. private investment in semiconductor materials R&D. The EPIC Center, short for Equipment and Process Innovation and Commercialization, sits on Applied's Silicon Valley campus and is backed by a planned $5 billion capital spend. SCREEN, a Kyoto-based supplier of process equipment for semiconductors and displays, brings wafer cleaning, wet etch, and surface preparation capabilities to a facility that already hosts TSMC, Advantest, and Broadcom teams working on deposition systems, advanced substrates, and chiplet integration for AI accelerators. The announcement came alongside Applied's record Q2 FY2026 results: $7.91 billion in revenue, up 11% year-over-year, with 49.9% gross margin and guidance for Q3 at $8.95 billion.
The real substance here is not the partnership announcement, it is what the EPIC Center's existence signals about where the semiconductor industry's actual constraint has moved. Advanced packaging has displaced node scaling as the primary lever for performance gains in AI chips. TSMC's 3nm and Samsung's 3GAE offer marginal gains over their 5nm nodes; what matters now is how many chiplets fit in a package, how tightly they connect, and whether the substrate can handle the power delivery and thermal load. The bottleneck is not lithography or etch, it is the materials, deposition chemistry, and equipment that enable hybrid bonding at sub-micrometer pitch, electrochemical deposition on panel-format substrates up to 510×515 millimeters, and interconnect metallurgy that scales from 300mm wafers to larger form factors. Applied recognized this shift before the foundries fully articulated it. The company acquired ASMPT's NEXX business to add panel-level electrochemical deposition tools. It released the Precision Selective Nitride PECVD and Trillium ALD systems in April specifically for angstrom-level materials control in Gate-All-Around logic at 2nm and beyond, tools that matter far less to designers than to process engineers trying to hit yield and throughput targets.
SCREEN's entry into EPIC reinforces the center's positioning as the de facto industry standards body for next-generation materials and equipment. SCREEN supplies wafer cleaning, wet etch, and surface preparation systems used in substrate preparation, steps that look unglamorous but directly determine whether hybrid bonding works at scale. By co-locating SCREEN engineers with TSMC's process teams, Applied ensures that TSMC's packaging roadmap will reflect SCREEN's process IP, creating a lock-in effect. TSMC does not design its own deposition or cleaning equipment; it works with suppliers like Applied, SCREEN, and Tokyo Electron. But the EPIC Center flips the traditional vendor relationship. Instead of TSMC specifying what it needs and Applied building it, TSMC and SCREEN are now co-developing at the same facility, under Applied's roof, with Applied's tools as the anchor platform. Advantest joined the same center to link front-end manufacturing and back-end test workflows, another vertical slice that used to be customer-supplier, now collaborative inside Applied's infrastructure.
This model works for foundries because they gain early access to next-generation equipment before competitors. It works for Applied because it embeds its process IP into foundry roadmaps at the design stage, making switching costs astronomical. TSMC, Samsung, and Intel each hold dominant patent positions in hybrid bonding and chiplet interconnect, but those patents mean nothing if the deposition equipment, thermal processing, and materials engineering happen inside Applied's EPIC Center. Applied is not just selling tools; it is selling the entire materials stack, pre-validated against the foundry's most demanding packages. The catch: everyone else pays the cost. Startups attempting panel-format packaging, emerging foundries like GlobalFoundries or PSMC, and equipment suppliers competing with Applied face a maturity gap they cannot close without equivalent capital and co-location rights. The EPIC Center, in effect, accelerates the consolidation of materials engineering expertise into Applied's ecosystem.
What to watch over the next eighteen months: EPIC's operational status by end-of-2026 (the center is on track but equipment delays or real-estate complications could slip the timeline); adoption rates of panel-format substrates by TSMC and Samsung (panel adoption validates the entire materials roadmap Applied is building, and delays signal either process immaturity or foundry hesitation); and whether Intel, Samsung, or any consortium builds an equivalent materials R&D facility in response (the absence of a competing center by Q4 2026 would suggest the industry has accepted Applied's platform dominance in packaging materials). Applied's Q3 guidance of $8.95 billion also matters as a signal of demand sustainability, if AI capex softens, EPIC's $5 billion spend could stretch into 2027 or face reductions that starve equipment development.