Artificial intelligence companies warn of bottlenecks
Taiwanese AI supply chain companies warned of emerging bottlenecks in semiconductor materials, memory chips, and advanced packaging during a Computex Taipei news conference, despite expecting strong industry growth driven by AI demand. Executives from Alchip Technologies, Nanya Technology, Msscorps, and Unimicron highlighted challenges like angstrom-era process scaling, DRAM bit growth exceeding 20 percent annually, and AI-related revenue accounting for over 60 percent of some firms’ total sales.
Taiwanese companies in the artificial intelligence (AI) supply chain expressed optimism about industry growth but warned of critical bottlenecks during a news conference at Computex Taipei. Executives from chip designer Alchip Technologies Ltd, memory maker Nanya Technology Corp, semiconductor analysis provider Msscorps Co, and substrate supplier Unimicron Technology Corp outlined challenges in memory chips, advanced packaging, and semiconductor material analyses as AI investment accelerates. Msscorps Co identified material analysis as the next scaling bottleneck, with CEO Gene Liu noting increased demand for advanced characterization technologies due to angstrom-era processes and AI-driven computing. The shift requires expanded laboratory capacity and proprietary expertise, according to Liu. Nanya Technology Corp expects AI demand to drive over 20 percent annual DRAM bit growth, prompting plans to increase R&D spending by up to 70 percent and double output through capacity expansion. The company recently secured $2.5 billion in private placement funding, adding to its $5.8 billion cash reserves to support growth. Unimicron Technology Corp reported AI-related business will account for over 60 percent of its revenue this year, driven by demand for advanced packaging materials like Ajinomoto Build-up Film substrates. Meanwhile, Alchip Technologies Ltd stated AI-related projects now make up 80 percent of its revenue, fueled by cloud providers seeking custom AI chips for efficiency gains. The Taiwan Stock Exchange (TWSE) hosted the event to highlight Taiwan’s semiconductor ecosystem, with plans to feature about 40 IPO applicants this year, 40 percent of which are AI-related. Executives also noted that slowing Moore’s Law progress has increased reliance on advanced packaging technologies like 2.5D packaging and chip-on-wafer-on-substrate solutions.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.