China has brought AI chips into its national security and reliability evaluation framework for the first time, turning what looks like a product certification process into something more consequential: an emerging gatekeeping system for AI computing infrastructure.
Synopsys reported divergent regional performance in the second quarter, with China showing sequential growth while North America and Europe declined, as demand patterns for semiconductor design tools and simulation software remained uneven across end markets.
Sino-American Silicon Products (SAS) held its shareholders meeting on May 26 and completed a full board overhaul, with founder Mingguang Lu stepping down as a director and Hsiu-lan Hsu being re-elected chairwoman. Lu will continue supporting the group as honorary chairman as SAS deepens its generational transition and diversified growth strategy.
Huawei's "Tau (τ) Scaling Law," unveiled at ISCAS 2026 on May 25, has become a global semiconductor flashpoint, drawing scrutiny from financial institutions, media outlets, chip analysts, and research firms.
Synopsys said the rise of agentic artificial intelligence and the integration of Ansys are creating new growth opportunities across electronic design automation (EDA) and simulation software, as semiconductor and industrial customers adopt more complex AI-driven engineering workflows.
Qualcomm has reportedly landed more than one ASIC customer, with a separate project for a US cloud service provider also taking shape, according to industry sources. The development follows reports from Bloomberg and other media that ByteDance, the developer of TikTok, is Qualcomm's first major client for its custom chip push.
Synopsys said growing demand for artificial intelligence chips from semiconductor companies and hyperscale data center operators is driving broader adoption of its electronic design automation (EDA), hardware-assisted verification, and intellectual property (IP) products, supporting a higher fiscal 2026 revenue outlook and reinforcing the company's position in AI-related semiconductor development.
On May 27, Marvell Technology reported record revenue in the fiscal first quarter of 2027 and raised its long-term outlook, citing accelerating demand for AI-related data center infrastructure, particularly in high-speed interconnect, Ethernet switching, and custom silicon solutions used in hyperscale cloud deployments.
Marvell Technology has said its custom silicon business and data center solutions are emerging as central pillars of its long-term growth strategy, driven by rising demand from hyperscale cloud customers building AI infrastructure, including XPUs, Ethernet switching systems, and high-speed optical interconnects.
Marvell Technology used its first-quarter fiscal 2027 earnings call to detail accelerating progress across optical, copper, and silicon photonics-based connectivity, signaling that AI infrastructure is shifting from compute-centric bottlenecks toward networking-driven architectures. The company's expanded interconnect portfolio, new silicon photonics acquisition, and emerging scale-up switch programs underpin a materially higher multi-year outlook.
Nvidia's new Taiwan headquarters is expected to begin construction by the end of 2026 and officially open in 2030, and CEO Jensen Huang is propagating that Taiwan "needs more electricity" to fund projects. With the AI cluster revolving around Nvidia, Taiwan Power Company (Taipower) estimates that power demand in the Beitou-Shilin Technology Park (BSTP) area will reach approximately 180MW.
More coverage
