In this edition of Silicon Brief, we're diving into the latest developments in AI chips and data centers. Nvidia has reported a significant $39.1 billion in data center sales, despite facing restrictions in China. Meanwhile, TSMC is expanding its global footprint with plans to open a chip design center in Munich. These moves highlight the ongoing global demand for advanced semiconductor technologies.
On the data center front, Google is making waves with a $7 billion investment in Iowa, while Microsoft has opened its first data center in Indonesia. These investments underscore the growing need for robust infrastructure to support AI advancements. Stay tuned as we explore these stories and more in this edition.
Energy & Sustainability
Soluna Holdings, Inc. has launched Project Annie, a 75 MW solar-powered data center in northeast Texas, co-located with a 114 MW solar farm. Pacific Gas and Electric Company projects that increased electricity demand from data centers could lower customer electric bills by 1-2%, collaborating with Westbank on projects in San Jose. xAI has installed 168 Tesla Megapack units at its Colossus supercomputer in Memphis to enhance power reliability. Recogni and DataVolt have partnered to create an energy-efficient AI cloud infrastructure. CenterPoint Energy has boosted its capital plan to $52.5 billion to support AI data center power needs.
Strider Technologies and SCSP report on China's strategic AI expansion, detailing over 250 AI data centers with military links. Texas Instruments partners with NVIDIA to develop power solutions for future AI data centers. Gcore appoints AI expert Prof. Dr. Feiyu Xu as Board Advisor. The AI-powered edge networking market is projected to reach $82.52 billion by 2029, according to The Business Research Company. GMI Cloud becomes a Reference Platform NVIDIA Cloud Partner, enhancing its AI project capabilities. MinIO supports AWS S3 Express API on its AIStor platform, improving AI workload performance. Drut Technologies partners with Ranovus for Co-Packaged Optics solutions in data centers. Nvidia plans an 800V HVDC power infrastructure for AI servers by 2027.
Applied Digital is planning an AI data center in Deuel County, South Dakota, which is expected to create jobs and stimulate local development, though there are concerns about resource usage. Nvidia reported a revenue of $44.1 billion for Q1 fiscal 2026, despite facing an $8 billion loss due to U.S. sales restrictions impacting its H20 product line. Meanwhile, EdgeCortix, a Tokyo-based AI chip startup, has secured ¥3 billion in subsidies from the Japanese government to advance energy-efficient AI chiplets for defense applications.
Google has committed $7 billion to expand its data center presence in Iowa, while Microsoft has launched its first data center in Indonesia. A CBRE report highlights a shortfall in AI-ready data centers in Asia Pacific, and Vena Global has launched a subsidiary to address this demand. Greykite is transforming a Warsaw site into a data center for EdgeConneX, and Humain seeks a US partner for its data center expansion. Meanwhile, Telus plans a C$70 billion investment in network and AI centers in Canada, and Nvidia and Wistron are considering Fort Worth for a new AI supercomputer plant.
Nvidia reported $39.1 billion in data center sales, despite $4.5 billion in charges due to Chinese export restrictions. TSMC is set to open a chip design center in Munich by 2025, focusing on high-performance AI chips. Saudi Arabia's Humain has partnered with Nvidia for 18,000 chips for a new AI data center. The DGX B200 Blackwell node set a world record with over 1,000 transactions per second per user. Tencent claims a sufficient chip stockpile for AI amid US restrictions. AMD unveiled the Radeon AI PRO R9700 GPU at Computex 2025. Vaire Computing developed an energy-recycling chip, while Silicon Data launched an AI chip rental cost index. Marvell Technology, Inc. introduced an advanced packaging platform for AI, and Aion Silicon secured $12 million for a RISC-V accelerator program. A report by Kearney and SEMI discusses AI's impact on semiconductors.