NeuReality's NR1 Appliance Debuts, Qualcomm Expands to Abu Dhabi, and US Eyes Nuclear for AI - Silicon Brief #18

May 19, 2025 - Silicon Brief
Hi there, In this edition of Silicon Brief, we're diving into the latest developments in AI chips and data centers. NeuReality has launched its new NR1 AI Inference Appliance, aiming to enhance AI processing capabilities. Meanwhile, Cognichip has secured $33 million to advance its AI chip design, reflecting the growing investment in AI hardware. On the data center front, NetActuate is expanding its operations in Seattle to boost AI and edge connectivity, while Hon Hai predicts its AI server revenue will surpass NT$1 trillion, highlighting the increasing demand for AI infrastructure. Additionally, Qualcomm is set to open an AI and IoT engineering center in Abu Dhabi, further cementing its commitment to AI technology. In a significant partnership, Corning and Broadcom are collaborating on AI data center optics, which could lead to advancements in data transmission efficiency. These stories underscore the rapid advancements and investments in AI technology and infrastructure, shaping the future of the tech industry.

Energy & Sustainability

US Department of Energy has identified 16 sites for nuclear power and data centers to meet AI energy demands, focusing on expedited permitting. Constellation Energy announced its support for the US AI data center sector's energy needs and called for regulatory clarity from FERC, investing in nuclear energy, including a partnership with Microsoft to revive the Three Mile Island plant. Gradiant has secured contracts to provide sustainable water solutions for data centers in the U.S. and Indo-Pacific, focusing on reducing environmental impact and addressing water consumption challenges.

Technology

Delta Electronics, Inc. introduced new AI data center networking solutions at COMPUTEX 2025, focusing on efficiency and scalability with a Co-Packaged Optics Ethernet switch and next-generation switch design. Supermicro launched its Data Center Building Block Solutions (DCBBS) for liquid-cooled AI data centers, offering customizable infrastructure and cost efficiency. AIC is set to reveal AI storage technologies and GPU integration solutions at COMPUTEX 2025, showcasing advancements like PCIe Gen6 and CXL technologies for enhanced data processing efficiency.

Regulation & Policy

TrendForce reports that US and Chinese cloud service providers are developing in-house ASICs due to AI server demand and US export controls, potentially creating two distinct global AI server ecosystems.

The Brazilian government is proposing new policies to enhance the data center industry, potentially attracting hyperscalers like Oracle and AWS. The Redata policy may include import tax exemptions, reducing costs for international businesses.

The US Commerce Department has issued guidance warning that using Huawei's Ascend AI chips could result in criminal penalties due to export control violations. The chips are believed to contain US technology, prompting a global caution against their use.

The US Department of Commerce has rescinded the AI Diffusion Rule, which aimed to limit GPU exports, citing concerns over hindering American innovation. The department also highlighted that using Huawei Ascend chips violates US export controls.

Data Centers

NetActuate has expanded its Seattle operations to enhance AI and edge connectivity, while Hon Hai predicts its AI server revenue will surpass NT$1 trillion in 2025. Qualcomm is set to open an AI and IoT engineering centre in Abu Dhabi, and Dell Technologies has introduced an AI system for data centres. Additionally, atNorth will host 6G AI Sweden's AI Cloud in Stockholm, and Corning has partnered with Broadcom on AI data center optics.

We hope you enjoyed this article.

Consider subscribing to one of several newsletters we publish like Silicon Brief.

Also, consider following us on social media:

Subscribe to Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.