This week on Tom's Hardware Premium: October 10, 2026
This week on Tom's Hardware Premium, we covered how the AI buildout is affecting the industry supporting it across all axioms, including in law, chip design, manufacturing, and finance.
This week on Tom's Hardware Premium, readers can expect a comprehensive update on the intersection of artificial intelligence and the semiconductor industry. The featured report delves into how AI utilization may influence electricity consumption in data centers, with experts weighing in on software as a potential solution to the growing demand.
The article highlights Tencent's $7 billion lease deal for high-performance AI chips from Oracle, noting that the per-hour cost is estimated to be 43% lower than standard H100 rental rates, potentially allowing Chinese firms to access export-controlled hardware.
The piece also explores Synopsys' new partnership with OpenAI, unveiling a specialized AI model called GPT-Synopsys. This model combines Synopsys' domain IP knowledge and Electronic Design Automation (EDA) tools with OpenAI's expertise in designing large language models (LLMs), potentially streamlining chip design development. Moreover, the story examines the global supply disruption of neon gas, a crucial component in the DUV chipmaking process, and the efforts by companies like Gigaphoton and Kanten Techno to develop recycling systems.
In a surprising turn of events, Qualcomm has reportedly entered into a patent licensing agreement with Huawei for the first time, granting access to Huawei's LogicFolding chip-stacking methods. This move, which includes Huawei's LogicFolding chip architecture, could potentially disrupt the established pattern of Huawei licensing Qualcomm patents.
The article concludes by mentioning two former Groq employees who have filed a class action lawsuit against the company's former board and CEO, alleging that they breached their duties and cost stockholders billions following Groq's $20 billion deal with Nvidia.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.