The Download: glass chips and “AI-free” logos

· Source: MIT Technology Review · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Cybersecurity & Data Privacy · Depth: Fundamental Awareness, short

Summary

Absolics, a South Korean company, will begin producing specialized glass panels in 2026 for next-generation AI chips, with Intel also exploring this technology. These glass-based chips aim to enhance computing power and efficiency while reducing energy consumption in data centers, consumer laptops, and mobile devices. Concurrently, there is a global push to establish an "AI-free" logo for human-made products, alongside a "QuitGPT" campaign. Senator Elizabeth Warren is seeking information regarding xAI's reported access to classified military networks. Other notable developments include Meta's planned layoffs affecting over 20% of staff, ByteDance delaying a video AI model due to copyright disputes, and a Chinese AI startup achieving an $18 billion valuation. Cybersecurity investigators also exposed a North Korean scam involving remote US jobs.

Key takeaway

For CTOs and VPs of Engineering evaluating future data center infrastructure, consider the potential of glass-based AI chips from companies like Absolics and Intel. This technology promises substantial energy efficiency gains and performance improvements, which could impact your hardware procurement strategies and operational costs by 2026. Additionally, monitor the evolving landscape of AI ethics and data access regulations, particularly concerning military data, to ensure your organization remains compliant and responsible.

Key insights

Glass substrates are emerging as a key technology for more efficient and powerful AI chips.

Principles

In practice

Topics

Best for: CTO, VP of Engineering/Data, Director of AI/ML, General Interest, Tech Journalist, Software Engineer

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT Technology Review.