AI Weekly Malaysia

Nvidia is reportedly testing reduced memory configurations for its upcoming Rubin Ultra AI accelerator due to HBM supply shortages, with designs including as little as 192 GB and a step back to HBM4 from a more advanced memory type. The report is based on supply-chain rumors, not official confirmation.

Why: If Rubin Ultra ships with less memory than originally planned, AI/ML teams building large-model inference or training pipelines should factor tighter VRAM ceilings into their 2026-2027 infrastructure roadmaps — especially in SEA where GPU access is already constrained by allocation priority. Founders budgeting for next-gen GPU rentals or cloud instances should not assume memory specs will scale up linearly from current Blackwell-class hardware.

Samsung Foundry has updated its semiconductor process roadmap, pushing its 1.4nm node to 2029 and planning to use high-NA EUV lithography for 1nm-class and smaller nodes from 2030 onward. The article text itself is almost entirely Tom's Hardware site navigation and subscription boilerplate, with the substantive detail confined to the headline.

Why: For software builders, this is a distant hardware supply-chain signal with no near-term action item. The only useful takeaway is that advanced node capacity (which affects future AI chip cost and availability) is slipping to late 2029–2030, meaning any assumption about dramatically cheaper inference hardware before then should be tempered. No decision needs to be made this week.