Cerebras' AI chips have seen their first mainstream adoption by none other than OpenAI, as the AI lab reveals that their latest Codex model has another compute provider alongside NVIDIA. OpenAI Has Managed to Achieve a Shocking 1,000 TPS Output, With Cerebras' Blazing-Fast Bandwidth Well, there has been an NVIDIA-OpenAI saga on the financing front, but it appears that, in the race for compute, OpenAI has taken an interesting route through its earlier partnership with Cerebras. In the company's recent Codex release, it is disclosed that the GPT‑5.3‑Codex‑Spark is powered by Cerebras' AI chips and, more specifically, that the benefit […]

Read full article at https://wccftech.com/openais-latest-codex-model-runs-on-cerebras-infrastructure-hinting-at-a-serious-second-option-for-ai-inference-beyond-nvidia/

One of the major barriers to the adoption of China's DRAM and NAND products was the inclusion of CXMT and YMTC in the Pentagon's Section 1260H list, but in a new update, they have now been moved out. YMTC & CXMT Can Now Participate In US DoW Contracts, But Business With OEMs Isn't Entirely Risk-Free [Update]: The Pentagon has withdrawn the document that suggested updates to Section 1260H, and we haven't seen any updates from the Department of War (DoW) on it yet. [Original]: Well, it appears the prospect of manufacturers sourcing Chinese memory and NAND products has become much […]

Read full article at https://wccftech.com/cxmt-ymtc-removed-from-pentagon-list-opening-door-for-chinese-dram-adoption/

A European chip lab based in Barcelona has developed a computing architecture featuring RISC-V and an Intel process, marking it the first step towards "sovereign infrastructure". BZL's Latest Chip Features "Tenary" Architectures Onboard, To Refine Workload-Specific Performance There's always a debate in the mainstream technology industry about Europe's role in the AI revolution, given that the region has few "independent" ventures that have contributed to the computing race. While we won't go into this debate for now, the Barcelona Zettascale Lab (BZL) has completed an "experimental" build of its Cinco Ranch TC1 chip. According to the lab's report, this marks […]

Read full article at https://wccftech.com/european-chip-startup-pulls-off-working-risc-v-solution-on-the-intel-3-node/

China's SMIC has recently said that memory shortages are affecting broader supply chains, and the CEO said many customers are looking to scale back chip orders. SMIC Claims That Ditching Chip Orders in Light of DRAM Shortages Isn't the Wise Move For Customers DRAM has become one of the most important commodities in the AI world, given its rapid integration across AI architectures and end products. We have been witnessing shortages for several quarters now, and, apart from their impact on the client segment, large-scale foundries like SMIC are also seeing the effects. Talking about quarterly earnings, SMIC's co-CEO, Zhao […]

Read full article at https://wccftech.com/even-china-largest-chip-company-smic-isnt-safe-from-dram-shortages/

Samsung's return to the HBM market is a sight to see, and according to the latest announcement, its HBM4 process offers blazing-fast speeds and is now commercially deployed. Samsung's HBM4 Reaches Up to 13 Gbps, Making It Ideal For Integration With Vera Rubin The HBM industry is certainly seeing a shift in market dynamics, mainly in the distribution of market share among SK hynix, Samsung, and Micron. In particular, the race with the HBM4 technology has become much more competitive, and the Korean giant has improved its offerings by a huge margin. In the company's latest post, Samsung has revealed […]

Read full article at https://wccftech.com/samsung-hbm4-officially-makes-its-way-into-nvidia-vera-rubin/

NVIDIA's Blackwell platform has brought new levels of token optimization to AI inference workloads, as the company reveals a massive milestone in the realm of tokenomics. NVIDIA's GB200 NVL72 Achieves 10x Better Tokenomics Than Hopper, Credited "Expert-Level" Parallelism While NVIDIA has been racing to build new infrastructure in the AI world, one of the company's biggest focuses has been improving the efficiency of the hardware it deploys. And, with the Blackwell-trained frontier AI models dropping in the industry, we have seen how NVIDIA has progressed with token output and costs, and now, in a new blog post, the company has […]

Read full article at https://wccftech.com/nvidia-has-managed-to-reduce-token-costs-by-10x-with-blackwell/

HP, one of the biggest PC manufacturers, has given its verdict on the ongoing DRAM shortages, and the solution is renting gaming laptops. Yeah, you read it right. HP Is Now Offering Four Different OMEN Laptop Configurations to Rent, And You'll Never Own the Device The DRAM supercycle has disrupted the PC supply chain, and it appears manufacturers have few options on board to mitigate the risks posed by rising DRAM prices. One solution discussed in the industry is sourcing Chinese memory modules, which is thought to be a cheaper option, but we won't go into that for now. In […]

Read full article at https://wccftech.com/hp-response-to-memory-shortages-isnt-switching-to-chinese-suppliers-or-cutting-specs-its-asking-gamers-to-rent-laptops/

Rapidus, one of Japan's biggest native chipmakers, now plans to take its 2nm production ambitions to new heights, aiming to reach a respectable production capacity by 2028. Rapidus Plans to Achieve 60,000 WPM By 2028 For Its 2nm Process, Making It Japan's Most Ambitious Chip Fab The semiconductor industry is heating up, not just in terms of the competition it faces, but also in the demand foundries like TSMC are seeing amid the AI frenzy. While most chip orders are currently focused on the Taiwan chip giant, it appears that multiple competitors are seeking an 'opening window', among them Rapidus. […]

Read full article at https://wccftech.com/rapidus-accelerates-2nm-push-and-plans-dramatic-production-expansion/