NVIDIAGB200 Mass Production in December; AMD Going with One Unified GPU Architecture for Consumer…
AI Brief Headlines 9/11/2024
NVIDIAGB200 Mass Production in December; AMD Going with One Unified GPU Architecture for Consumer and Data Center Segments; iPhone 16 RAM Upgrade to 8GB; TSMC Revenue Up 33%;
AI Brief Headlines 9/11/2024
- NVIDIA Resolves Blackwell GPU Production Challenges, Mass Production Expected in December
- AMD Announces Unified UDNA Architecture to Challenge Nvidia’s CUDA Ecosystem
- iPhone 16 RAM Upgrade and Apple’s AI Push
- TSMC’s August Revenue Climbs 33% as AI Chip Demand Holds Up
NVIDIA Resolves Blackwell GPU Production Challenges, Mass Production Expected in December
A report from Commercial Times revealed that NVIDIA made changes to the 6-layer GPU mask of its Blackwell series, resolving production issues without needing to restart the taping-out process. This will help minimize delays, with the updated B200 expected to be completed by late October. Mass production of the GB200 chips is planned for December, and large-scale deliveries to ODMs are expected in the first quarter of next year. Previous delays were attributed to yield rates in advanced packaging, but recent adjustments have stabilized production. NVIDIA plans to use CoWoS-S for the GB200 this year.
AMD Announces Unified UDNA Architecture to Challenge Nvidia’s CUDA Ecosystem
At IFA 2024 in Berlin, AMD’s Jack Huynh, Senior VP and GM of the Computing and Graphics Business Group, revealed that AMD will unify its RDNA (consumer) and CDNA (data center) architectures into a single microarchitecture called UDNA. This unified architecture is designed to simplify development and improve performance, allowing AMD to better compete with Nvidia’s CUDA ecosystem. The shift is part of AMD’s strategy to deprioritize high-end gaming graphics cards and focus on expanding its market share, particularly in the data center and AI sectors.
iPhone 16 RAM Upgrade and Apple’s AI Push
The iPhone 16 and iPhone 16 Plus will come with an additional 2GB of RAM compared to the iPhone 15 series. However, the iPhone 16 Pro and Pro Max will not receive a memory boost. If RAM capacity is a key factor, waiting for the iPhone 17, rumored to include 12GB of RAM, might be wise. Meanwhile, Apple is advancing in AI technology, with its “Apple Intelligence” feature set launching next month via iOS 18.1, iPadOS 18.1, and macOS Sequoia 15.1, focusing on enhancing tasks such as writing, photo organization, and notifications. The AI features will debut in U.S. English, with additional languages and capabilities rolling out later.
TSMC’s August Revenue Climbs 33% as AI Chip Demand Holds Up
TSMC (Taiwan Semiconductor Manufacturing Company) saw its revenue increase by 33%, primarily due to sustained demand for AI chips. TSMC has benefited from the rising interest in artificial intelligence and related technologies, positioning it as a key supplier for companies developing AI hardware. This growth comes as part of a broader trend in the semiconductor industry, with increasing investments in AI and data-driven technologies.
메타데이터
- post_id
- 6ecbc0813eff
- slug
- nvidiagb200-mass-production-in-december-amd-going-with-one-unified-gpu-architecture-for-consumer-6ecbc0813eff
- url
- https://medium.com/byte-sized-ai/nvidiagb200-mass-production-in-december-amd-going-with-one-unified-gpu-architecture-for-consumer-6ecbc0813eff
- canonical_url
- https://medium.com/byte-sized-ai/nvidiagb200-mass-production-in-december-amd-going-with-one-unified-gpu-architecture-for-consumer-6ecbc0813eff
- author_url
- https://medium.com/@donmoon
- status
- ok
- fetched_at
- 2026-06-27 10:08:50