- Apple is in talks with Micron, Tata Group and other Indian chipmakers to buy $12 billion worth of Indian-made chips, according to Indian media reports.
- According to Tom’s Hardware, at IFA 2024 in Berlin, Germany, AMD announced that the company will unify the consumer-centric RDNA and data center-centric CDNA architectures into a microarchitecture called UDNA, which will lay the foundation for the company to more effectively deal with Nvidia’s entrenched CUDA ecosystem. At the time of this news, AMD has decided to deprioritize high-end gaming graphics cards to accelerate market share growth.
- TSMC released its revenue report for August 2024, with revenue of US$7.792 billion, a decrease of 2.4% from July and an increase of 33% from the same period in 2023, setting the second highest monthly record in history. The total revenue in the first eight months of 2024 reached US$55.102 billion, an increase of 30.8% from the same period in 2023.
- According to foreign media reports, David Dahan and Ran Halutz, co-founders of Habana Labs, which Intel acquired for $2 billion on December 16, 2019, have left Intel and founded another artificial intelligence company called Touch in Tel Aviv, Israel, co-founded by Avigdor Willenz, David Dahan and Ran Halutz.
- Synopsys announced the industry’ first complete UCIe IP solution, running at up to 40 Gbps per pin to meet the growing computing performance requirements of the world’s fastest artificial intelligence data centers.
- SK Hynix announced that it has developed the high-performance solid state drive (SSD) product ‘PEB110 E1.S’ for data centers.
- The EU plans to reduce additional tariffs on Chinese-made electric vehicles, with Tesla’s tariff reduced to 7.8%.
- TSMC to set up factory in Germany, German Economic Office: at least 25 supply chain companies are interested in following suit.
9. Micron has launched a mass-produced 12-layer HBM3E chip for the next-generation artificial intelligence (AI) GPU, with a memory stack capacity of up to 36GB and a speed of over 9.2Gb/s. Micron said that it is currently sending production-capable 12-layer fifth-generation HBM (HBM3E) to important partners in the AI industry chain for verification procedures.
