➀ AMD unveiled a series of AI products including a new flagship AI chip, server CPU, AI network card, DPU, and AI PC mobile processor; ➁ The flagship AI chip AMD Instinct MI325X GPU uses HBM3E high bandwidth memory, with AI peak computing power reaching 21PFLOPS; ➂ AMD disclosed the latest AI chip roadmap with the MI350 series using CDNA 4 architecture, expected to be launched next year; ➃ AMD launched the world's first AI network card supporting UEC and the Pensando Salina 400 DPU, doubling performance; ➄ The fifth-generation EPYC server CPU supports up to 192 cores and 384 threads, with a top price of 14813 USD per core; ➅ AMD's third-generation commercial AI mobile processor Ryzen AI PRO 300 series is expected to be on the market with more than 1 million units by 2025.

➀ TSMC reported revenue of $23.62 billion in the July-September period, a 36% year-on-year increase; ➁ September sales reached $7.74 billion, up 39.6% from the same period last year; ➂ Half of TSMC's revenue now comes from HPC, with growth driven by AI ICs, and demand for which depends on large AI investors finding paying customers for AI products.

➀ BrainChip introduces the Akida Pico NPU, an ultra-low-power neural processing unit for battery-powered devices; ➁ The Akida Pico consumes less than a milliwatt and is designed for applications like hearing aids and medical equipment; ➂ The NPU supports popular AI frameworks like TensorFlow / Keras and PyTorch, enabling quick deployment of AI solutions.

➀ Shanghai Siyuang Technology Co., Ltd. (Siyuang Technology) has completed the listing guidance registration with Shanghai Securities Regulatory Bureau and is preparing for an IPO. Siyuang Technology is a super unicorn in the AI chip field, founded by Zhao Lidong, a graduate of Tsinghua University's Radio and Telecommunication Engineering Department 85th class. The company has raised nearly 7 billion yuan in financing, with Tencent being one of the largest investors. According to Hurun Research Institute, Siyuang Technology is valued at 16 billion yuan, ranking 482nd in the Global Unicorn List. ➁ Siyuang Technology has chosen to focus on AI training chips, a field with high difficulty and high barriers. The company has developed a series of AI training and inference chips, providing underlying computing power for Internet companies, cloud service providers, operators, and research institutions. ➃ The domestic AI chip market is facing challenges such as high cost of migration and technical support. Siyuang Technology needs to consider how to commercialize its computing centers. The article discusses the challenges and opportunities for Siyuang Technology and other domestic AI chip companies as they prepare for the IPO.

➀ Cerebras Systems, a company specializing in AI training and inference chips, has filed for an IPO on the NASDAQ with the ticker symbol 'CBRS'. It is known for its wafer-scale engine (WSE) chips, the largest in the world.➁ The company reported a net loss of 3 billion USD (about 21 billion RMB) from 2022 to 2023, but its revenue increased nearly 1500% to 1.364 billion USD in the first half of 2024.➂ Cerebras' chips have four technical advantages for AI: large-scale, tightly coupled computing; fast and efficient interconnection; ultra-large on-chip memory and memory bandwidth; and local hardware acceleration for sparsity.

➀ The Mecpow X4 Pro offers a powerful 22-watt diode laser for engraving and cutting; ➁ It comes with an array of safety features and an integrated webcam; ➂ The device lacks an integrated display and has issues with the acrylic enclosure and software experience.

➀ Huawei has started testing its new AI chip, Ascend 910C, with potential clients in China; ➁ The chip is an upgraded version of the 910B and is being offered to large Chinese internet firms; ➂ Huawei's chips are seen as alternatives to Nvidia's GPUs in multiple industries, with the Ascend solutions used in training over half of China's top large language models.

➀ Intel's Gaudi 3 AI accelerator is now generally available, aiming to compete in the scale-out AI market; ➁ The product is expected to be deployed in systems from several vendors, including Dell and HPE; ➂ Intel is focusing on building clusters with up to 1024 nodes or 8192 accelerators to cater to enterprise needs.

➀ This research explores the potential of two-dimensional silk fibroin films in electronics; ➁ The study investigates the growth mechanisms of silk fibroin films and their interactions with substrates; ➂ The findings highlight silk fibroin's potential as a versatile biomaterial for electronic applications.

➀ Tesla's BEV market share dropped from 19% to 17% in the first five months of 2024; ➁ BYD is just one point behind Tesla at 16%; ➂ The BEV portfolio of legacy automakers like Volkswagen and Geely-Volvo stood at 7 and 8 percent, respectively.

1. Nvidia's second-quarter sales surged by 122.4% to $30.04 billion, significantly outperforming expectations. 2. The company provided strong sales guidance for the third quarter, anticipating a 2.5% increase above analyst estimates. 3. Nvidia's dominance in the AI chip market, particularly with its H100 GPU, positions it well for continued growth as major tech companies like Google, Meta, and Microsoft heavily invest in its technology.

➀ AMD MI300X features Infinity Fabric Advanced Package for high-speed interconnects and supports FP8 operations. ➁ Intel Gaudi 3 introduces a 256x256 matrix multiplication engine and enhanced interconnectivity. ➂ SambaNova SN40L offers large on-chip memory and a hybrid Mesh/Ring interconnect structure. ➃ FuriosaAI introduces a Tensor Contraction Processor based on Einstein summation notation. ➄ Tenstorrent's architecture emphasizes asynchronous memory access and a scalable 2D/3D Torus network. ➅ BRCM focuses on CPO technology for high-density optical interconnects. ➆ IBM Telum 2 integrates a DPU and AI accelerator in its next-gen mainframe processor.

➀ SurgeForce Technology, a Shanghai-based AI chip startup, is preparing for an IPO on the A-share market with a valuation of 16 billion yuan. ➁ The company has secured nearly 7 billion yuan through ten rounds of financing, with Tencent participating in six consecutive rounds. ➂ SurgeForce specializes in cloud computing power products for the AI sector, aiming to create a computing power foundation for general artificial intelligence.

➀ SambaNova introduces SN40L, an AI chip system supporting 5 trillion parameter models, with performance 3.1x and 2x that of NVIDIA's H100 in inference and training, respectively, at only 1/10 the cost. ➁ The SN40L chip, built on TSMC's 5nm process, features 1020 billion transistors and 1040 custom 'Cerulean' RDU cores, offering 638TFLOPS (BF16). ➂ SambaNova's system includes 520MB of on-chip SRAM, 64GB of HBM, and 1.5TB of external memory, enabling support for large-scale model training and inference.

➀ Suyi Technology, a Shanghai-based AI chip company, has filed for IPO and is under辅导备案 with the China Securities Regulatory Commission. ➁ Founded in 2018, Suyi focuses on cloud and edge computing for artificial intelligence, serving major clients including Tencent and China Mobile. ➂ The company has completed over 51 billion yuan in financing, with Tencent as its largest shareholder, holding a 21.37% stake.

➀ TSMC has acquired a plant and equipment from Innolux Corp. for NT$17.14 billion. ➁ The facility is located in the Southern Taiwan Science Park, covering an area of 317,445 square meters. ➂ This acquisition aims to boost TSMC's AI chip packaging capacity.

➀ Huawei's new AI chip, Ascend 910C, is compared to NVIDIA's mid-to-low-end chips, but not yet capable of challenging high-end models. ➁ NVIDIA's path to the Chinese market is blocked by Huawei's strategy. ➂ The pricing of Huawei's 910C chip is expected to be around $30,000, which is competitive against NVIDIA's similar offerings at nearly $40,000.

➀ A seasoned AI chip salesperson, Junjie, decides to leave the industry due to repeated failures in selling AI chips, attributing the challenges to NVIDIA's dominant ecosystem and inferior performance of domestic chips. ➁ Another AI chip salesperson, Shuhao, transitions from a research role, believing in the potential of AI and hoping to gain more control over his career. ➂ Many in the AI chip industry face despair, confusion, and numbness, but some remain hopeful, striving to build a domestic AI chip ecosystem despite the difficulties.

➀ NVIDIA's next-generation Blackwell AI chips are delayed due to design flaws, affecting major clients like Meta, Google, and Microsoft. ➁ The production was initially planned for Q3 2024 but was halted by TSMC engineers discovering issues. ➂ Despite initial optimism, the delay is more significant, impacting the deployment plans of companies like Supermicro.

➀ The semiconductor industry is increasing silicon usage for AI problems; ➁ Ausdia is focusing on ensuring proper timing in chip functionality.