1. AMD's share price has declined by -53.4% over the last year compared to Nvidia's +22.3% return. 2. AMD's MI325X platform may outperform Nvidia's Hopper H200 GPUs in some inference applications. 3. AMD is well-positioned to benefit from the growth of the AI inference market with the release of the MI350 architecture and the MI400 platform.
Recent #NVIDIA news in the semiconductor industry
➀ TSMC is in talks with Intel, Nvidia, Broadcom, Qualcomm, and AMD about a joint venture to acquire Intel Foundry;
➁ TSMC plans to take a stake of no more than 50% in Intel Foundry to maintain US ownership, as required by the White House;
➂ The deal requires approval from the White House.
➀ Neil Druckmann discussed the evolving role of storytelling in video games with screenwriter Alex Garland, reflecting on the challenges and inspirations around narratives in games.
➁ Games like BioShock and The Last of Us demonstrated the full capability of narrative in gaming, challenging the industry's skepticism.
➂ Druckmann shared his early inspirations for narrative-driven games, including Monkey Island 2 and Half-Life 2.
➀ The scarcity of NVIDIA's GeForce RTX 50 series GPUs has led to inflated prices on platforms like eBay.
➁ Smugglers in China are selling illegal GeForce RTX 5090 graphics cards at high prices to meet the demand for DeepSeek's R1 AI model.
➂ Despite US sanctions, the RTX 5090D variant is available in China, but it has lower AI performance.
➀ Infineon is now supporting Nvidia Tao AI models on its Psoc Edge microcontrollers, which combine Arm Cortex-M55 CPUs with Arm Ethos-U55 NPU.
➁ The combination of these microcontrollers with Tao models and a fine-tuning toolkit greatly simplifies the creation, customisation, optimisation and deployment of high accuracy vision AI models.
➂ Infineon claims that the integration of Nvidia Tao features will become available for the entire PSOC Edge family along with a development ecosystem, including tools, libraries and documentation.
➀ The specifications of Nvidia's upcoming budget GPUs, RTX 5050, RTX 5060, and RTX 5060 Ti, have been leaked, showing minimal changes and a focus on VRAM.
➁ The RTX 5060 Ti is expected to come in 8GB and 16GB flavors with a 180W TGP.
➂ The RTX 5060 is rumored to have an 8GB configuration and a 150W TGP.
➀ A team from Saarland University's Faculty of Engineering and Medicine is developing smart implants at the university and university hospital to monitor and promote healing in body bone fractures.
➁ The robotic implants can stiffen and soften through shape-memory technology, allowing for permanent monitoring of fracture healing.
➂ The research team is miniaturizing the technology for use in market screws as part of an EU project.
➀ NVIDIA将为Stargate AI超级计算机提供64,000个新的GB200 AI GPU，以支持OpenAI和Oracle；
➁ 这些GPU预计将在2026年交付，总价值22.4亿美元；
➂ Stargate AI超级计算机位于德克萨斯州阿比林，将与Elon Musk的xAI超级计算机竞争。
➀ NVIDIA plans to release two GeForce RTX 5060 Ti models with 8GB and 16GB of GDDR7 memory;
➁ The cards will be powered by the GB206-300-A1 RTX Blackwell chip with 4608 CUDA Cores;
➂ The performance is expected to be similar to the RTX 4070, with a rumored launch in April or May 2025.
1. Nvidia reported impressive quarterly results, with revenues reaching $39 billion and a forecast of $43 billion for FQ1'26, significantly beating analyst estimates. 2. Despite substantial AI data center spending by major tech companies, Nvidia's stock trades lower than tech peers like Apple and Broadcom. 3. Nvidia's gross margins are expected to normalize to around 75% with the Blackwell chip ramp-up, which will significantly impact EPS and reinforce the bullish investment thesis. 4. The stock is undervalued at 20x FY27 EPS targets despite strong AI chip growth and robust revenue forecasts.
➀ A shipping manifest suggests NVIDIA is preparing to announce a new generation of workstation graphics cards based on the Blackwell architecture.
➁ The RTX Pro X Blackwell features a 96GB GDDR7 memory and a 512-bit bus.
➂ The non-X model has a 48GB GDDR7 memory on a 384-bit bus.
➀ Nvidia has released another Hotfix for the RTX 50-series GPUs to address crashes and clock speed issues;
➁ The GeForce Hotfix Driver Version 572.75 tackles specific problems with the RTX 5090 and 5080 GPUs;
➂ Users are advised to manually download and install the Hotfix as it is not automatically installed.
➀ New rumors suggest an RTX 5050 SKU is in the works for Nvidia's RTX 50 series; ➁ The budget card could offer a more affordable entry point for gamers; ➂ The RTX 5050 is expected to be priced between $199 and $249 and could feature 8 GB of memory.
➀ Nvidia is reportedly preparing to announce the RTX 5060 series in the next two weeks; ➁ The series will include three models: RTX 5060 Ti 16 GB, RTX 5060 Ti 8 GB, and RTX 5060 (8 GB); ➂ The availability dates have not been shared yet, but they are expected to be released in April.
➀ Micron and Astera Labs demonstrate the world's first PCIe 6.0 SSDs at DesignCon 2025;
➁ The SSD achieves read speeds of 27.14 GB/s with Astera's Scorpio PCIe 6.0 switch;
➂ PCIe 6.x promises to be the next horizon for enterprise and consumer products.
➀ Reputable leaker Billbil-Kun reports that Death Stranding 2 pre-orders are expected to start in March, with a release date likely announced at SXSW 2025.
➁ The game will have a Standard Edition at $69.99 and a Collector's Edition at $229.
➂ The leak also confirms that Death Stranding 2 will launch with a physical Standard Edition, priced at $69.99 USD / 69.99 EUR, along with a Collector's Edition planned for North America and Europe, retailing at $229 USD.
➀ NVIDIA可能会在下周或下下周初宣布GeForce RTX 5060 Ti和RTX 5060 GPU；
➁ RTX 5060系列预计将在3月中旬发布，具体日期为3月17日之后；
➂ RTX 5060系列包括基础RTX 5060（8GB）和两种RTX 5060 Ti版本（8GB和16GB），预计在发布后一个月左右上市。
➀ Nvidia's RTX 50 series ROP issue persists, affecting laptop manufacturers; ➁ These companies are implementing additional testing measures to ensure defect-free GPUs; ➂ The issue is believed to be from the same production line as affected desktop models, leading to extended working hours for laptop manufacturers.
➀ The HPE ProLiant DL145 Gen11 is reviewed, highlighting its features and performance as an AMD EPYC 8004 edge server;
➁ The server's design, I/O options, and expandability are discussed;
➂ The review also touches on the server's power efficiency and potential use cases.
➀ NVIDIA initially stated that a 'rare issue' with missing ROPs affected only 0.5% of GeForce RTX 5090/5090D and GeForce RTX 5070 Ti cards.
➁ Reports emerged that GeForce RTX 5080 cards could also be affected, leading to NVIDIA revising its statement.
➂ The latest episode reveals that notebook manufacturers are aware of missing ROPs on laptop variants of the GeForce RTX 50 series, with NVIDIA checking inventories for affected models.