一、市场: ●12月20日收盘：美股收高纳指涨逾300点 AI板块重拾升势 北京时间12月20日凌晨，美股周五收高，科技股领涨，纳指涨超300点。道指、纳指、标普500指数均收涨，本周三大股指涨跌不一。软件巨头甲骨文因 TikTok美国业务交易股价大涨，此前曾因债务和AI支出问题股价承压。英伟达收高，美光科技延续涨势。不过，市场或因"四重巫师日"出现剧烈波动，本周 五有超7.1万亿美元名义价值期权到期。此外，耐克因大中华区营收下滑及关税压力，股价收跌10.55%。有策略师认为AI相关企业发债潮或影响市场至2026 年，但企业有融资能力。 ●12月20日美股成交额前20：FTC批准英伟达投资英特尔 周五美股成交额前20的股票中，英伟达成交额居首，收高3.93%，成交580.42亿美元，FTC批准其对英特尔投资50亿美元，该交易或影响竞争格局。券商伯 恩斯坦称英伟达估值具吸引力，给予"跑赢大盘"评级。博通收高3.18%，特鲁斯特证券上调其目标价。特斯拉收跌，美最高法院恢复马斯克薪酬方案。谷歌 A类股收高，与派拓网络拓展合作。此外，美光、甲骨文、AMD等股票也有不同表现，如美光创历史新高，AMD苏姿丰访华交流。 ● ...

By Echo Wang NEW YORK, Dec 19 (Reuters) - AI chip maker Cerebras Systems is preparing to file for a U.S. initial public offering as soon as next ​week, targeting a second-quarter 2026 listing, people familiar with the matter said. The company, ‌which develops high-performance processors for artificial intelligence workloads, withdrew its prior IPO filing in October, days after announcing it ‌had raised over $1 billion in a fundraising round that valued the AI chipmaker at $8 billion. Cerebras declined ...

AI chip maker Cerebras Systems is preparing to file for a U.S. initial public offering as soon as next week, targeting a second-quarter 2026 listing, people familiar with the matter said. ...

Core Insights - Cerebras Systems has signed a Memorandum of Understanding (MOU) with the U.S. Department of Energy (DOE) to collaborate on next-generation AI and high-performance computing (HPC) technologies aimed at enhancing scientific research and national security [1][2][3] Collaboration Framework - The MOU establishes a framework for information sharing, joint research and development, and future agreements to accelerate the development of secure, scalable, and energy-efficient AI infrastructure [3][4] - Strategic collaborations will focus on developing world-leading Cerebras wafer-scale AI supercomputers and advanced AI models, as well as converged AI+HPC workflows [4][5] Historical Partnership - Cerebras has a longstanding partnership with the DOE, having worked together for over a decade on AI, HPC, and converged AI+HPC research [6] - DOE laboratories were early adopters of the Cerebras Wafer-Scale Engine, which is recognized as the world's largest and most powerful commercially available AI processor [6] Technological Advancements - The collaboration has led to significant technological advancements, including the development of large-scale data sets for scientific research and new computing architectures [7] - Breakthrough AI models have been developed for various scientific domains, including genomics and clean energy, with notable recognition such as the Gordon Bell Prize [8][11] Performance Capabilities - Cerebras systems demonstrate faster-than-exascale performance on key DOE mission workloads, enabling scientific modeling at unprecedented speeds compared to traditional processors [11] - Revolutionary hardware and memory co-design initiatives are underway, which could potentially increase wafer-scale system capacity for scientific simulations and large-scale AI workloads by 100 times [11]

SUNNYVALE, Calif.--(BUSINESS WIRE)--Cerebras Systems, in partnership with G42's Inception and MBZUAI's IFM, today announced the release of Jais 2, the leading open-source Arabic LLM – the first frontier language model both trained and deployed for inference on Cerebras Systems. The organizations combined their expertise with leading machine learning techniques, uniquely enabled on Cerebras wafer-scale clusters, to achieve state-of-the-art quality on Jais 2, using only a fraction of compute used. ...

Core Insights - Cerebras Systems has been awarded Demo of the Year for its AI Inference technology at the 2025 TSMC North America Technology Symposium, highlighting its significant innovation in the AI infrastructure space [1][3]. Group 1: Technological Achievements - Cerebras has developed a wafer-scale processor, the CS-3, which is 50 times larger than conventional processors, enabling AI workloads to run over 20 times faster than GPUs [2][8]. - The company’s flagship technology, the Wafer Scale Engine 3 (WSE-3), is the largest and fastest AI processor, outperforming the largest GPU by 56 times while consuming less power per compute unit [8]. Group 2: Market Adoption and Partnerships - Cerebras AI Inference is utilized in demanding environments globally, available through major cloud platforms such as AWS, IBM, and Hugging Face, and is adopted by sectors including healthcare, biotech, finance, and design [4][6]. - The technology supports critical national scientific research at U.S. Department of Energy laboratories and the Department of Defense, showcasing its versatility and reliability in high-stakes applications [4]. Group 3: Performance Metrics - Cerebras is recognized as the fastest platform for AI coding, generating code over 20 times faster than competing solutions, and consistently achieving the fastest inference speeds verified by independent benchmarks [5][8]. - The company serves trillions of tokens monthly across its cloud and on-premises deployments, indicating robust demand and operational scale [6].

Core Insights - Vulcan Elements, a rare earth startup supported by Donald Trump Jr.'s venture capital firm, received a significant loan commitment of up to $620 million from the U.S. Department of Defense to expand industrial magnet supply [1][4] - This loan represents the largest ever issued by the Pentagon's Strategic Capital Office and is part of a broader financing agreement valued at $1.4 billion, which includes over $550 million in private financing and other federal incentives [1][4] - The transaction has raised concerns about potential conflicts of interest, given the connections to Trump Jr. and the timing of the loan [1][2] Government Contracts and Investment Landscape - Since Trump Jr. joined 1789 Capital in 2024, the firm has invested in at least four companies that collectively secured over $735 million in government contracts this year [2] - 1789 Capital, founded by Trump-supporting donors, has a portfolio that has expanded from initially focusing on anti-"woke culture" businesses to include industries likely to benefit from Trump administration regulations [2] - Other companies in the 1789 Capital portfolio that received government contracts include Firehawk Aerospace, PsiQuantum, and Cerebras Systems, with contracts exceeding $10 million each [2] Denials of Political Influence - All parties involved have denied any political interference in the loan process, with Vulcan Elements' CEO stating no contact with Trump Jr. and emphasizing the value-based nature of the loan [3] - A Pentagon spokesperson confirmed that neither Trump Jr. nor 1789 Capital participated in discussions regarding the loan commitment [3] - Despite these denials, Trump Jr. previously indicated his deep involvement in strategic decisions at 1789 Capital [3] Financing Structure Analysis - The $620 million financing for Vulcan Elements is strategically significant, aimed at supporting the production of critical materials for defense equipment such as drones and nuclear submarines [4] - As part of the deal, the U.S. Department of Commerce will acquire a $50 million equity stake in Vulcan, and the Pentagon will receive warrants for Vulcan Elements and its partner ReElement Technologies [5] - Vulcan expects to increase its workforce from approximately 30 to 50 employees by the end of the year [5]

Core Insights - Cerebras Systems has launched "Cerebras for Nations," a global initiative aimed at assisting governments in developing and scaling their sovereign AI projects [1] Group 1: Initiative Overview - The "Cerebras for Nations" program will collaborate with international governments and their private sector partners in data centers, cloud, and AI ecosystems [1] - The initiative focuses on three key pillars of sovereign AI, emphasizing co-design and building world-class infrastructure [1]

Cerebras Systems is teaming up with the government of Guyana to build the area’s first data center https://t.co/j0sXAXqW0G ...

Core Insights - DeepSeek has released a new multimodal model called DeepSeek-OCR, which has sparked significant discussion in the industry regarding its potential applications in optical and quantum computing [1] - The model's visual encoder enables efficient decoding, providing a clear technical pathway for integrating optical computing into large language models (LLMs) [1] Group 1: Contextual Optical Compression - DeepSeek has introduced "Contextual Optical Compression" technology, which processes text as images to achieve efficient information compression, theoretically allowing for infinite context [3] - This technology can compress tokens by 7 to 20 times; for instance, converting a page of text that typically requires 2000-5000 tokens down to just 200-400 visual tokens [3][4] - The model maintains 97% decoding accuracy at 20x compression, with 60% accuracy still achievable at 20x compression, which is crucial for implementing LLM memory's forgetting mechanism [4] Group 2: Optical Computing Integration - By transforming text problems into image problems, DeepSeek's OCR technology may pave the way for the integration of optical computing chips into large language models [5] - Optical computing chips are seen as a potential technology for the "post-Moore era," leveraging light-speed transmission, high parallelism, and low power consumption for AI and other computation-intensive tasks [5] - The DeepEncoder component of DeepSeek-OCR is particularly suited for execution by optical co-processors, while the text decoding will still be handled by electronic chips [5] Group 3: Challenges and Industry Landscape - Current challenges for optical computing include advanced optoelectronic integration and the maturity of the software ecosystem, which hinder large-scale development and optimization [6] - Key players in the domestic market include companies like Xizhi Technology and Turing Quantum, while international competitors include Lightmatter and Cerebras Systems [6][7] - Turing Quantum has made significant progress in the mass production of thin-film lithium niobate (TFLN) products, but it may take 3 to 5 years to compete with GPUs in data centers due to engineering, cost, and ecosystem challenges [7]