By Zaheer Kachwala and Krystal Hu Jan 14 (Reuters) - OpenAI will purchase up to 750 megawatts of computing power over three years from chipmaker Cerebras as the ChatGPT maker looks to pull ahead in the AI ​race and meet the growing demand, the two companies said on Wednesday. This deal is worth more than $10 billion ‌over the life of the contract, according to a source familiar with the matter. The ChatGPT maker plans to use the systems built by Cerebras ‌to power its popular chatbot in what is the lates ...

Core Insights - OpenAI has entered into an agreement to purchase up to 750 megawatts of computing power from Cerebras over a three-year period [1] - The deal is valued at more than $10 billion, indicating a significant investment in computing resources by OpenAI [1] Company Summary - OpenAI is making a strategic move to enhance its computing capabilities through a substantial agreement with Cerebras [1] - Cerebras will provide the necessary computing power, which is crucial for OpenAI's operations and future projects [1]

The ChatGPT-maker is racing to secure more computing power, especially for responding to user queries. ...

Among its collection of 2026 predictions in the data centre space, GlobalData is anticipating that rack densities will continue to grow as AI workloads move into the enterprise space. Outside the enterprise, super dense systems mainly targeted at hyperscale customers and neoclouds will hit the market in 2026. AMD’s 72-GPU Helios rack, running on the Instinct MI400 architecture, and Nvidia’s NVL72 rack systems based on the Blackwell Ultra chips, will disrupt the landscape, introducing significant performa ...

Core Viewpoint - Nvidia has clarified that it has not acquired Groq but has obtained a non-exclusive license for Groq's intellectual property and hired key engineering talent from Groq to enhance its AI technology offerings [1][2]. Group 1: Nvidia's Engagement with Groq - Nvidia has entered into a non-exclusive licensing agreement with Groq regarding its inference technology, with Groq's founder and key team members joining Nvidia to advance the licensed technology [1][2]. - Groq will continue to operate independently, with Simon Edwards taking over as CEO, and its cloud services will remain unaffected by this collaboration [1][2]. - Nvidia's response counters previous reports claiming a $20 billion acquisition of Groq, emphasizing the focus on talent acquisition and technology licensing rather than outright purchase [1][2]. Group 2: Groq's Technology and Market Position - Groq, founded by former Google employee Jonathan Ross, specializes in AI chips for cloud computing, having developed the GroqChip capable of achieving 750 TOPS with 16 interconnected chips [2][3]. - The company has introduced the "Language Processing Unit" (LPU) concept, claiming its chips are ten times faster than Nvidia's H100 at a fraction of the cost, addressing the demand for real-time AI inference services [2][3]. - Groq's technology utilizes SRAM, which is significantly faster than the memory used in GPUs, allowing for quicker production and deployment of its chips [2][3]. Group 3: Market Dynamics and Competitive Landscape - Groq has rapidly gained attention in the AI chip market, achieving a valuation of $6.9 billion after multiple funding rounds, positioning itself as a strong competitor to Nvidia in the inference market [3][6]. - Nvidia maintains a leading position in the training segment of AI but faces increasing competition in inference from various companies, including Groq and Cerebras, which are exploring different architectures to capture market share [3][6]. - The market is witnessing a shift in focus from training to inference, providing opportunities for companies like Groq to capitalize on their technological advancements [3][6]. Group 4: Strategic Implications for Nvidia - By integrating Groq's technology and talent, Nvidia aims to strengthen its capabilities in the AI inference domain, potentially reducing reliance on external suppliers like TSMC for advanced packaging and memory [6][7]. - The acquisition of Groq's engineering team is seen as a strategic move to enhance Nvidia's existing ecosystem and address gaps in real-time inference capabilities [7]. - This transaction reflects a growing trend in Silicon Valley towards "acqui-hire deals," where companies acquire startups primarily for their talent rather than their products [8].

Core Viewpoint - Nvidia announced a record $20 billion cash deal with AI chip startup Groq, which initially created a stir in Silicon Valley, marking Nvidia's largest transaction to date, surpassing the $7 billion acquisition of Mellanox in 2019 [2][3]. Group 1: Transaction Details - Shortly after the announcement, both Nvidia and Groq clarified that the deal was not an acquisition but a non-exclusive technology licensing agreement [6][7]. - The transaction involves technology licensing and the integration of Groq's core team into Nvidia, while Groq will continue to operate independently [10][13]. - Groq's founders and key executives will join Nvidia to advance the application of the licensed technology [13]. Group 2: Strategic Intent - Nvidia's CEO Jensen Huang indicated that the strategic intent behind the deal is to integrate Groq's low-latency processors into Nvidia's AI infrastructure to support broader AI inference and real-time workloads [27]. - Groq specializes in high-performance AI accelerator chip design, particularly in the inference domain, with its language processing unit (LPU) claimed to be ten times faster than traditional solutions while consuming only one-tenth of the energy [28][30]. Group 3: Market Context - The deal reflects a trend in the tech industry known as "acqui-hire," where companies acquire talent and technology while avoiding traditional merger regulatory hurdles [16]. - Other tech giants have engaged in similar transactions, indicating a competitive landscape where major players are absorbing potential disruptors [17][45]. - Nvidia's cash reserves have significantly increased, providing ample resources for further acquisitions and investments in AI and energy infrastructure [34][35]. Group 4: Industry Developments - Other AI chip startups, such as SambaNova and Cerebras, are also in the spotlight, with potential acquisition interests from major companies like Intel [38][41]. - The competitive environment is tightening as large firms leverage capital to absorb emerging players, potentially narrowing the window for other competitors [45].

Core Viewpoint - Nvidia announced a record $20 billion cash deal with AI chip startup Groq, marking its largest transaction ever, surpassing the $7 billion acquisition of Mellanox in 2019 [2][4] Group 1: Transaction Details - Shortly after the announcement, both Nvidia and Groq clarified that the deal is not an acquisition but a non-exclusive technology licensing agreement [4][5] - The agreement allows Nvidia to integrate Groq's products into its future offerings while Groq continues to operate independently [7][8] - Groq's core team, including CEO Jonathan Ross and President Sunny Madra, will join Nvidia to advance the application of the licensed technology [7][12] Group 2: Strategic Intent - Nvidia's CEO Jensen Huang indicated that the strategic intent behind the deal is to integrate Groq's low-latency processors into Nvidia's AI infrastructure, targeting the AI inference market [12][14] - Groq specializes in high-performance AI accelerator chips, with its technology reportedly running large models ten times faster than traditional solutions while consuming only one-tenth of the energy [12][14] Group 3: Market Context - The transaction reflects a trend in Silicon Valley known as "acqui-hire," where companies acquire talent and technology without triggering antitrust issues [6][8] - Other tech giants have engaged in similar transactions, indicating a competitive landscape where major players are absorbing potential disruptors [8][22] - Nvidia's financial position is strong, with $60.6 billion in cash and short-term investments, enabling it to pursue significant acquisitions and investments in the AI sector [16]

英伟达官宣：以200亿美元现金与AI芯片初创公司Groq达成交易。 消息一出迅速引发市场轰动，因为这是英伟达有史以来最大规模的一笔交易， 远超2019年收购Mellanox的70亿美元 。 但仅仅几小时后，画风突变。 英伟达和Groq双双发表声明，对交易性质进行了澄清，并非收购。Groq在官方博客中写道： 梦晨 发自 凹非寺 量子位 | 公众号 QbitAI 平安夜老黄没有休息，一项 200亿美元 创纪录芯片收购消息，轰动硅谷。 我们与英伟达签订了一份非排他性技术许可协议。 英伟达方面也明确表态： 我们不是在收购Groq这家公司，我们只是获得技术授权，并将Groq的产品整合到未来的产品中。 看起来老黄也学会了"人才收购"这招：重金掏空一家公司的人才和核心资产，但又避免触发反垄断。 所以这200亿美元到底买了什么？ 席卷硅谷的"人才收购" 答案是：技术授权，加上一整支核心团队。 最先爆料的是Groq主要投资方Disruptive，其CEO透露英伟达已同意以200亿美元现金收购Groq的资产，交易进展非常迅速。 就在9月，Groq刚刚完成了一轮7.5亿美元的融资，估值达到约69亿美元。 自2016年Groq成立以 ...

来源：环球市场播报 欧洲股市再创新高 政策乐观情绪延续 欲览更多环球财经资讯，请移步7×24小时实时财经新闻 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 12月20日美股成交额前20：FTC批准英伟达投资英特尔 12月20日热门中概股多数上涨 小鹏汽车涨6.80%，满帮跌1.99% 特朗普称不排除与委内瑞拉开战后，周五油价上涨 现货黄金涨0.13%，报4338.32美元/盎司 宏观 2025年美国人求职不易 感觉就像是衰退 明年也将同样艰难 九家药企与特朗普达成协议 更多药企将跟进 特朗普称将召集保险公司开会讨论降价问题 特朗普五天圣诞长假期不会惠及大多数美国私企打工人 特朗普给联邦政府加假两天 但美国国债按原计划标售 美国防部连续第八年未通过财务审计 美制裁马杜罗家族成员及相关人士 加大对委政权施压 美国二手房销量小幅上升 房价涨幅放缓提供助力 谈判久拖不决 美国贸易代表批评欧盟和印度 公司 马斯克上诉获胜，恢复2018年560亿美元的特斯拉薪酬协议 存在溜车风险 福特在美召回超27万辆电动和混动汽车 受韦德布什下调评级并警示自动驾驶出租车冲击风险影响，Lyft股价下跌 星巴克聘请亚 ...

Market - US stock market closed higher on December 20, with the Nasdaq rising over 300 points, driven by a rebound in the AI sector [2][6] - Popular Chinese stocks mostly increased, with Xpeng Motors up 6.80% and Manbang down 1.99% [2][6] - Gold prices rose by 0.13%, reaching $4,338.32 per ounce [2][6] - European stock markets hit new highs, reflecting continued optimistic policy sentiment [2][6] Company - Ford recalled over 270,000 electric and hybrid vehicles in the US due to risks of rolling away [7] - Lyft's stock price fell following a downgrade by Wedbush, which warned of risks from autonomous taxi services [7] - Starbucks appointed Anand Varadarajan, a former Amazon executive, as the new Chief Technology Officer [7] - Google filed a lawsuit against a data scraping company, alleging it used false searches to steal online content [7] - Cerebras, an AI chip company, plans to apply for a US IPO after delays [7] - Rivian's stock price surged, reaching its highest level since December 2023 [7] - Biotech company Horizon Therapeutics plans to acquire Amicus Therapeutics for $4.8 billion, focusing on rare diseases [7]