This article first appeared on GuruFocus. Tesla (NASDAQ:TSLA) Chief Financial Officer Vaibhav Taneja sold 3,000 shares of the electric-vehicle maker's stock in a recent transaction that was described as tax-related, while retaining about 130,000 shares including indirect holdings. The sale brought in about $1.3 million, based on a share price near $450. Tesla shares had been trading above that level in recent weeks before slipping 4.7% on May 15 during a broader market selloff. Tesla shares have also been volatile this year as investors have tracked the company's efforts to win approval to sell its Full Self Driving feature in China. The stock had climbed on that expectation before pulling back. Tesla's FSD system can handle most driving tasks most of the time, and U.S. customers currently pay $99 a month for the service. Tesla says it has 1.3 million subscribers.

German economy minister's China trip to include CEOs of BASF, Thyssenkrupp Reuters

(RTTNews) - Osisko Development Corp. (ODV), on Thursday announced the pricing of a $275 million aggregate principal amount offering of 4.125% convertible senior notes due 2031 in a private placement. The notes were priced with a 25% conversion premium. The offering includes $225 million of notes to qualified institutional buyers. Double Zero Capital, LP, agreed to purchase $50 million aggregate principal amount of affiliate notes. The company said that the Initial purchasers were also granted an option to purchase up to an additional $25 million aggregate principal amount of notes. The company expects net proceeds from the offering of about $215.9 million, or about $240 million if the option is exercised in full, after discounts, commissions, and expenses. The company expects net proceeds of about $50 million from the private placement. The proceeds will be used to fund development of the Cariboo Gold Project, pay capped call transaction costs, and for general corporate purposes. The company said the capped call transactions carry an initial cap price of $5.88 per share, representing a 100% premium to the last reported New York Stock Exchange share price of $2.94 on May 20. The notes will bear interest at 4.125% annually and mature on June 15, 2031, unless earlier converted, redeemed, or repurchased. In the pre-market trading, Osisko Development is 15.56% lesser at $2.4700 on the New York Stock Exchange. The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.

格隆汇5月21日｜白宫官员：特朗普将推迟实施拜登时期出台的制冷剂管控规则。

据两名知情业内高管透露，为应对旗下人工智能产品日益增长的市场需求、扩充算力资源，Anthropic公司目前正洽谈租用搭载 微软 自研 AI 服务器芯片的算力服务器。 若成功拿下Anthropic这一客户，对微软而言将是重大利好。此前微软自研芯片项目在去年遭遇研发延期。英伟达 AI 芯片几乎占据了微软面向 AI 开发者出租的全部算力设施，如今微软意图效仿 谷歌 、 亚马逊 等云服务竞品，打造可替代英伟达芯片的自研芯片产品。 各大云服务商高管均表示，英伟达硬件采购成本居高不下，研发定制化自研芯片，是稳住并提升企业利润空间的必要举措。 微软迈亚芯片能为Anthropic运行克劳德大模型提供全新算力选择，Anthropic还有望推动该芯片新一代产品按自身业务需求定制优化。迈亚芯片主打提速运行现有大模型，性能优于英伟达同类产品，但并不像英伟达芯片那样，适配客户全新大模型的研发搭建工作。 微软今年 1 月宣布，已在旗下 Azure 云数据中心正式投产最新款迈亚 200 芯片。Azure 业务负责人斯科特・格思里表示，借助迈亚 200 芯片，微软有效降低了旗下智能助手 Copilot 的运行成本，该助手依托开放人工智能研究中心与Anthropic的大模型提供服务。不过相较于谷歌、亚马逊的自研 AI 芯片，迈亚芯片目前仍处于发展初期。 知情人士称，双方此次芯片算力合作洽谈尚处于初步阶段，最终未必能达成正式合作协议。 长期以来，Anthropic始终与其他 AI 企业走差异化路线：通过采购多品牌服务器芯片开展技术研发与业务运营，摆脱了开放人工智能研究中心、xAI 等企业高度依赖英伟达硬件的困境。该企业目前同时使用亚马逊、谷歌及英伟达旗下 AI 服务器芯片，还在接洽一家英国新兴初创企业的芯片产品。 Anthropic本季度营收接近 110 亿美元，远超今年一季度营收规模，预计单季将斩获 5.6 亿美元净利润，业绩表现亮眼。与此同时，该企业已敲定向美国三大头部云服务商合计投入至少 3300 亿美元算力采购资金。从理论层面来看，选用三大云服务商定制化 AI 芯片，相比普通芯片性价比更高，若云服务商愿意承担部分成本补贴，效益还将进一步提升。 脸书母公司元宇宙平台、开放人工智能研究中心等一众 AI 同行，如今纷纷效仿Anthropic的多芯片布局策略，接连与多家算力供应商达成合作，同时还各自自...

格隆汇5月21日｜据澎湃，5月20日，在快速且几乎全面完成对“自主业务”分拆后，巴斯夫宣布“制胜之道”战略的下一阶段将重点强化并发展其“核心业务”。为此，巴斯夫新设立“核心业务”转型团队（Core Transformation Office），并由Julia Raquet出任总裁，负责推动跨部门转型项目的落地。项目覆盖所有“核心业务”、服务部门及企业中心。该项目命名为“CoreShift”，旨在以2024年为基准，到2029年实现核心业务的净现金固定成本降低至多20%。Julia Raquet在新岗位上将直接向巴斯夫集团执行董事会主席凯礼博士汇报。

This article first appeared on GuruFocus. Nvidia Corporation (NASDAQ:NVDA) said it approved an additional $80 billion share repurchase authorization as the chipmaker continued to generate strong cash flow from artificial-intelligence demand. Nvidia said the new authorization comes on top of roughly $39 billion remaining under its existing buyback program. The company also raised its quarterly dividend to $0.25 per share from $0.01. Nvidia generated about $48.55 billion in free cash flow during the fiscal first quarter ended April 26. The company said revenue rose to $81.61 billion, supported by continued strength in its data center business tied to AI infrastructure spending. Nvidia projected second-quarter revenue of about $91 billion, plus or minus 2%, ahead of analyst expectations.

Russia announced plans to sell government bonds denominated in Chinese yuan for the second time after President Vladimir Putin returned from a two-day visit to China. Investors will be offered fixed-coupon bonds with a maturity of 10 years and a face value of 10,000 yuan per note, the Finance Ministry said in a statement Thursday. The ministry will collect orders on May 28, with placement on the Moscow Exchange scheduled for June 3. Bond purchases and coupon payments will be available in either Chinese yuan or Russian rubles at the investor’s choice, according to a statement. The plan to raise yuan for the state coffers follows talks in Beijing between Putin and Chinese President Xi Jinping , as the two leaders sought to underscore the strength of their ties amid global tensions over the wars in Ukraine and Iran. Read more: Xi and Putin Wrap Talks Showcasing Ties Amid Global Unrest Russia first sold local bonds in yuan late last year , raising a total of 20 billion yuan ($2.9 billion) in two tranches as the government sought funding to cover budget shortfalls driven by military spending and weaker oil revenues. The 12 billion-yuan bonds due February 2029 carry a 6% coupon, while an 8 billion-yuan tranche due June 2033 pays 7%. Finance Minister Anton Siluanov said at the time that the ministry aimed to build a benchmark yield curve in the Chinese currency to help corporate borrowers tap yuan funding.

This article first appeared on GuruFocus. SpaceX filed its S-1 with the U.S. Securities and Exchange Commission on Wednesday for one of the biggest initial public offerings in history. The company will trade on Nasdaq under the ticker SPCX. Class A shares carry one vote each and Class B shares carry 10. No price range or share count was disclosed in the preliminary prospectus. The filing reveals a company that generated $18.7 billion in revenue in 2025 and lost $4.9 billion, driven by $20.7 billion in capital expenditures. Starlink accounts for most of the economics. The segment reported $11.4 billion in revenue, $4.4 billion in operating income, and 10.3 million subscribers across 164 countries. SpaceX's February 2026 acquisition of xAI brought Grok and X into the consolidated entity going public. The S-1 also discloses that Anthropic signed a compute services agreement in May 2026 to pay $1.25 billion per month through May 2029 for access to SpaceX's COLOSSUS data centers. The company also disclosed ownership of 18,712 Bitcoin, acquired at approximately $35,000 per coin and now valued at $1.6 billion at year-end 2025. Elon Musk holds approximately 6.4 billion shares, including 5.6 billion Class B shares that give him 85.1% of combined voting power. The Tesla (NASDAQ:TSLA) CEO could become the world's first trillionaire if the company prices at or above the expected valuation.

Kirk Fisher/iStock Editorial via Getty Images The U.S. is filled with aging housing stock. According to the American Planning Association , the median age of owner-occupied houses is about 40 years. As I’ve written about previously , the fundamentals of old housing stocks should be a boom for the home improvement retailers, given the pure necessity of continuing repairs and maintenance (“R&M”). Results from Lowe’s Companies ( LOW ) show that this isn’t exactly the case. In the current environment, it appears the average DIYer is putting away their toolbox, at least for the time being. LOW’s Q1 came in ahead of expectations on both the top and bottom lines. Forward guidance, however, was light, albeit reaffirmed. The guidance didn’t surprise me. In my last piece on LOW, I wrote that I believed the market would remain soft for the entirety of the year and that LOW was fairly valued for the environment. Shares in LOW have declined by about 14% since that update . This compares to a 8.8% gain in the S&P 500 ( SPY ) over the same period. Seeking Alpha - LOW's Share Performance Since Last Author Update While I do see shares as more attractive today than in my last update earlier in the year, I still believe shares in LOW are a ‘hold’ in today’s operating environment, which likely will remain soft for an extended period of time. LOW Q1 Results Recap In reviewing Q1, I would note that LOW reported a better-than-expected quarter, though the report still reflected the difficult environment facing the home improvement space. At the topline, revenues increased 10.3%, while comparable sales rose 0.6%. This marked the company’s fourth consecutive quarter of positive comps. Adjusted EPS, meanwhile, came in at $3.03/share, ahead of expectations and up 3.8% from the prior year period. LOW Q1 Earnings Presentation - Snapshot Of Operating Results The quarter was once again driven by the Pro business, where the company continued to see continuing demand from its core small- and medium-...

Japan-listed Kura Sushi stock gained as much as 5.4% after disclosures showed Trump bought between $1 million and $5 million in Kura Sushi USA shares

European stock markets could rise another 5% by the end of the year, driven by “an explosion of earnings,” said Beata Manthey , head of European equity strategy at Citigroup Global Markets Ltd. “If you look at what is happening on the earnings front and the parts of the market that are actually driving the index level, they are doing pretty well,” Manthey said in an interview with Bloomberg Television. “We’ve been seeing an explosion of earnings in the last reporting season, so from that perspective I don’t see risks.” Stocks can move higher even in the face of increased bond yields, she said, as long as the pace of inflation doesn’t pick and economic growth, as measured by purchasing manager indexes, remains solid. “If PMIs start rolling over that could trigger, or historically have triggered, a selloff in equity markets,” she said. Europe faces greater impact from higher oil prices than other regions, and Citigroup last month shifted its view to favor stocks in the US versus the rest of the world. “There are no miracles,” she said. “Economies and parts of equity markets exposed to higher energy prices will see downgrades, and we are already starting to see them.” Yet Citi hasn’t changed its target for European stocks nor its earnings growth forecast, because an “explosion of earnings in energy” is offsetting downgrades elsewhere, she said. The market rally has occurred “in this very narrow fashion,” Manthey said. Still, Manthey warned of increasing frothiness around artificial intelligence. A Middle East ceasefire would trigger a squeeze in European equities as underweight investors rush in, though a Ukraine resolution “will potentially be even bigger than the Middle East,” she said. Citigroup recommends technology, materials and financials via banks, while maintaining caution on sectors exposed to higher inflation. This story was produced with the assistance of Bloomberg Automation.

(RTTNews) - While reporting financial results for the fourth quarter on Wednesday, luxury home furnishing retailer Williams-Sonoma, Inc. (WSM) reiterated its revenue growth guidance for the full-year 2026. For fiscal 2026, the company continues to expect net revenue growth of 2.7 to 6.7 percent, with comparable brand revenue growth of 2 to 6 percent. Over the long term, the company continues to expect mid-to-high single-digit annual net revenue growth with an operating margin in the mid-to-high teens. In Thursday's pre-market trading, WSM is trading on the NYSE at $186.00, up $5.76 or 3.24 percent. For more earnings news, earnings calendar, and earnings for stocks, visit rttnews.com The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.

Home Depot (NYSE:HD) sits at an inflection point. The stock has slid from a September 2025 peak near $422.71 to $302.44, hugging the lower end of its 52-week range as housing affordability and tepid consumer sentiment weigh on big-ticket projects. Our 24/7 Wall St. price target for Home Depot is $367.94 over the next 12 ... Where Will Home Depot’s Stock Price Be In 2027?

Flipper Devices, maker of the Flipper Zero hacking device, today announced a new gadget called Flipper One that has multiple network connectivity chops and can act as a Linux PC (is this the year?). The company has sold over a million Flipper Zero units and has generated over $150 million in sales. However, the new device is not a successor as it operates on a different layer than the Flipper Zero, the company said. The Flipper Zero device is popular in the hacker community, which can connect to radios like Bluetooth, RFID, NFC, a sub-1GHz transceiver, and Infrared. The device could act like a key fob or an entry pass, but could also be used in cases like spamming nearby iPhones. Meanwhile, the new Flipper One device relies on network connectivity through 2x Gigabit Ethernet, USB Ethernet (5 Gbps), and Wi-Fi 6E (2.4/5/6 GHz). What’s more, the device has an M.2 port, which can be used to connect a modem for 5G connectivity or other devices like SDR modules, AI accelerators, SSDs (NVMe or SATA), and Wi-Fi cards via adapters. The device is still in development, and the company is merely announcing the project at the moment. The device will run two processors along with an 8GB RAM. The first is an eight-core RK3576 chip that runs open Linux along with a Mali-G52 GPU and an NPU to run local AI models. Image Credits: Flipper Devices Image Credits:Flipper Devices The company said that it worked with open-source software consulting firm Collabora to push this chip’s support into the mainline Linux Kernel, so anyone can download it from Kernel.org and tinker with it. The second chip is a two-core Raspberry Pi RP2350 microcontroller. This powers display, buttons, and touchpad, LEDs, and the power subsystem, so even when the Linux part of the device is turned off, users can operate the device. Flipper Devices CEO Pavel Zhovner said that the company is also developing its own Linux-based flavour. He said in a blog that while Raspberry Pi OS is fluid and he enjoys using it, it i...

Under Secretary of War for Research and Engineering Emil Michael reportedly said that conversations with Anthropic ( ANTHRO ) have stopped, while many other companies are very willing to work with the Pentagon. Micheal cited recent deals signed with Microsoft, Nvidia, OpenAI, and SpaceX while speaking on Bloomberg TV. The under-secretary also stated that he puts Mythos in the same category as other cyber models. Michael added, “You’ll see us, as the U.S. government, looking at all these models and working with them. We have to, as a government, treat all these new capabilities in a different way than we’ve treated anything like this before to make sure our networks are patched.” Anthropic did not immediately respond to Seeking Alpha’s request for comment. In other AI news of note, U.S. President Donald Trump is expected to sign an executive order on AI and cybersecurity as soon as Thursday, as pressure grows from parts of his political base to increase oversight of new AI models, such as Anthropic’s Mythos, Reuters reported, citing sources. The White House was working to bring AI company CEOs to a signing ceremony with President Donald Trump, the report said. Earlier this month, the Department of War, also known as the Department of Defense, inked agreements with ChatGPT maker OpenAI ( OPENAI ); Alphabet’s ( GOOG ) ( GOOGL ) Google; Elon Musk’s SpaceX (SPACE), which owns xAI ( X.AI ), the maker of AI chatbot Grok; Microsoft ( MSFT ); Amazon’s ( AMZN ) Amazon Web Services; Nvidia ( NVDA ); and startup Reflection AI to deploy their advanced AI capabilities on the department’s classified networks for lawful operational use. The DoD said that SpaceX, OpenAI, Google, Nvidia, Reflection AI, Microsoft, and AWS will provide resources to deploy their capabilities for both Impact Level 6 (IL6) and Impact Level 7 (IL7) environments. More on Anthropic Anthropic Is Taking Over Enterprise Wall Street Lunch: Anthropic Tries To Contain Claude Code Instruction Leak Anthropic's IPO: ...

Getty Images Introduction Per my February article , Nvidia Corporation ( NVDA ) has four key advantages which are relatively low stock-based compensation (“SBC”), light capex, high margins, and extraordinary growth. We now have the fiscal Q1 '27 numbers and immense amounts of new information from Nvidia’s March 2026 GPU Technology Conference (“GTC”). One factor driving these economic advantages is their internal use of AI. My thesis is that Nvidia uses AI internally to make the best AI hardware. Their approach compounds into a structural advantage in cost, speed, and quality. This is not just about better chips, it is about compressing the design loop. There are other companies who use AI internally for chip design but it is beyond the scope of this article to go into details on them. One company in this category is Ricursive Intelligence and they have partial backing from Nvidia along with some former employees. Nvidia employees use design tools from companies who have AI built into the products. The 3 biggest companies in this area are Synopsys ( SNPS ), Cadence ( CDNS ) and Siemens EDA ( SIEGY ). Another company using AI internally for chip design is Alphabet/Google ( GOOG , GOOGL ) and one of the highlights of the March 2026 GTC was the Advancing to AI’s Next Frontier session given by Nvidia Chief Scientist Bill Dally and Google Chief Scientist Jeff Dean. I wrote about Google’s progress in this area on May 5. Nvidia Design Automation Research Director Haoxing (“Mark”) Ren gave a 2022 presentation on ML for chip design methodology and electronic design automation (“EDA”) optimization. Chip designs are described in a hardware description language like Verilog at an abstraction level called the register-transfer level ("RTL") which is basically the source code of a chip. For the digital flow of chip design, we have architecture, RTL, and physical design. Within physical design, we have synthesis, placement, clock tree synthesis, optimization, and routing. I like to...

Getty Images Introduction Per my February article , Nvidia Corporation ( NVDA ) has four key advantages which are relatively low stock-based compensation (“SBC”), light capex, high margins, and extraordinary growth. We now have the fiscal Q1 '27 numbers and immense amounts of new information from Nvidia’s March 2026 GPU Technology Conference (“GTC”). One factor driving these economic advantages is their internal use of AI. My thesis is that Nvidia uses AI internally to make the best AI hardware. Their approach compounds into a structural advantage in cost, speed, and quality. This is not just about better chips, it is about compressing the design loop. There are other companies who use AI internally for chip design but it is beyond the scope of this article to go into details on them. One company in this category is Ricursive Intelligence and they have partial backing from Nvidia along with some former employees. Nvidia employees use design tools from companies who have AI built into the products. The 3 biggest companies in this area are Synopsys ( SNPS ), Cadence ( CDNS ) and Siemens EDA ( SIEGY ). Another company using AI internally for chip design is Alphabet/Google ( GOOG , GOOGL ) and one of the highlights of the March 2026 GTC was the Advancing to AI’s Next Frontier session given by Nvidia Chief Scientist Bill Dally and Google Chief Scientist Jeff Dean. I wrote about Google’s progress in this area on May 5. Nvidia Design Automation Research Director Haoxing (“Mark”) Ren gave a 2022 presentation on ML for chip design methodology and electronic design automation (“EDA”) optimization. Chip designs are described in a hardware description language like Verilog at an abstraction level called the register-transfer level ("RTL") which is basically the source code of a chip. For the digital flow of chip design, we have architecture, RTL, and physical design. Within physical design, we have synthesis, placement, clock tree synthesis, optimization, and routing. I like to...

车企斥资数百亿美元扭转连年市场份额下滑颓势 周四，斯特兰蒂斯向投资者公布转型复兴方案。这家拥有多品牌矩阵的汽车集团表示，美国市场投资重心将聚焦吉普与公羊两大品牌，同时推出更多售价 4 万美元以内的车型。 吉普、公羊与欧洲主流品牌标致、菲亚特并列为集团四大核心品牌。首席执行官安东尼奥・菲洛萨推出总额 700 亿美元复兴计划，四大品牌将包揽公司 70% 的产品研发投入。 集团计划推出更多亲民价位新车，低价车型匮乏一直是困扰其美国经销商的突出问题。北美市场将推出 11 款全新车型，其中 7 款售价低于 4 万美元，2 款售价不足 3 万美元，以此抢夺更多市场份额。 经典品牌克莱斯勒目前仅在售一款厢式 MPV，后续将新增三款紧凑型 SUV，至少一款车型起售价低于 3 万美元。 历经数年经营低迷后，车企力图重回增长正轨。美国作为核心盈利市场，斯特兰蒂斯在此份额持续缩水，2025 年录得 260 亿美元净亏损。巨额亏损主要源于公司调整长期电动车投资战略产生的资产减值损失。 菲洛萨表示，这份规划凝聚了公司数月严谨筹备的成果，旨在推动企业实现长期盈利增长。 斯特兰蒂斯于 2021 年由 菲亚特克莱斯勒 与法国 PSA 集团合并组建，旗下汇聚吉普、公羊、克莱斯勒、道奇等美式品牌，以及标致、雪铁龙、菲亚特、欧宝、阿尔法・罗密欧等欧洲品牌。 此次合并并未达成投资者预期，未能打造出高效协同的跨国汽车巨头。分析师也诟病品牌架构冗余，核心市场销量萎缩的同时，各品牌还相互争夺资金与资源。 近一年前上任的菲洛萨，将以这份最新方案作为自身履职考核标尺。 前任首席执行官卡洛斯・塔瓦雷斯于 2024 年 12 月离任，其任职期间美国市场表现持续走弱，企业电动车投入规模庞大。疫情期间利润大涨后，公司主攻高定价、高利润的皮卡与 SUV 车型，最终导致经销商库存积压，平价车型供给短缺。 菲洛萨上任后调整产品策略，重启吉普切诺基常规款 SUV 车型，一改此前仅保留大切诺基大型 SUV 的规划；公羊皮卡也重新搭载经典大排量八缸 HEMI 发动机，推翻塔瓦雷斯时期停产该机型、改用六缸发动机的决策。 斯特兰蒂斯以往常年难以达标美国燃油经济性法规，如今特朗普政府大幅放宽相关管控标准，企业也顺势调整产品路线。 今年年初，公司停产克莱斯勒大捷龙 MPV、吉普牧马人 SUV 的插电混动版本。本周，公羊品牌正式推出全新燃油高性...

00:00 Speaker A First, we got to talk about Nvidia. 00:03 Speaker A And I'm pleased to say we are joined today by Nvidia CFO Colette Kress, joining us um on the phone. And uh Colette, thank you so much for joining us. Um the quarter was strong in a word. 00:20 Speaker A Um and in particular on the call, I think one of the things that investors and analysts were excited to hear was the breakout and more detail on the CPU data center business. Um and that's where I wanted to begin today. And and if you could talk talk us a little bit more through how Nvidia is thinking about that business and the size of that opportunity. 00:48 Colette Kress Thank you. Thank you so much. Yes, we had a great quarter, uh talking about a lot of different things. Uh important to see uh the growth that we've seen in the AI and the different types of AI. Uh we're in a phase right now of agentic AI, uh and that has really driven a significant amount of compute uh that we've been able to put in market. 01:14 Colette Kress The diversifying of our customers continues to grow and all great types of uh folks buying our compute. But most importantly now is a focus in terms of the CPU and how important this is to the workloads as well. 01:29 Colette Kress And we have the ability not only to just be inside uh Grace Blackwell as well as what we have in terms of our systems with Vera Ruben, we now have an opportunity to also have what we call a standalone uh CPU and selling that. We have customers already looking at that opportunity to provide standalone along with our Vera Rubin and our Grace Blackwell together. That together, we believe for this year, can be a total of about $20 billion worth of CPUs. 02:08 Speaker B So when when you talk about that 20 billion, is it is it the 20 billion as in the CPUs included with Vera Ruben and then the standalone units as well? Or is it the standalone uh on its own that's the 20 billion? How does that that kind of work out? 02:30 Colette Kress It's actually both...

NVIDIA Corp. NVDA — the undisputed global leader of generative artificial intelligence (AI)-powered graphical processing units (GPUs) — reported solid first-quarter fiscal 2027 earnings results. Adjusted earnings per share came in at $1.87, surpassing the Zacks Consensus Estimate of $1.77 and the year-ago figure of $0.81. Revenues came in at $81.62 billion, outpacing the Zacks Consensus Estimate by 3.63% and the year-ago figure of $44.06 billion. Adjusted gross margin remained strong at around 75%. The chart below shows the price performance of NVDA year to date. Zacks Investment Research Image Source: Zacks Investment Research Data Center Revenues Maintain Momentum Data center revenues reached $75.2 billion, surged 92% year over year. NVDA reported 50% of these revenues came from hyperscalers. This segment will continue its northward journey as four AI hyperscalers — Meta Platforms Inc. META, Alphabet Inc. GOOGL, Microsoft Corp. MSFT and Amazon.com Inc. AMZN — raised their AI capex budget to $725 billion in 2026 from $670 billion estimated earlier. The most important fact is that the remaining 50% of data center revenues came from various other sources like AI Clouds, industrial, enterprise, and sovereign customers. This indicates that NVDA is strategically diversifying its customer base as AI tools are gradually being adopted in mainstream business operations. Transformation From AI GPU to AI CPU Beside being the generative-AI-powered GPU leader, NVIDIA is aiming to become a leading supplier of AI CPUs (Central Processing Units). The so long dormant AI CPU space has charged up with the advent of agentic AI tools. In this regard, NVDA’s new AI superchip — Vera Rubin — may become a game-changer. This innovative rack-scale system will deliver 10 times more performance per watt than its predecessor, Grace Blackwell. The Vera Rubin system comprises 1.3 million components, including 72 Rubin GPUs and 36 Vera CPUs. NVIDIA will start the shipment of Vera Rubin in the seco...

Credo Technology Group Holding Ltd CRDO shares surged 8.3% in trading yesterday and closed at $182.98. The catalyst behind the rally was Credo’s collaboration with Rebellions. By integrating Credo’s ZeroFlap active electrical cables (AECs) into Rebellions’ RebelPOD AI infrastructure platform, the two companies aim to deliver scalable, turnkey AI factories optimized for enterprise inference workloads. In high-utilization clusters, brief instability can interrupt data flows, reduce GPU efficiency and force administrators into time-consuming troubleshooting cycles. Credo’s ZeroFlap AEC technology is designed specifically to eliminate these intermittent disruptions. AECs are engineered to provide stable high-speed connectivity, reduce signal integrity issues, enable plug-and-play deployment, lower maintenance complexity and improve operational uptime. By integrating these cables into the RebelPOD architecture, Rebellions can offer enterprises AI infrastructure with greater resilience and predictability. The collaboration strengthens Credo’s position within the AI inference market. Rebellions is developing AI accelerators focused on inference workloads rather than training. Inference, the process of running AI models in production environments, is expected to become one of the fastest-growing segments of the AI semiconductor market. By integrating Credo’s connectivity solutions into Rebellions’ production-ready AI infrastructure stack, the partnership creates a more comprehensive offering for enterprises seeking scalable AI deployments. This also expands Credo’s exposure beyond traditional hyperscale customers and into enterprise AI markets. The 8% stock gain suggests that investors believe the partnership could accelerate Credo’s revenue growth opportunities in AI infrastructure. CRDO has already benefited from surging demand for AI networking solutions, particularly among hyperscale cloud providers. Strategic alliances like this could diversify its customer base while ...

格隆汇5月21日｜WTI原油突破102美元/桶，日内涨3.86%。

格隆汇5月21日｜据国家数据局，2026年5月19日下午，国家数据局在广东省广州市白云国际会议中心组织召开数据基础设施建设现场会，总结数据基础设施建设先行先试工作进展，交流典型经验，部署下一阶段重点任务。会议强调，聚焦技术攻关，加快可信数据空间、隐私计算、区块链、数场等技术路线动态优化升级，积极探索数据编织、数据流式处理、本体论等新技术应用，加强数据领域核心技术攻关。

Joe Raedle/Getty Images News SpaceX’s ( SPCX ) blockbuster IPO is expected to become one of the largest public offerings in history, but analysts say the deal may reveal little about the overall health of the IPO market, Reuters reported Thursday. The company’s planned listing, which could raise more than $75 billion at a valuation near $1.75 trillion, has generated intense investor interest because of its scale, Elon Musk’s profile and the company’s combination of satellite internet, launch services and AI-related ambitions. Still, market observers told Reuters that SpaceX ( SPCX ) is too unusual to serve as a reliable indicator for other companies considering IPOs. Analysts said investor demand remains focused on a narrow group of sectors including AI infrastructure, defense, energy and biotechnology rather than the broader market. Some bankers also warned that the sheer size of the offering could absorb investor capital and temporarily sideline other IPO candidates, particularly larger deals that depend heavily on institutional U.S. investors. Several companies, including AI chipmaker Cerebras and advisory firm Lincoln International, have recently moved ahead with listings, with some issuers said to be accelerating timelines ahead of SpaceX’s expected debut. Retail enthusiasm is also expected to play a major role in the transaction. Analysts interviewed by Reuters said many individual investors are being drawn more by Musk’s reputation and the rarity of the opportunity than by traditional valuation metrics. Reuters previously reported that SpaceX has discussed allocating as much as 30% of shares to retail investors, far above the norm for major IPOs. The company’s filing showed that most of its $18.67 billion in 2025 revenue came from its Starlink satellite internet business, while its AI operations remain unprofitable. Analysts said the broader IPO market recovery will likely depend more on economic stability, reduced geopolitical tensions and calmer equity mark...

騰訊控股(00700.HK)5月21日回購5.01億港元,年內累計回購121.60億港元 富途牛牛

bymuratdeniz Crypto services provider Blockchain.com Group Holdings said it has confidentially submitted a draft registration statement on Form S-1 with the U.S. Securities and Exchange Commission. The company has proposed an IPO of class A shares, with the number of shares to be offered and price range yet to be determined. One of the oldest crypto exchanges, Blockchain.com began operations in 2011. It has over 95M wallets, more than 43M verified users, and has facilitated over $1.1 trillion in crypto transactions. More on IPOs OpenAI could pull off biggest IPO ever, Deutsche Bank says Cathie Wood loads up on Cerebras after blockbuster IPO debut Pershing Square USA opens ~18% below IPO price in NYSE debut

This article first appeared on GuruFocus. Advanced Micro Devices (AMD, Financials) said it will invest more than $10 billion across Taiwan's semiconductor ecosystem to support its next-generation artificial intelligence chips. The investment will focus on expanding strategic partnerships and scaling advanced packaging manufacturing, a critical step in building high-performance AI processors. Taiwan remains one of the most important centers in the global chip supply chain. For AMD, deeper ties there could help secure capacity as demand for AI infrastructure grows. The move also shows how chipmakers are investing beyond design and into the manufacturing networks needed to deliver more powerful processors at scale. AMD is competing with Nvidia and other semiconductor companies for a larger share of AI spending from cloud providers, enterprises and governments. Investors will watch for more details on timing, partners and whether the investment helps AMD convert AI demand into stronger revenue growth.

Advanced Micro Devices (AMD, Financials) said it will invest more than $10 billion across Taiwan's semiconductor ecosystem to support its next-generation artificial intelligence chips. The investment will focus on expanding strategic partnerships and scaling advanced packaging manufacturing, a critical step in building high-performance AI processors. Taiwan remains one of the most important centers in the global chip supply chain. For AMD, deeper ties there could help secure capacity as demand for AI infrastructure grows. The move also shows how chipmakers are investing beyond design and into the manufacturing networks needed to deliver more powerful processors at scale. AMD is competing with Nvidia and other semiconductor companies for a larger share of AI spending from cloud providers, enterprises and governments. Investors will watch for more details on timing, partners and whether the investment helps AMD convert AI demand into stronger revenue growth.

Welcome to Bloomberg’s Banking Monitor . Every Thursday we’ll deliver you the top news of the global banking industry with emerging trends, winners and losers and market opportunities. Sign up now if you’re not already on the list. The lull in releases of annoying new economic data and the more-or-less lull in Iran have left Wall Street free to get back to its day job of raising hundreds of billions of dollars for companies going public and buying each other. Goldman Sachs and Morgan Stanley won lead roles in the upcoming mega initial public offering for Elon Musk’s SpaceX. The duo are reportedly angling for the same role in an almost-imminent IPO filing for OpenAI , the creator of ChatGPT. Their counterparts on credit teams are lining up about $50 billion of loans and bonds to help clinch Paramount Skydance’s acquisition of Warner Bros. Discovery. It’s one of the biggest financing efforts in recent years and it comes with another $6.2 billion for Warner Bros. to clean up some current debt. The credit spigots are so wide open that a unit of Blue Owl Capital was able to raise $400 million for the second time in a month. There was more progress on US deregulation, at least from Wall Street’s point of view. The Federal Reserve terminated enforcement actions against UBS Group and Credit Suisse stemming from their dealings with Archegos Capital Management. Regulators are also looking at pausing cyber-related exams , letting big new issuers duck many of the agency’s strongest disclosure rules for their IPOs and overhauling the secret grades known as CAMELS that examiners use to gauge the soundness of banks. Some frantic lobbying headed off a new Trump regulation that would have made firms debank their customers if they didn’t have the right citizenship papers, which by one estimate would include nearly half the US population. Investors of a certain age may recall watching helplessly as stock brokers “traded through” their obviously superior bids and left them empty handed...