LecturaInvestigaciónRadarMarco de inversión
Iniciar sesión / Registrarse
TodoTecnologíaProductosInversiónBolsaOtros

Lectura

2026-11-11104 publicaciones

Actualizar

US October CPI Rises 0.4% MoM, Beating Consensus; December Rate Hike Odds Climb to Nearly 10%

Material
Verificado 2026-11-11 07:51 GMT+8

The U.S. Bureau of Labor Statistics (BLS) released data on November 10, 2026, showing that the Consumer Price Index (CPI) rose 0.4% month-over-month in October, higher than the 0.3% consensus estimate surveyed by Bloomberg. The core CPI, which excludes food and energy, increased 0.3% month-over-month, also surpassing the expected 0.2%. This marks another upward trend in inflation following a significant 0.6% rise in September.

Component details indicate that the shelter index rose 0.3% in October, accounting for over thirty percent of the monthly all-items increase. The food index increased 0.5%, with food at home up 0.4% and food away from home up 0.5%. The energy index continued to climb by 0.7% after a 5.2% surge in September, driven by a 1.0% rise in gasoline prices. Notably, due to missing data collection during the 2025 government shutdown, the BLS could not publish 12-month percent changes for most series.

Following the release, financial markets quickly repriced the Federal Reserve's policy path. According to CME futures data, as of the afternoon, the probability of a 25 basis point rate hike at the December meeting rose from 0% to 9.8%. Simultaneously, U.S. Treasury yields generally increased, reflecting heightened investor concerns about sticky inflation and its impact on monetary policy. This shift signals pressure to revise the previous trading logic based on expectations of cooling inflation.

Fuentes:Bureau of Labor Statistics

Investigación

Meituan Open-Sources 1.6T Parameter Multimodal Model LongCat-2.5 Under MIT License

Material
2026-11-10 13:14 GMT+8

Meituan's LongCat team officially open-sourced its native multimodal large model, LongCat-2.5, on the evening of November 9. The model features 1.6 trillion total parameters, activates approximately 48 billion parameters per token, and supports a context window of 1 million tokens. Released under the permissive MIT license, it allows for commercial inference deployment and secondary training without additional authorization requirements.

Technically, LongCat-2.5 focuses on optimizing visual programming, complex codebase development, and long-horizon agent tasks. To address the limitation of domestic AI accelerators having less than 80GB of VRAM, Meituan provided NPU deployment instructions based on the SGLang-FluentLLM project. The release includes a W4A4 quantized version, which quantizes expert weights and activations to 4 bits. According to official data, this version improves prefill and decode throughput by 43.3% and 42.77% respectively compared to the W8A8 INT8 baseline.

In benchmark comparisons against eight other major models including Claude Opus 5 and GPT-5.6-Sol, LongCat-2.5 ranked first in four specific tests: BrowseComp (web retrieval), ALE v1.1 (cross-domain agents), ResearchRubrics (deep research), and 3DCodeBench (text-to-3D code generation). However, its performance in some general coding tests still trails behind competitors like DeepSeek-V4.1-Flash.

Regarding API pricing, LongCat-2.5 currently offers a limited-time rate of 2 RMB per million input tokens and 8 RMB per million output tokens, significantly lower than comparable overseas models, though its price advantage over top domestic models is less pronounced. Early developer feedback indicates strong capabilities in generating interactive 3D scenes and game frontends, but also notes issues with circular reasoning that can reduce execution efficiency.

Fuentes:zhidx.com

Investigación

Amazon Acquires Fauna Robotics to Rebuild Home Robot Strategy

Tema · 亚马逊个人机器人组Material
2026-11-08 21:00 GMT+8

According to CNBC, Amazon is restructuring its home robot efforts by acquiring New York-based startup Fauna Robotics and integrating it into a newly formed "Personal Robotics Group." This team also includes staff from Lab126, the hardware lab behind Kindle and Echo, as well as former members of the Astro project.

Fauna Robotics developed Sprout, a 3.5-foot-tall humanoid robot capable of walking, gripping, dancing, and holding conversations. Currently sold to research labs for approximately $50,000, Sprout counts Disney among its early customers. An Amazon spokesperson stated that the acquisition reflects the company's commitment to "tackling the challenges that can unlock the personal robotics category."

This move marks a strategic shift in Amazon's consumer robotics approach. Its first home robot, Astro, launched in 2021 as a wheeled tablet but never moved beyond invite-only beta testing, reaching only a few thousand homes. Additionally, Amazon's plan to acquire vacuum maker iRobot collapsed in early 2024 due to EU regulatory scrutiny, resulting in a $94 million breakup fee.

Advances in AI and robotics have positioned humanoids as a key inflection point. Barclays forecasts that the global humanoid market could grow from $2-3 billion today to roughly $40 billion by 2035. Amazon CEO Andy Jassy recently cited AI and robotics as among the "biggest inflections of our lifetime" in his annual shareholder letter.

Fuentes:CNBC

Investigación

US Investment Firm Rhumbline Advisers Discloses $133.8 Billion US Equity Portfolio Heavy on AI Infrastructure

Material
Verificado 2026-11-11 03:50 GMT+8

US investment firm Rhumbline Advisers has filed its Form 13F-HR with the U.S. Securities and Exchange Commission (SEC) for the quarter ended September 30, 2026. The filing discloses approximately $133.84 billion in reportable US equity securities. It is important to note that this figure represents only the market value of specific stocks held in US markets as of the quarter-end snapshot; it does not reflect the firm's total Assets Under Management (AUM), which remains undisclosed in this form.

The portfolio reveals substantial exposure to the artificial intelligence sector. Key holdings include significant positions in Broadcom Inc. (approximately $2.89 billion), Meta Platforms Inc. (approximately $2.79 billion), and Advanced Micro Devices Inc. (approximately $1.72 billion). Other notable tech and AI-related holdings include Micron Technology ($2.11 billion) and Cisco Systems ($743 million). These figures indicate deep institutional capital participation in the AI infrastructure and platform layers.

13F filings are quarterly snapshots required from large institutional investors, reflecting holdings at quarter-end rather than real-time trading activity or complete asset portfolios (excluding shorts, detailed derivatives, and non-US assets). Despite the lag, Rhumbline's disclosure provides concrete quantitative evidence of how top-tier institutional capital is positioned within the AI ecosystem.

Fuentes:sec.gov

Investigación

Smart Ring Maker Oura Acquires AI-Native Primary Care Startup Counsel Health for $625M After Postponing IPO

Tema · Oura HealthMaterial
2026-11-11 03:23 GMT+8

US-based smart ring manufacturer Oura Health announced on November 10, 2026, that it will acquire AI-native primary care startup Counsel Health for $625 million. The deal comes just a few weeks after Oura indefinitely postponed its up to $2.2 billion IPO, citing "uncertainty in the IPO market."

Counsel Health is an AI-powered healthcare platform that combines a chat-based medical experience with access to licensed physicians. Its service was integrated into the Oura app in June, allowing users to get personalized medical advice via AI and chat with doctors when needed. The acquisition aims to expand their partnership across preventative, acute, specialty, and chronic care.

"Oura helps members understand their physiology and, with Counsel, they can take action on their health across a range of needs, from quick questions and prescriptions to lab results, chronic care, and referrals," said Oura CEO Tom Hale. According to Bloomberg, Counsel Health CEO Muthu Alagappan will join Oura as chief clinical AI officer, along with the company's 70 employees.

Financial disclosures reveal that Oura's revenue grew 153% year-over-year to $533 million for the quarter ended September 30. For fiscal year 2026, Oura generated approximately $1.75 billion in revenue, marking 92% year-over-year growth. The company reached an $11 billion valuation following a $900 million late-stage funding round in October 2025. Counsel Health had raised a $25 million Series A led by Andreessen Horowitz (a16z) and GV last year, with Pitchbook estimating total fundraising at around $36 million.

The deal is expected to close before the end of 2027. Oura stated that Counsel will continue serving existing and new enterprise partners post-acquisition, aiming to expand its AI-native care model beyond Oura's membership base.

Fuentes:TechCrunch

Investigación

AI Firm Anthropic Files Updated S-1, Seeks Up to $88.5 Billion in IPO and Private Placement

Estructural
2026-11-11 02:51 GMT+8

U.S.-based artificial intelligence company Anthropic submitted an updated S-1 registration statement on Tuesday, revealing details of its upcoming initial public offering (IPO) and concurrent private placement. According to Crunchbase News, the company plans to sell over 800 million shares in the IPO portion at a price range of $70 to $78 per share, expected to raise approximately $62.5 billion. Additionally, Anthropic anticipates pulling in another $26 billion from a private placement involving investors such as AMD, Broadcom, Google, Micron Technology, Microsoft, Nvidia, and TSMC. Combined, this totals up to $88.5 billion, potentially marking the largest-ever fundraise tied to a public market debut.

The updated filing sets a fully diluted valuation cap of up to $1.4 trillion for the San Francisco-based company. This figure comes in below earlier projections where many investors reportedly forecasted closer to $2 trillion. Nevertheless, even at $1.4 trillion, Anthropic would go public at the second-highest valuation in history, topped only by SpaceX's June debut. The company is expected to list on Nasdaq under the ticker symbol ANT later this month.

Financial disclosures show that Anthropic generated $33.2 billion in revenue for the first nine months of the year, up from $2.5 billion in the same period a year ago. Despite this rapid growth, the company is not yet profitable, posting an operating loss of $1.3 billion and a net loss of $104 billion for the nine-month period. The substantial capital raise is intended to support its expensive expansion and infrastructure investment plans.

Fuentes:Crunchbase News

Investigación

Norwegian Pension Fund KLP Discloses $32.3 Billion US Equity Portfolio Heavy on AI Sector

Material
Verificado 2026-11-11 03:00 GMT+8

Norwegian pension fund manager KLP Kapitalforvaltning AS has filed its Form 13F-HR with the U.S. Securities and Exchange Commission (SEC) for the quarter ended September 30, 2026. The filing discloses approximately $32.33 billion in reportable US equity securities. It is important to note that this figure represents only the market value of specific stocks held in US markets as of the quarter-end snapshot; it does not reflect the fund's total Assets Under Management (AUM), which remains undisclosed in this form.

The portfolio reveals substantial exposure to the artificial intelligence sector. Key holdings include significant positions in Apple Inc. ($2.25 billion), Alphabet Inc. ($1.62 billion across Class A and C shares), Amazon.com Inc. ($978 million), Advanced Micro Devices Inc. ($438 million), and Broadcom Inc. ($674 million). Other notable tech and AI-related holdings include Cisco Systems Inc., Analog Devices Inc., Arista Networks Inc., CrowdStrike Holdings Inc., and Constellation Energy Corp.

13F filings are quarterly snapshots required from large institutional investors, reflecting holdings at quarter-end rather than real-time trading activity or complete asset portfolios (excluding shorts, detailed derivatives, and non-US assets). Despite the lag, KLP's disclosure of such a massive portfolio with significant AI exposure provides concrete quantitative evidence of how top-tier institutional capital is positioned within the AI infrastructure ecosystem.

Fuentes:sec.gov

Investigación

OpenAI and Upwork Study: AI Increases Freelance Project Task Diversity by 45%, Raises Budgets

Material
2026-11-10 15:00 GMT+8

OpenAI and Upwork, a US-based online marketplace for freelancers, published a joint study on November 10, 2026, analyzing the impact of AI on the freelance market. The research finds that as AI usage grows, the structure of freelance demand is shifting significantly: many activities common in ChatGPT (such as writing) have lost share on Upwork, while less common activities have gained share. Of the 63 work categories analyzed, 43 followed this pattern.

Notably, AI is not simply replacing human work but changing how work is "packaged." The study shows that between early 2023 and late 2025, the mix of tasks in AI-related postings became 45% more diverse—equivalent to an increase from about 15 equally common tasks to about 21. Clients are increasingly asking freelancers to combine AI capabilities with other skills, such as requesting AI-generated video, editing, and motion graphics in a single project brief.

Furthermore, AI-related projects exhibit larger budgets and involve more work hours. While posted hourly rates were similar for AI-related and other work, AI-related projects had larger fixed-price budgets and involved more hours. This suggests that AI may be expanding the scope of projects rather than just accelerating individual tasks.

The study relies on Upwork marketplace data (through April 2026) and ChatGPT usage data (Jan 2025–Jun 2026). OpenAI notes that these findings do not point to a simple substitution story, but rather indicate that AI is shifting both what clients seek from freelancers and how they assemble projects.

Fuentes:OpenAI

Investigación

AMD Announces General Availability of 6th Gen EPYC Processors for Agentic AI Infrastructure

Material
2026-11-10 23:39 GMT+8

U.S. semiconductor company AMD announced on November 10, 2026, the general availability of its 6th Gen EPYC™ server SP7 processors, based on the new “Zen 6” architecture. This generation aims to provide a computing foundation for growing agentic AI deployments and demanding enterprise applications. AMD stated that systems powered by the new CPUs are shipping from all leading server vendors, while cloud instances from Microsoft Azure and Oracle Cloud Infrastructure (OCI) are coming soon.

Dan McNamara, Senior Vice President and General Manager of Compute and Enterprise AI at AMD, noted that AI is significantly expanding the role of the CPU across every layer of the data center. The 6th Gen EPYC portfolio is workload-optimized for different roles, including general-purpose services, GPU host nodes, and agent sandboxes. For instance, high-frequency host-node processors offer up to 5.3 GHz and claim to deliver up to 1.8x the LLM output throughput of previous generation EPYC processors; for hosting agent sandboxes, high-density EPYC processors claim an advantage of ~3.3x agents-per-CPU-watt over Nvidia Vera.

It is important to note that these performance comparisons are AMD internal estimates or based on specific benchmarks (such as SPECrate®2017_int_base) and represent vendor self-reported results, not independently verified real-world production performance. Previously, in October, AMD released the Helion rack-scale reference design blueprint, which only outlined future availability; this announcement confirms the actual delivery status of the underlying silicon, marking the official market entry of the new generation of x86 server CPUs.

Fuentes:AMD Newsroom

Investigación

Microsoft lobbies against potential FCC ban on Chinese optical transceivers, warning of stalled AI data center rollout

Tema · FCC中国光模块禁令Material
2026-11-11 01:51 GMT+8

According to Reuters, Microsoft is lobbying the US government against a plan by the Federal Communications Commission (FCC) to potentially ban imports of Chinese optical transceivers. The company has reportedly met with the White House Office of Science and Technology Policy and the National Security Council, working alongside the Information Technology Industry Council (ITI).

Optical transceivers are critical components in data centers, converting electrical signals into light signals to transmit data at high speeds between AI GPUs and server racks via fiber optics. Chinese companies currently supply 66% of the global market. A ban would more than halve the available supply in the US, exacerbating existing shortages.

John Miller, Executive Vice President of Policy and General Counsel at ITI, stated that the move "really threatens to arguably grind (the data center rollout) to a halt." He noted that without these transceivers, purchased AI chips would sit idle waiting for connectivity. While there are security concerns regarding potential data theft or remote disruption from Chinese equipment, only two US-based companies, Coherent and Lumentum, produce the necessary optical transceivers, and they lack the immediate scale to replace Chinese supply.

A compromise floated by the Trump administration involves banning only Chinese optical transceivers capable of hitting 3.2 terabits per second. Although this technology exists, mass production is not expected to expand until 2028, with installation beginning in 2029. This proposal aims to give American producers time to ramp up output but remains under consideration.

Fuentes:Tom's Hardware

Investigación

IBM and Microsoft Launch 'Imagine Enterprise' AI Agent Platform for Regulated Industries

Material
2026-11-10 21:00 GMT+8

US tech giants IBM (NYSE: IBM) and Microsoft (NASDAQ: MSFT) jointly announced on November 10, 2026, the launch of a new AI agent platform named "Imagine Enterprise." The platform is designed to help enterprises build, run, and govern AI agents for executing industry-specific processes, with a particular focus on regulated sectors such as financial services, insurance, and retail.

According to the joint press release, IBM Consulting contributes more than 6,000 pre-built AI agents and assets, along with delivery experience from its IBM Consulting Advantage platform. Microsoft provides the underlying AI, data, cloud, and security technologies. Forward-deployed engineers from both companies will support qualifying customers in moving from experimentation to production.

The platform's core advantages are described as instant expertise (leveraging pre-built agents), fitting the enterprise (connecting to existing data and systems), and acting as an intelligence multiplier (integrating enterprise knowledge). Specific technical components include Microsoft Foundry, GitHub Copilot, Copilot Studio, Fabric, Entra, Purview, Defender, and Azure, alongside IBM's Process Studio, Context Studio, and Advantage Core.

Financial services firm SEI (NASDAQ: SEIC) was cited as an early adopter. Sneha Shah, Chief AI Strategist at SEI, stated that the collaboration aims to explore how emerging capabilities can support critical business processes and scale responsibly. Mohamad Ali, Senior Vice President at IBM Consulting, noted that clients are no longer just looking for faster productivity but want to build an "agentic business."

Imagine Enterprise initially targets financial services, insurance, and retail, with plans to expand into manufacturing, energy, public sector, and mobility. Clients can engage IBM Consulting to deploy and operate the platform or choose self-management options.

Fuentes:prnewswire.com

Investigación

OpenAI Chief People Officer KC: AI Renders the Distinction Between Junior and Senior Talent Meaningless

Material
2026-11-10 08:00 GMT+8

KC (formerly Chief Information Officer at Palantir), Chief People Officer at OpenAI, recently made a judgment regarding organizational talent assessment in a public sharing session: the distinction between junior and senior levels is losing its meaning.

His reasoning begins with the drastic shortening of the validity period of experience. KC described the leaps in model capabilities over the past four years using three milestones: two years ago, models were still stuck in parlor tricks like counting how many 'R's are in 'strawberry'; a year ago, reasoning models could solve competition-level mathematics; this September, OpenAI’s agents claimed to have solved the Navier-Stokes Millennium Prize problem within 88 hours (a result that remains controversial in the mathematical community). At such a pace, KC believes that professional knowledge accumulated two years ago may already be obsolete. He suggests completely eliminating the words 'junior' and 'senior,' because in future work contexts, perhaps no one will have experience relevant to that specific context.

KC further depicted the modern work structure he envisions. The model is the brain, while tools like Codex or Claude Code serve as exoskeletons—providing skills, tools, and cross-agent contextual runtime. Above this sit various incarnations: first assistants, then experts. He pointed out a common issue in existing enterprises: most companies treat employees as middleware for their own systems, completing the work first and then logging it into the system, which he calls 'work on work.'

Once system updates are delegated to agents, humans begin collaborating with other humans. KC believes this redefines everyone's title: add a suffix—'agent manager.' He gave an example, stating that he is not a salesperson, but a sales agent manager.

He simultaneously emphasized that emotional intelligence will become humanity's core differentiating capability. He recounted a story from an oncologist's funeral: a patient had seen several doctors who provided the same diagnosis and treatment plan, but ultimately chose this one for his attitude. KC believes that when everyone can access tools surpassing their own intelligence, what still distinguishes people is how they show up as human beings, how they demonstrate integrity, and how genuinely humble they are in their desire to learn.

KC summarized the choices facing companies into two paths: use AI to complete the same tasks faster, or reimagine what is possible. The first path increases profits, while the second changes the trajectory of the enterprise. Initially, the slopes appear identical, but geometric compounding eventually creates explosive separation.

It should be noted that this is KC's personal judgment based on internal observations at OpenAI and has not yet been validated across industries. The 'agent manager' framework assumes that AI agents are mature enough to undertake system maintenance work, whereas most organizations are currently in the early stages of this transition. Furthermore, the rate of experience decay varies significantly across fields—in highly regulated industries or those heavily reliant on tacit knowledge, the applicability of this logic may be greatly diminished.

Fuentes:Tomasz Tunguz

Investigación
Siguiente →
Iniciar sesión / Registrarse
LecturaInvestigaciónRadarMarco de inversión