LectureRechercheRadarCadre d'investissement
Connexion / Inscription
ToutTechnologieProduitsInvestissementBourseAutres

Lecture

2026-10-30115 publications

Actualiser

Intel Stock Plunges 20% After Weak Q4 Guidance Cites PC Market Slowdown

Sujet · 英特尔18A工艺Matériel
2026-10-30 04:04 GMT+8

US chipmaker Intel released its third-quarter earnings report on October 29, 2026. Although adjusted earnings per share and total revenue slightly exceeded analyst expectations, the company's guidance for fourth-quarter revenue fell significantly short of market forecasts, causing shares to plunge approximately 20% in extended trading.

According to CNBC's coverage of the company's data, Intel reported an adjusted EPS of 52 cents for Q3, beating the 40-cent estimate from analysts polled by LSEG. Revenue came in at $16.51 billion, roughly matching the $16.5 billion expectation. However, revenue from the Client Computing and Physical AI group declined 5% year-over-year to $8.1 billion, missing estimates, while Data Center and AI revenue surged 78% to $7.3 billion.

"We see the PC slowdown continuing at least through the first half of next year," said Intel CFO David Zinsner in the earnings release. This statement directly impacted market confidence in the company's near-term performance. For the current quarter, Intel expects adjusted EPS of just 25 cents on revenue between $15 billion and $16 billion, well below the analyst consensus of $17.1 billion in revenue and 43 cents in EPS.

Despite headwinds in the PC business, Intel CEO Lip-Bu Tan emphasized optimism about the broader portfolio, noting robust long-term demand for CPUs and all computing platforms. The rise of agentic AI applications, such as Meta's Muse and OpenAI's Dots, which rely on CPUs for task orchestration, is seen as a potential growth area for Intel amidst the GPU-dominated AI wave.

Sources :CNBC

Recherche

US EV maker Rivian licenses AI voice assistant to Volkswagen Group, expanding $5.8 billion joint venture

Matériel
2026-10-30 04:05 GMT+8

US electric vehicle maker Rivian Automotive Inc. is expanding its $5.8 billion joint venture with German automaker Volkswagen Group to include artificial intelligence. According to Rivian’s third-quarter investor presentation released Thursday, the company has begun licensing elements of its in-house AI voice assistant technology to Volkswagen.

This move marks a shift in the partnership from integrating electrical architectures and software platforms to actively licensing specific AI application layers. While Rivian did not disclose licensing fees, payment terms, or which Volkswagen vehicles will use the technology, speculation suggests the AI assistant could debut in Volkswagen’s planned low-cost EV, the ID EVERY1. This small four-door hatchback is slated to be the first VW vehicle to utilize Rivian’s software and electrical architecture.

Rivian founder and CEO RJ Scaringe stated at TechCrunch Disrupt 2026 that the company built its own agentic framework capable of understanding user intent and routing actions to the appropriate parts of the vehicle’s software. Scaringe noted this capability could eventually allow Rivian vehicles to run AI agents created by other companies, including Anthropic. Rivian described the expansion to include AI assistant elements as “a proof point for the extensibility of the platform we’ve built in-house.”

Previously, Volkswagen committed to investing up to $5.8 billion in Rivian by 2027 to secure its next-generation electrical architecture and software. The addition of AI assistant licensing further validates the commercial spillover of Rivian’s vertical integration strategy, though its specific financial contribution remains unknown due to undisclosed terms.

Sources :TechCrunch

Recherche

JPMorgan Initiates Coverage on Palantir with Overweight Rating and $262 Price Target

Sujet · PalantirMatériel
2026-10-30 00:41 GMT+8

U.S. investment bank JPMorgan Chase & Co. initiated coverage on AI software company Palantir Technologies Inc. (NASDAQ: PLTR) on October 29, 2026. Analyst Brian Essex assigned an 'Overweight' rating and set a 12-month price target of $262.

This target implies approximately 31.98% upside from Palantir's recent trading levels. Essex cited accelerating operational momentum and renewed demand for artificial intelligence as the basis for the bullish thesis. JPMorgan highlighted that Palantir has achieved 12 consecutive quarters of revenue growth, with the most recent quarter showing a 93% year-over-year increase to $1.93 billion.

The initiation comes just before Palantir's scheduled third-quarter earnings report on November 2, 2026. Other institutions, including Barclays and Oppenheimer, have recently reiterated buy ratings or raised price targets for the stock. According to TipRanks, the average 12-month price target among 25 Wall Street analysts surveyed over the past three months stands at $218.95.

Sources :Finbold

Recherche

US EV Maker Rivian Raises Full-Year Delivery Guidance to At Least 67,000 Units as R2 SUV Becomes Best-Seller

Sujet · Rivian交付量Matériel
2026-10-30 04:02 GMT+8

US electric vehicle manufacturer Rivian Automotive announced on October 29 that it has raised the lower bound of its full-year 2026 EV delivery guidance from 65,000 to 67,000 units. This marks the second time the company has increased its delivery expectations this year. Rivian attributed the adjustment primarily to the strong sales performance of its newly launched R2 SUV in the third quarter, although the company did not disclose the exact number of R2 units sold.

In the third quarter, Rivian produced 19,751 vehicles and delivered 19,248, representing an approximately 45% increase year-over-year and a significant jump from the second quarter. The company stated that the R2 model was responsible for much of the quarter-over-quarter gain. This increase in deliveries signals that Rivian is on track to break a three-year stalemate during which it struggled to ship more than 50,000 vehicles annually.

Financially, Rivian reported third-quarter revenue of $2.2 billion and a net loss of $913 million, narrowing from a $1.2 billion loss in the same period last year. In a letter to shareholders, the company noted that with the US automotive marketplace starved for high-quality EV choices, the attractively priced R2 is resonating with a broad set of consumers and is expected to be a game changer for Rivian’s long-term growth and profitability.

The R2 SUV launched in June with a starting price just under $60,000. A more affordable 'premium' trim is scheduled for later this year at $53,990, with even cheaper versions slated for 2027 starting at $44,990. As the US federal EV tax credit ends and regulatory shifts cause traditional automakers to pivot back to hybrids, pure-EV companies like Tesla and Rivian face a changing competitive landscape.

Sources :TechCrunch

Recherche

a16z Podcast: How Defense Demand Finances US Manufacturing Capacity

Sujet · 国防需求驱动制造产能Matériel
2026-10-29 21:50 GMT+8

US venture capital firm Andreessen Horowitz (a16z) published the podcast episode "How to Build the Next SpaceX or Tesla" on October 29, 2026. Hosted by Elena Burger, the episode features a16z partners Erin Price-Wright and Connor Love. The discussion centers on the AD50 list, exploring how the booms in American defense, energy, manufacturing, and robotics are converging into a story of rebuilding the country's industrial base.

According to the episode description, the partners discuss how defense demand can help finance manufacturing capacity that ultimately serves much larger commercial markets. They analyze why AI is changing the economics of American factories and how a new generation of founders is applying lessons from SpaceX, Tesla, Anduril, and Palantir to mining, energy, manufacturing, and defense. They also make the case that these companies can become some of the most valuable in the world, as private capital becomes increasingly willing to fund physical infrastructure and production before government demand is guaranteed.

For readers interested in AI industry investment, this resource provides a specific perspective on the "defense-driven manufacturing" investment logic. While no full transcript is provided, the official summary lists key topics such as vertical integration strategies, DJI's transition from drones to robots, and why hardware has become the most valuable game. This helps contextualize the differences in capital structure and market dependency between current hard-tech startups and traditional software companies.

Sources :youtube.com

Recherche

Ukraine Deploys Remote-Controlled Wheeled Robots for Battlefield Wire Laying; Unmanned Missions Surge Nearly Threefold in Six Months

Sujet · 乌克兰无人地面车辆部署Prise rapide
2026-10-30 03:20 GMT+8

According to TechRadar, Ukraine's Ground Forces 16th Separate Support Brigade has begun deploying remote-controlled wheeled unmanned ground vehicles (UGVs) to carry out battlefield wire-laying tasks. These robots are equipped with specially designed frames to hold coils of wire, releasing material gradually as they move forward. This allows defensive obstacles to be constructed without requiring military engineers to leave protected positions and expose themselves to direct enemy fire.

"Installing barriers in an open area is always a risk for combat engineers. But today, this hard work is done by ground-based drones," the Ground Forces of Ukraine stated on social media. Currently, the specific manufacturer and model of the robot remain unknown, and key performance metrics such as carrying capacity, operational distance, and the amount of fencing deployed per mission have not been disclosed.

This deployment extends Ukraine's expanding use of unmanned terrestrial systems. Data from Ukraine's DELTA military management platform shows that recorded robotic operations have increased sharply: from 7,511 in January 2026 to 19,940 in July. By early September, the total had reached approximately 112,000 missions. These figures cover all types of robotic operations rather than just wire installation, indicating that remotely operated equipment has become deeply integrated into routine military logistics and engineering support.

Sources :TechRadar

Recherche

Microsoft Reports Q1 FY27 Results: Revenue Up 19%, Azure Grows 48%, Commercial RPO Surges 73%

Sujet · 微软AI发债融资Structurel
Vérifié 2026-10-30 03:11 GMT+8

US technology company Microsoft Corp. announced its financial results for the first quarter of fiscal year 2027, ended September 30, 2026, on October 28, 2026. Total revenue for the quarter was $92.3 billion, an increase of 19% year-over-year. Operating income rose 21% to $45.9 billion. On a GAAP basis, net income increased 32% to $36.6 billion, with diluted earnings per share at $4.92. Excluding the impact from investments in OpenAI, non-GAAP net income was $36.4 billion, up 18%.

Cloud services remained the primary growth driver. Microsoft Cloud revenue reached $63.9 billion, growing 30% year-over-year. Specifically, Azure revenue increased by 48%, reflecting sustained acceleration in demand for artificial intelligence infrastructure. Additionally, commercial remaining performance obligations surged 73% to $677 billion, a key metric often viewed as a forward-looking indicator of future revenue visibility.

In terms of capital return, Microsoft distributed $10.2 billion to shareholders through dividends and share repurchases during the quarter. CEO Satya Nadella stated that the company is helping organizations build and own their own intelligence, translating innovation into real-world impact across industries. CFO Amy Hood noted that results exceeded expectations across revenue, operating income, and earnings per share.

Sources :sec.gov

Recherche

Ben Carlson, U.S. Wealth Manager at Ritholtz Wealth Management: A Primary Residence Is a Consumer Good; Comparing Its Returns to the S&P 500 Is Asking the Wrong Question

Matériel
2026-10-30 02:01 GMT+8

A certified public accountant in his late twenties posed what seemed like a rational question to Ben Carlson of Ritholtz Wealth Management in the United States: Mortgage rates are above 7%, home insurance premiums are rising, it takes longer to save for a down payment, property taxes are higher due to increased home values, and home prices themselves refuse to decline—meanwhile, the stock market continues to boom, bond yields are higher, and options strategies offer double-digit returns. From a purely numerical perspective, is buying a house now the worst financial asset?

Carlson’s response was unexpectedly direct: I think you’re asking the wrong question. A primary residence is consumption, not investment. You have to live somewhere anyway. Therefore, comparing buying a home to the S&P 500 is not a realistic benchmark—you can’t live inside an index fund, and you can’t invite friends over for dinner in an ETF.

This categorical distinction does not mean housing lacks financial attributes. Carlson acknowledges that real estate can serve as a store of value, accumulating equity over time; it can be borrowed against for renovations; and it offers favorable tax treatment. The leverage effect also means that when you borrow significant funds to purchase, prices do not need to rise much to generate a solid return on your down payment. However, while index fund investment expenses are transparent and straightforward, real estate carries numerous hidden costs—realtor fees, moving expenses, closing costs, property taxes, insurance, maintenance, renovations, lawn care, and more. Housing is also an illiquid asset, which is both a benefit (forcing savings) and a drawback (limiting flexibility).

Carlson did not deny that affordability is indeed worsening. He calculated the average monthly mortgage payment using the median price of existing U.S. homes, assuming a 30-year fixed-rate mortgage with a 20% down payment, and found that payments have more than doubled since the late 2010s. Home prices rose by over 50% in the 2020s, interest rates more than doubled, and insurance costs increased by nearly 50%. The Case-Shiller National Home Price Index indicates that this year may mark the second consecutive year where national home price growth lagged behind the inflation rate. He also noted that the market performance of the 2020s has already pulled forward approximately a decade’s worth of gains, implying that future returns should logically be lower.

However, he simultaneously acknowledged uncertainty: I do not know if housing will provide steady appreciation in the future. If forced to guess, keeping pace with inflation would be a reasonable baseline expectation. Demographically, young people remain in their home-buying years, but in the 2030s, as the baby boomer generation passes away, downsizes, or moves to Florida, a significant amount of inventory will likely enter the market. Bullish arguments for housing include the fact that construction volumes are genuinely insufficient, and once interest rates fall to more reasonable levels, housing activity could surge—unless, of course, the reason for falling rates is that we are experiencing a financial crisis.

Sources :A Wealth of Common Sense

Recherche

Epsilon (Publicis Groupe Data Technology Company) Platform Team: AI Test Trajectory Replay Skips Redundant Model Inference

Sujet · EpsilonPrise rapide
2026-10-30 02:20 GMT+8

When UI testing is delivered as a managed platform capability to multiple product teams, each team performs high-frequency repeated executions of the same set of test workflows. If every execution relies on large language models for step-by-step reasoning, latency and costs grow linearly with execution frequency, making the economic viability of platform delivery difficult to sustain. The Agentic Platform team at Epsilon (50 engineers, part of Epsilon, a data and marketing technology company under Publicis Groupe) described their solution in an AWS blog post, built on Amazon Nova Act and the Amazon Bedrock AgentCore Browser Tool, where the trajectory replay mechanism is the key design that makes this model economically feasible.

Trajectory replay divides the testing lifecycle into two states. In the draft state, Nova Act calls large language models to reason through each natural language step, determining how to perform actions in the browser; after typically four to five rounds of prompt iteration, a successful trajectory is saved. In the active state, the system no longer calls large language models to reason, but instead replays the saved trajectory directly as a deterministic sequence of operations. Epsilon reports that saved trajectories achieve a 100% pass rate during repeated executions—this is its internal benchmark data, published in an AWS customer case study blog post, and has not been independently reproduced.

The applicability conditions of this mechanism need to be clarified. The 100% pass rate describes the deterministic re-execution of stable workflows that have already been successfully explored, rather than Nova Act's ability to handle new scenarios or interface changes. If the target page undergoes structural changes, the saved trajectory may no longer match, requiring a return to the draft state for exploration. Therefore, trajectory replay reduces the cost of regression-style repeated executions, not the cost of adaptive testing. For products with frequently changing interfaces, the model inference overhead during the exploration phase will still occur periodically.

Trajectory replay becomes a necessary component of the platform architecture due to Epsilon's multi-tenant delivery model. Product teams consume UI testing capabilities through a unified API interface, with each team running tests under scope-limited IAM roles within their respective AWS accounts, while the platform team does not own the browser infrastructure. In this model, if every execution triggers a complete model inference loop, the platform-side inference costs will accumulate with the number of teams and execution frequency. Trajectory replay reduces the cost of repeated executions from model inference to trajectory lookup and deterministic operations, enabling economic feasibility for the multi-tenant platform at scale.

Sources :Amazon Web Services

Recherche

MoEngage's Engineering Practice: Turning 400 Private APIs into Agent-Discoverable Tools with Cedar Policies and Semantic Search

Sujet · MoEngageMatériel
2026-10-30 00:54 GMT+8

When engineering teams begin building AI Agents, the capabilities those agents need usually already exist as internal APIs. This was exactly the case for MoEngage: this customer engagement platform serving over 1,350 consumer brands relies on roughly 400 internal REST APIs, maintained by 25 engineering teams and running within private Virtual Private Clouds (VPCs) across seven AWS Regions. The real challenge was not building new functionality, but enabling these APIs to be called safely by Agents without exposing them to the public internet, without giving models control over identity, and without flooding context windows with hundreds of tool definitions.

The initial intuitive approach was for each team to expose its own MCP server. After estimating the costs, MoEngage rejected this: having over 20 MCP servers meant over 20 deployments to monitor, patch, and keep protocol-compatible; cross-domain Agents would need to connect to multiple servers and coordinate between them; and every team would need MCP expertise unrelated to their actual product work. The principle established here ran through all subsequent designs: engineering teams do not implement MCP, they only build REST APIs and declare which operations should be treated as tools; the AI Platform team owns the MCP layer end-to-end; and a team's responsibility ends at writing a good OpenAPI specification.

This principle immediately created three requirements: a central tool surface that Agents connect to just once, server-side authorization enforced by the Agent, and private connectivity that does not need to be rebuilt for every new API. Building a central MCP server from scratch could solve the twenty-servers problem, but it would still require manually building a tool-generation pipeline, policy file formats, alias registries for renames, and circuit breakers—all undifferentiated work for the platform team that would not make MoEngage’s Agents smarter or improve campaign outcomes for customers.

The four attributes of the AgentCore Gateway map precisely to these constraints. First, OpenAPI in, MCP tools out: registering a specification automatically generates tools in the gateway; when the spec changes, the tools change, requiring no deployment or protocol expertise. Second, Cedar policies are evaluated before catalog serialization: the gateway uses partial evaluation to filter the response to tools/list per principal. Each Agent authenticates with its own IAM role, and its policy determines which tools it can see and call. This is what makes 400 tools manageable—not client-side post-hoc filtering, but server-side scoping before the response is built. A campaign QA bot, for example, can only see read-only tools; write operations simply do not appear in its tools/list response.

Third, managed egress via Amazon VPC Lattice: the gateway reaches internal services through AgentCore managed egress, while inbound access is only possible through VPC interface endpoints (private DNS), leaving no public path. Onboarding a new API is merely adding a target, not changing network topology—this breaks the operational explosion caused by multiplying by seven Regions. Fourth, semantic search as a gateway capability: Agents discover tools by describing tasks rather than memorizing tool names. MoEngage’s client library exposes only one meta-tool, , to the model. The model describes its needs, the gateway returns candidates, and only those candidates become callable for the remainder of the turn. The full catalog never enters the context.

Sources :Amazon Web Services

Recherche

AMD Releases ROCm HRX Loom SDK v0.1: A Unified Compiler Toolchain for CPUs, GPUs, and NPUs

Sujet · AMD Loom SDKPrise rapide
2026-10-29 18:22 GMT+8

AMD released version 0.1 of the ROCm HRX Loom SDK on October 29, 2026. This marks the first tagged release of the SDK, which aims to provide developers with a specialized compiler and intermediate representation (IR) stack optimized for agentic workloads across AMD's CPU, GPU, and NPU architectures.

According to the documentation, Loom compiles kernels and asynchronous device programs, links model programs, and specializes them for execution on multiple AMD hardware types. It carries source facts, target facts, configuration, correctness cases, and compiler reports through a single authoring and execution path.

Currently, Loom supports targeting AMDGPU HSA emission for GPU execution, SPIR-V/Vulkan for portable execution, and x86 and WebAssembly support that is still maturing. Developers can access the SDK via GitHub or install it using Python's pip as the hrx-loom package for experimentation.

Sources :phoronix.com

Recherche

AI Expert Network Startup SuperMe Raises $20M Led by Forerunner

Sujet · SuperMePrise rapide
2026-10-30 00:19 GMT+8

According to TechCrunch, AI startup SuperMe announced on October 29, 2026, that it has raised $20 million in funding. The round was led by Kirsten Green, a partner at Forerunner Ventures, with participation from Greylock Partners.

SuperMe's core product is an "AI expert network." It creates digital AI avatars of real-world industry experts using public sources like LinkedIn, YouTube, and Substack, as well as user-authorized private data such as Google Drive and Notion. When users ask career or business questions, the platform matches them with the most suitable expert AIs and generates a group chat featuring multiple AI personas to provide answers.

The company was founded in 2024 by Casey Winters and Ludo Antonov. Winters previously held executive roles at Grubhub, Pinterest, and Eventbrite, while Antonov led engineering teams at Pinterest, Lyft, and Whatnot. SuperMe currently has about 13 employees and plans to use the funds to expand its user base, increase expert supply, and improve AI models for better question-to-expert matching.

SuperMe is currently free for users but plans to cap the number of questions and transition to a paid model in the future. This funding event reflects venture capital interest in converting human professional knowledge into scalable AI assets, although long-term profitability and user willingness to pay remain to be validated by the market.

Sources :TechCrunch

Recherche
Suivant →
Connexion / Inscription
LectureRechercheRadarCadre d'investissement

Carlson’s final advice is pragmatic: Always crunch the numbers when deciding between renting and buying; in some areas, renting and investing the difference in the stock market may be financially superior; in others, locking in a 30-year fixed-rate mortgage, refinancing when yields drop, and slowly building equity makes sense. Regional differences are vast, and location matters greatly. But regardless of the choice, do not evaluate your house the same way you evaluate portfolio holdings. Reasons to buy include putting down roots, stability, hedging against inflation, eliminating your largest monthly expense in retirement, and leaving something for heirs. Do the math, but don’t let perceived returns from real estate dictate your decisions. Housing is more consumption than investment.

search_tools

Identity handling is another key design decision. Every internal API call carries tenant, user, role, and database context. If this information were passed as LLM-visible tool parameters, the model might get it wrong or be induced to modify it. MoEngage passes identity as request headers rather than tool parameters: the client library converts the Identity object into headers prefixed with X-MOE-*, which the model cannot see or set. Discovery and execution have different identity requirements—listing and searching tools carry no tenant identity, only tools/call does—keeping discovery lightweight and ensuring an unambiguous audit trail.

The onboarding process is fully automated: teams add an x-moe-mcp block to their existing OpenAPI specification to declare tool descriptions and parameter overrides. Upon push, the CI pipeline validates the declaration, registers the target with the gateway, and applies Cedar policies, requiring no manual steps or tickets to the platform team. MoEngage found that directly reusing API descriptions led to poor tool selection—the model picked tools based on implementation details (e.g., returning paginated lists) rather than intent (e.g., finding campaigns by status or channel). The tool_description and param_overrides fields exist specifically to write model-facing text without altering developer-facing documentation.

It should be noted that this article was co-written by AWS and MoEngage; the estimated saving of six months of engineering effort and the reduction of onboarding time from days to hours are self-estimates by the parties involved, with no independent measurement. The solution is deeply tied to AWS infrastructure (AgentCore, VPC Lattice, SigV4), and cross-cloud portability is not discussed. The rename tolerance brought by semantic search (matching by description rather than identifier) is a practical benefit, but it also means that the quality of tool descriptions directly determines whether an Agent can find the correct tool—a burden requiring continuous maintenance. The search-first mode changed how MoEngage thinks about Agent design: new Agents start with zero hardcoded tool references, describe intent, receive candidates, and call the tools they find; when teams rename operations or release new ones, existing Agents can pick them up without code changes.