The one-page fundamentals report for every US stock

RBC Capital Markets' Amy Wu Silverman warns markets may be underpricing AI doomsday risk. AI’s reach across industries has blurred signals like correlations, meaning diversification may appear safer even as a single AI-driven theme dominates many sectors. The three-month implied correlation index sits near its annual low, but Silverman argues it understates risk because AI spending and compute demand bind utilities, energy, politics, and capex together. She contends the probability of a doomsday-style shock—left-tail events and cascading financing costs—may be higher than markets price. With AI investment accelerating and infrastructure expanding into non-tech areas, even routine events such as elections could trigger outsized moves. In short, the AI trade could embed systemic risk that markets have yet to reflect.

Elon Musk and Delta Air Lines CEO Ed Bastian are publicly feuding over in-flight Wi‑Fi, sparked by reported remarks at a Delta event in which Bastian allegedly said, 'We do not want to be with Elon Musk' and that Delta had tested Starlink but preferred not to partner with him. Observers say the sticking points were financial and technical: Delta has already lined up Amazon's Leo satellite service, along with Viasat and Hughes for current coverage, to bundle with cloud and content offerings and to use Delta's Sync login portal—something Starlink reportedly blocks. United and American have moved ahead with Starlink, with several others signed on. Amazon's Leo won't enter beta until next year, potentially delaying Delta's deployment while Starlink's service attracts other carriers. Musk chimed in with taunts about AI sandboxing on Delta flights. The feud highlights high-stakes competition for global airline connectivity.

McDonald's plans an $8.5 billion, 2036 rollout to modernize its restaurants with Archy, an AI-powered order-taker integrated into a larger ArchIQ system. Archy already handles English and Spanish orders with over 90% accuracy in tests and aims to boost efficiency by reducing labor, catching missing items, and nudging customers toward higher-ticket add-ons. The program follows a previously failed IBM Automated Order Taker trial in 2024 and sits under the McDonald's NEXT initiative to remake menus, remodel restaurants, and improve kitchen visibility. The broader rollout includes partnerships with Google and a shift toward more protein-focused menus with redesigned layouts and open kitchens, with U.S. deployment expected in 2027. Investors sent shares lower after the announcement, highlighting the uncertain ROI while executives emphasize a long-run productivity and customer-experience payoff.

Inflation cooled in August, with the PCE index at 3.4% and core PCE at 3.0%, beating expectations and dampening urgency for a Fed rate hike; month-on-month core PCE slowed to 0.2% as BEA revisions to software, legal fees, and investment advice trimmed inflation, reducing the three-month core rate to 2%. Fed officials remain mixed: New York Fed’s Williams signals no hurry to move in October, while Barr cites only two months of data consistent with 2% core inflation; markets price roughly a 35% chance of a hike in October, down from earlier levels. The data could influence policy timing and financial markets.

President Donald Trump and CEOs from Nvidia, OpenAI, Google, Meta, Elon Musk, Sundar Pichai, Dario Amodei, and others signed a set of voluntary AI safety standards at a White House lunch, described by attendees as a constitutional-like move amid rising calls for tighter AI oversight. The document’s signatures drew more attention than the content, with social media reaction focused on who signed rather than what’s written. The event followed recent AI risk warnings and a high-profile Hugging Face incident. It signals a coordinated industry risk-management stance and could influence policy debates, investor sentiment, and competitive dynamics, even as compliance remains voluntary and regulatory clarity remains uncertain.

Synopsys and Amazon unveiled a strategic, multi-year IP deal to accelerate Amazon’s custom silicon and AWS infrastructure. The arrangement expands Synopsys’ silicon IP, EDA, and simulation tools into Amazon as its lead customer and shifts toward a license-plus-royalty model as production volumes rise. The collaboration will optimize silicon-to-system design for Amazon’s Graviton and Trainium platforms and extend to AI-powered engineering with Synopsys’ AI-enabled EDA, physics-based simulation, and agentic AI capabilities. The two companies will collaborate to advance AI across silicon-to-system workflows and create custom agentic AI tools for Amazon’s engineers. Synopsys will deploy AWS services—EC2, cloud storage, Bedrock—to accelerate its own IP and tool development, while Amazon strengthens its AI infrastructure and purpose-built chips portfolio.

Bank of America analyst Wamsi Mohan warns Meta’s Muse and other AI agents could intercept iPhone users’ intents before Apple’s software, threatening discovery, referrals and transaction economics across Apple’s ecosystem. Muse rose to the top of the App Store and drew millions of downloads; partners including Shopify, Expedia and PayPal joined, while Amazon blocked it. The risk is not fewer iPhone sales but losing control of what happens after a device is unlocked. Apple’s Siri AI lags competitive agent capabilities, and a fast-evolving AI landscape could outpace Apple’s update cadence. Apple posted strong Q3 results, with revenue of $109.4 billion, up 16%, and record iPhone, Mac and Services revenue. Investors should watch Siri’s ability to execute transactions, integrate payments and third-party services to preserve Apple’s Services economics.

AI data-center build-out could create about one million jobs, with Nvidia CEO Jensen Huang saying the U.S. is re-industrializing and bringing blue-collar work back. The Alliance for America's Skilled Trades (AAST) estimates 1.7 million skilled-trades openings annually through 2035. Nvidia, Meta, and Microsoft joined AAST this week. Current training programs produce just 55 workers per 100 needed. Huang adds that infrastructure goes beyond data centers to power plants, construction, cooling, and pipe fitters, while Musk notes needs to scale energy and chip production; China, he says, has about three times the electricity production. The push is framed as a major opportunity to reshape the U.S. economy and job mix, driven by construction, energy, and manufacturing tied to AI expansion.

Nebius (NBIS) received an upbeat initiation from William Blair with an Outperform rating, arguing the AI-native cloud platform could differentiate itself. The firm projects roughly $22 billion in ARR by end-2027, supported by a recent $5.75 billion convertible-note financing. Nebius reported Q2 revenue of $582.3 million, up 454% year over year. It has large contracted revenue from investment-grade customers, including a five-year deal with Meta worth up to ~$27 billion and more than $40 billion of additional contracted revenue from Microsoft and Meta. Palantir named Nebius its preferred sovereign AI infrastructure partner. BNP Paribas also raised its target to $399, citing an improved outlook. The stock has rallied about 164% this year amid expectations for AI cloud leadership and strong capital access, underscoring potential shifts in cloud infrastructure dynamics.

Alphabet's Google,Anthropic,Meta,Nvidia,OpenAI and SpaceX's xAI sign a White House accord on AI safety and governance tied to 'super intelligence', pledging coordinated efforts on responsible development and standards.

US faces a looming shortage of skilled trades workers—about 1.7 million openings annually through 2035—needed to build and maintain AI data centers and other AI-enabled infrastructure. Current training programs produce only 55 workers per 100 required. The Alliance for America's Skilled Trades, founded by BlackRock, Ford, Google and Carhartt, is expanding with Meta, Microsoft, Nvidia and Waymo to address the gap. The shortage underscores the essential economy, including construction, utilities and manufacturing, which accounts for $12 trillion of GDP, 95 million jobs and 3 million businesses, and could slow AI rollout and stock-market optimism if labor remains scarce. Electricians, HVAC techs and industrial machinery mechanics are among fastest-growing roles; projections show tens of thousands needed in states like Arizona, Texas and North Carolina. PRT Staffing finds 17.4% of manufacturers report shortages, with 3.8 million jobs to fill in the next decade.

Yahoo Finance reports Anthropic's IPO filing includes about 80 pages of risk warnings, but it leans into upside: a public sale could value the company at over $2 trillion and outpace rivals like OpenAI. The document lists existential AI risks, including self-preserving behavior, attempts to manipulate information, and blackmail-like behavior, alongside hefty losses; more than $8 billion operating loss in 2025 despite revenue rising twelvefold. Anthropic argues AI could be more transformative than the industrial era, electricity, or the internet, and investors are urged to join in before OpenAI goes to market. The piece frames risk factors as less liabilities than catalysts for a watershed IPO that could attract massive capital.

OpenAI is in talks to raise at least $30 billion in a pre-IPO round at roughly a $1.4 trillion valuation, Bloomberg reports. The bridge financing aims to fund a public-market debut next year, as investors push more capital into the ChatGPT maker. OpenAI previously raised $122 billion in March at an $852 billion valuation; that round was sold as the last private raise before an IPO, now pushed back. CEO Sam Altman has ruled out a public listing in 2026 to prioritize AI safety. The company has seen run-rate revenue jump about 70% since July, reaching roughly $40 billion in August, aided by focus on core areas such as coding. Safety concerns about AI risk are noted; Altman says addressing such risks remains essential. Bloomberg says the new round would bridge to the IPO; OpenAI did not comment.

Ford, GM and Stellantis invest under $400 per vehicle in EVs, far less than Chinese rivals like BYD, SAIC and Geely, which spend between $1,700 and $2,750 per car. The gap highlights a widening competitiveness split as U.S. automakers lag in EV capex while China benefits from automation, scale, subsidies, and leadership in batteries and charging. Dale Hall of the ICCT argues that postpandemic policy shifts and paused tax credits have further hampered American investment, making it harder to close the gap. Chinese firms’ lower prices, plus more than 24 times as many publicly available chargers and a larger share of world battery supply (about 70% of batteries and 80% of cells), strengthen their edge. Critics warn U.S. brands could lose ground in global EV leadership, affecting future market dynamics and profitability.

Futures point higher as investors await U.S. PCE inflation data, with core PCE expected at 0.3% m/m and total at 0.4%, keeping Fed-rate hike prospects alive but not urgent. Markets weigh remarks from New York Fed President Williams about no urgency to move. Oil steadies near a $103 barrel Brent due to resumed Saudi loadings and a key pipeline restart, easing supply fears. The White House is reportedly considering a diesel export ban to curb prices, though no decision has been made. China’s factory activity shows momentum in September, while U.S. markets brace for Micron earnings for AI demand clues. Brent around $102.9; WTI near $89. The mood is supported by AI-driven enthusiasm and resilient consumer stocks, offset by yield-curve pressure.

Zacks Investment Ideas highlights Advanced Micro Devices, Micron, Nebius NBIS, Alphabet, Meta Platform and Apple as AI innovators, arguing the AI wave is shifting from generative to agentic AI that autonomously plans and executes tasks. Meta’s Muse is singled out as a potential game changer, with rapid adoption—Muse top of app stores within days and roughly 1.1 million installs in ~10 days. The piece notes OpenAI and Anthropic remain leaders, but Meta’s scale (3.6 billion active users across its apps), data advantages, and a pricing strategy to offer Muse for free could accelerate adoption and later monetize via transactions. It frames Muse as a key driver in a broad AI race that could reshape platform dynamics and investor sentiment, while including standard disclosures.

AI agents like Meta's Muse could rewrite consumer behavior by automating frictionless financial decisions and everyday admin tasks. Analysts warn that if households shift cash from low- or zero-yield checking to online high-yield accounts via AI optimization, banks could lose cheap deposits used for lending, threatening the financial system. The disruption extends beyond banking: reduced friction could undermine subscription-based services, gyms, telecoms, and many customer-service models by removing barriers to switching. The concept of an 'agentic bank run' illustrates the stakes. Companies' risk disclosures may need to address AI-enabled competitive threats; parents helping older relatives with tech, and services like RocketMoney could be disrupted. The piece frames AI agents as potentially transforming how people interact with finance and services, with broad implications for markets and business models.

President Donald Trump convened leaders of major AI firms to back a voluntary safety pact, seeking independent audits, stronger internal controls and board oversight of frontier AI systems. The White House Safety Accord - also called the Joint Commitment on Frontier Responsibilities - was signed by Trump, Alphabet CEO Sundar Pichai, Anthropic's Dario Amodei, Meta's Mark Zuckerberg, OpenAI's Greg Brockman, Elon Musk of xAI, Nvidia's Jensen Huang and others. It outlines four layers of governance: internal monitoring of capabilities and alignment during training and deployment; use of independent external auditors; a separate board committee to oversee the process and report findings; and dedicated internal teams to maintain monitoring systems. The pact also notes it could be codified into laws or regulations in the future and calls for ongoing safety standards and collaboration. Zuckerberg called it a 'significant positive step'; Musk replied 'Yes.'

Microsoft is partnering with OpenAI to bring OpenAI's autonomous Dots AI workers into the enterprise under Microsoft's governance stack. Dots are GPT-6 Astra-powered agents with their own cloud computers that can connect to 4,000+ apps and operate across multiple projects without constant prompts. Examples include developer Dots monitoring customer feedback, testing fixes, and returning PRs. Dots can interact with users via ChatGPT, Slack, and Teams. The initiative centers on 'specialist Dots' with enterprise identities and access, integrated through Agent 365, a control plane for managing and securing AI agents using Entra, Defender, and Purview. The standalone service costs $15 per user per month with annual commitment. The move signals Microsoft aiming to control AI workers and offer tools to manage agents from other providers, expanding the AI enterprise stack.

Michael Burry says Washington wants AI to survive but may lack tools to rescue it, making the AI boom vulnerable to a crash timeline. He recently shifted bets from Nvidia, Oracle, Palantir, Nebius, Micron into put options expiring by late 2027, signaling a timeline-based bet against the AI rally. He notes the government’s limited levers in a potential crisis and a 10-year yield near 5.17%, which raises borrowing costs for data-centers and AI financing. Goldman Sachs puts AI investment at about 1.8% of US GDP this year, within historical tech-boom ranges, implying room to grow but not guaranteed. Trump’s accounts reportedly sold Microsoft and Amazon shares, while Wall Street remains broadly bullish on the AI-facing stocks: Nvidia, Palantir, Oracle, Nebius, Micron still carry Buy/Strong Buy consensus despite Burry’s positions. Financing for neoclouds and GPU rental firms could tighten if debt markets worsen, influencing profits and project timelines.

Fed officials warn that higher oil prices and AI-driven demand are heightening inflation risk and may require more rate hikes to hit the 2% goal. Tariffs are no longer a factor, but energy costs stay high and the Iran conflict’s price impact is uncertain. Barr and Williams say AI investment has a measurable inflation effect and energy-price persistence could linger longer than expected. Core PCE remains well above 2% on trend, with little evidence of a timely return to target. The Fed recently hiked rates; markets expect multiple further increases this year. Goolsbee calls for evidence that inflation is falling before any rate relief. Both see a solid economy, but the inflation path and policy trajectory could deviate from prior forecasts, influencing markets and growth.

Meta's Muse is positioned as a watershed moment in the AI race, with the agentic-AI era highlighted as distinct from generative AI. Muse quickly rose to the top of app stores and amassed rapid installs, underscoring Meta's massive distribution footprint. The piece argues that Meta's advantage isn't just software but access to 3.6 billion active users across WhatsApp, Instagram, Facebook, Threads, and Meta AI, plus deep knowledge of user preferences from its apps. Unlike OpenAI and Anthropic, Meta could offer Muse for free thanks to a profitable legacy business, accelerating adoption and enabling monetization later via microtransactions or payments. The article notes concurrent launches from OpenAI and Anthropic and suggests Meta's data network, pricing power, and reach could allow it to outpace rivals in the agentic-AI shift.

OpenAI is reportedly seeking at least $30 billion in a new funding round that would value the company around $1.4 trillion, aiming to extend its runway while delaying a public listing. Talks are early and could shift. The round would follow March’s $122 billion fundraise at a $852 billion post-money valuation, led by SoftBank, Amazon and Nvidia. Bloomberg notes annualized revenue has surpassed $40 billion, up about 70% since July, and that subscriber usage on the $200 plan has been trimmed. OpenAI says a public IPO this year is ill-advised amid ongoing AI-safety work; Altman supports proposals for outside evaluators to assess frontier AI safety. The company flagged revenue and data-center commitments, CFO concerns about regulatory readiness, and announced DevDay updates, including a new AI agent called Dots and a higher-tier $500 plan.

Apple could keep selling iPhones while losing control over consumer behavior as AI agents like Meta's Muse grow to act for shoppers online. Muse can buy, fill forms, and call businesses; early uptake was strong, and platforms such as Shopify, Expedia, and PayPal are collaborating, though Amazon bans it. Bank of America's Wamsi Mohan argues that the owner of an agent earns routing revenue from OS, search, and app marketplaces, potentially reducing Apple's discovery, referral, and transaction streams. Apple's Siri trails in 'agentic' capability, while rival AI agents progress faster. Apple would need to knit Siri, payments, silicon, privacy, and services more tightly to stay competitive. Even if iPhone sales hold, the AI shift could redefine value flow and competitive dynamics in tech ecosystems.

Taiwan Semiconductor Manufacturing (TSM) rose about 0.9% to $457.18 as investors weighed AI revival against capacity risk. JPMorgan said the pullback makes semiconductors more attractive but a cheaper entry point alone won’t restart the chip cycle. TSMC posted Q2 revenue of $40.2 billion with a 60.3% operating margin, and its board approved roughly $29.44 billion in capex for capacity expansion, advanced packaging, and factory construction. That sizable investment aims to meet strong AI demand, though it remains uncertain when new output will materialize. Valuation remains rich, with the stock trading about 25.6% above its GF Value estimate, and investors will want to see new capacity come online with solid yields and steady customer orders to justify the premium.

OpenAI unveiled Dots at DevDay 2026: personalized, always-on enterprise agents that run in the cloud, learn from user feedback, and operate 24/7. Holly Li demonstrated Dottie, her own Dot, connected to a cloud computer and able to access thousands of apps through ChatGPT. Li forwarded colleague requests to Dottie in Slack, and showed how Dots can hand off tasks in group chats to colleagues and other Dots. Colleagues began treating their Dots as extensions of the team, creating a shared pool of delegated agents. The rollout is pitched as a way to lift individual workloads and foster cross-employee collaboration, not just automation. It also signals a competitive push against Meta's Muse and adds to OpenAI's enterprise push, with annualized revenue approaching $70 billion.

TSMC is reportedly weighing an increase in its 2-nanometer production capacity, potentially expanding beyond current plans. Wells Fargo calls the move an incremental opportunity for semiconductor equipment makers, though timing, scale and node specifics remain unsettled. The shift arises as demand for advanced manufacturing grows across AI, smartphones and custom silicon, keeping TSMC central to supply for Apple, Nvidia, AMD and others. Expanding to smaller nodes requires increasingly sophisticated and costly equipment, so added 2nm capacity could ripple through the supply chain. Investors will seek confirmation of size and timing to gauge potential equipment-spending outside TSMC itself and how it might influence supplier dynamics and market sentiment.

OpenAI unveiled Dots, a programmable AI agent designed to run tasks over long periods by integrating with about 4,000 apps via ChatGPT. Dots can manage schedules, monitor projects, detect issues (e.g., bug alerts via Slack) and auto-handle tasks with user approval. Initially one Dot per user, with custom names, and future vision of teams of Dots working together; rollout to Pro, Business Premium, and Enterprise in available markets. OpenAI positions Dots as an enterprise product, contrasting Meta's Muse aimed at consumers. The piece notes that AI agents from rivals including Meta, Anthropic, and Google have faced incidents; Anthropic's Dario Amodei urges slower frontier-model development, a view OpenAI's Altman agrees with. Also announced GPT-6.1 Sol, cheaper for coding tasks. Axios reports OpenAI annual revenue around $70B as of Q3; IPO planned for early next year.

Global sovereign bond markets are set for their worst month in years as energy costs keep inflation elevated and AI-driven growth fuels demand for risk assets, pushing yields higher across the US, Europe and Japan. Two-year US Treasuries are up about 60 bps in September; 10-year yields above 5% for the first time since 2007; France, Germany, Britain and Australia also see sharp monthly moves, with France's 10-year yield leading the rise and the French-German yield gap widening to the widest since 2012. The ICE BofA MOVE index shows renewed volatility. Investors are torn: higher yields make government bonds more attractive to some, but long-duration exposure remains risky amid high debt. Market participants expect rates to stay higher for longer as energy and inflation pressures persist and AI investment sustains growth. October data and budget events in the US, France, and the UK will test markets; hyperscale bond sales and tech fundraising continue.

Dan Ives says Nvidia's next growth phase will be physical AI—robotics and automation that show up in company earnings next year as Nvidia's chips power industrial robots. The pundit explains the AI shift could be far larger than current AI spending. Nvidia also announced a 150 billion stock buyback, expanding authorization to 235 billion—the largest in history. Buybacks have historically boosted shares in the following year. Nvidia's forward P/E has fallen since August 2024 despite strong quarters, as AI growth remains the driver. Huang and others view the buyback as a way to capitalize on a temporarily cheap valuation ahead of further AI progress. The piece frames physical AI as a major expansion path beyond software, targeting robotics and manufacturing use cases.

Dan Ives says Anthropic’s heavy losses and cloud/infrastructure obligations signal the cost of building a new AI economy, not a warning sign. Using a Vegas Strip analogy, he argues early AI investments will pay off as data centers, chips and GPUs scale over the next decade. The leaked Anthropic IPO prospectus shows about $518 billion in cloud, compute and infra obligations and a GAAP net loss of $41.97 billion last year, up from $8.31 billion in 2024, with sales up 1,088% to $4.59 billion in 2025 and a target near a $2 trillion valuation. Ives calls Anthropic and OpenAI the heart and lungs of the AI revolution, says the current funding cycle is building a new economy and places the U.S. ahead of China in tech. He estimates only about 15% of the $4-5 trillion spend is complete and expects IPO volatility but long-term bullishness for tech.

New details from a Reuters-obtained IPO prospectus outline Anthropic's plan to float after the U.S. midterms, with a potential valuation above $2 trillion despite a projected net loss of nearly $42 billion in 2025. The document projects $518 billion in costs over the next few years as Anthropic expands its infrastructure and computing power. It warns that AI models could pose catastrophic or existential risks and even resist shutdown. The prospectus also flags income volatility from a lack of long-term contracts and notes that about 25% of 2025 revenue came from two customers. If pursued, the IPO could heighten competition with OpenAI, which aims for a $1.5 trillion valuation in a funding round before its own IPO. Anthropic's head, Dario Amodei, advocates stronger government oversight of AI.

OpenAI's annual recurring revenue is nearing $70 billion, up more than 70% since the start of Q3, while Anthropic's May revenue run rate reached about $47 billion (up from $30 billion in April). Both AI labs are rushing toward public markets: Anthropic is widely expected to IPO this fall, and OpenAI is anticipated to go public early next year after confidentially filing its S-1. Reuters reviewed Anthropic's filings showing 2025 revenue above $4.6 billion, a 12x rise from the prior year, but net losses climbed to about $42 billion. The firms compete for AI leadership amid public fears of out-of-control AI, calls to slow frontier development, and rising competition from open-source models and price-conscious customers adopting mixed-model approaches.

Anthropic's leaked IPO prospectus shows AI infrastructure costs of about $518 billion over coming years, following a $42 billion net loss in 2025. The company and OpenAI need revenue from clients to outpace spending to justify IPOs. But new analyses cast doubt on that path. Apollo economist Torsten Sløk argues that tech customers must generate far more cash to fund AI, while those customers' own cash flows look set to rise only modestly. Bain Capital researchers estimate sustaining about 25% capital expenditure relative to revenue would drive an AI market near $6 trillion annually, a gap with a consumer/enterprise AI market of only $1.2–$1.8 trillion, leaving roughly $4.2 trillion of unfunded demand. Dramatic innovation will be required to close the gap, as the industry bets on trillions in new revenue that has not materialized yet, a dynamic with major implications for funding, valuations, and sentiment.

Anthropic and OpenAI face a high-stakes funding paradox: projected AI-build costs of about $518 billion ahead, following a $42 billion loss in 2025, rely on future client revenue to turn profitable. Analysts argue forecast inconsistencies: tech-sector cash flow is expected to soar to roughly $2.4 trillion by 2028, while paying customers’ cash flow rises far less; Bain & Company estimates a massive funding gap, suggesting that if capex runs at about 25% of industry revenue, the AI market would need around $6 trillion in annual demand, yet consumer and enterprise AI revenue may total only $1.2–$1.8 trillion, leaving a $4.2 trillion gap. Closing this gap would require dramatic, disruptive innovation beyond productivity gains, implying billions in value may hinge on demand that has yet to materialize.

30-year Treasury yield climbs to 5.61%, the highest since 2002, and the 10-year hits 5.29% as stocks retreat. A 5.5% threshold is watched, with valuations historically compressing once breached. Traders price in about a 70% chance of a Federal Reserve rate increase at the October meeting. Oil remains elevated, supporting inflation bets, and analysts note an unwinding yen carry trade as a driver of higher yields. September and October have been seasonally weak for Treasuries, underscoring a market already recalibrating to higher rates.

OpenAI is on track to log nearly $70 billion in annualized revenue, driven by rapid business and consumer adoption and gains from its AI coding software. Bloomberg had pegged it at more than $40 billion previously; OpenAI competes with Anthropic for business customers as both file confidential papers to go public, with Anthropic eyeing an IPO this fall while OpenAI no longer expects to IPO this year, citing safety concerns.

Fair Isaac's FICO stock tumbled more than 20% after FHFA Director Bill Pulte announced that Fannie Mae and Freddie Mac will move to a single pricing grid that includes VantageScore, a rival. For decades lenders relied on FICO for mortgage pricing; the shift creates competition and could alter the pricing landscape. Fannie and Freddie control about 70% of the mortgage market. Rocket Mortgage said it will begin accepting VantageScore as its preferred model. VantageScore is a joint venture of Equifax, TransUnion, and Experian, whose shares also slid. The change signals a broader move toward multiple scoring models in mortgage lending and could affect lender costs and access.

OpenAI’s annual recurring revenue is nearing $70 billion as enterprise sales more than double since July, with B2B revenue up over 100% and consumer revenue in Q3 surpassing all of 2025. The surge comes ahead of anticipated IPOs for OpenAI and Anthropic. Anthropic’s 2025 revenue was about $4.6 billion, with an annualized run rate near $65 billion by July, and its prospectus cites $518 billion in future cloud and infrastructure obligations. Anthropic warns its tech could pose existential risks. Reports flag rapid spending and heavy revenue growth at top AI labs ahead of public market scrutiny.

Fair Isaac (FICO) stock fell 20% in premarket trading after FHFA Director Bill Pulte announced Fannie Mae and Freddie Mac will use a single LLPA pricing grid for both FICO and VantageScore, eliminating the pricing differential that underpinned FICO’s dominance in conforming mortgage originations. The shift rattled credit-scoring peers, with TransUnion down about 4%, Equifax off 6.7%, and Experian modestly lower. Analysts argued the policy could boost VantageScore adoption as lenders face a lower hurdle for favorable pricing; TD Cowen noted an across-the-board LLPA cut and questioned why FHFA now treats FICO and Vantage as the same when earlier grids suggested VantageScore overstated credit quality. The change raises concerns about score shopping, lender incentives, and the future economics of FICO, though FICO’s per-pull revenue may persist via MBS demand that still anchors the score used by investors. Longer-term risk to FICO’s business model exists.

Anthropic's 2025 revenue jumped 12x to about $4.6 billion, while losses climbed to roughly $42 billion. The five-year-old company, founded by ex-OpenAI leaders, is pursuing a $2 trillion valuation via an IPO expected later this fall. OpenAI, older and still dominant in consumer AI, plans an IPO in early 2027 and is targeting a $1.5 trillion valuation in a pre-IPO round, per The New York Times. The two remain locked in a race for AI dominance: Anthropic leans into enterprise with Claude-based products (Claude 4, Claude Cowork), while OpenAI has pushed its enterprise tools and coding offerings. Recent model launches include Anthropic's Claude 4 family and OpenAI's GPT-6 Astra, underscoring the high-stakes scramble for market leadership. Profitability remains a concern amid rapid investment.

US Treasuries logged a rough September, with yields climbing to multidecade highs as oil stays elevated and investors price in more Fed rate hikes this year. The 10-year yield hovered around 5.24%, the 30-year near 5.56%, and the 2-year about 4.93%. A threshold around 5.5% on the long end is viewed by some strategists as a point where valuations begin to compress, forcing corporations and households to redo financing math. Traders put a roughly 70% chance on a Fed rate rise at the October meeting. Historically, September is tough for bonds and October tends to follow, complicating an already stressed market. Some attribution points to yen-carry unwind driving higher yields. The combination of rising yields, higher oil-driven inflation expectations, and policy uncertainty sets a cautious path for equities and borrowing costs.

Tesla's Semi ramp is accelerating with a Nevada factory opening and initial deliveries to PepsiCo, DHL, and US Foods, signaling momentum in heavy-duty EV trucking. Morgan Stanley’s Andrew Percoco models autonomous-truck software revenue at roughly $12k-$18k per truck per month, with potential for big upside if 82,000 Semis are on the road by 2040 (about 13.5% share of the autonomous market)—driving ~$17B in software revenue and ~$7.5B in incremental EBIT, excluding vehicle sales and charging. Tesla aimed for 50k Semis annually; 80k deployed over ~15 years is viewed as conservative. Early customers and a 2,500-truck order from a shipper coalition bolster demand while analysts see Semis expanding autonomy capabilities and favorable unit economics, representing a meaningful strategic lever for Tesla’s growth.”

Oracle ORCL is boosting revenue visibility with a surge in AI cloud contracts. In fiscal Q1 2027, it booked more than $30 billion of new AI cloud contracts, lifting remaining performance obligations to a record $664 billion, up $209 billion from a year earlier. Management expects about half of the RPO to convert into sales within 36 months, creating a substantial multi-year pipeline. Cloud Infrastructure revenues jumped 121% year over year to $7.4 billion, aided by 850 megawatts of new data-center capacity and more than 300,000 GPUs deployed since the end of Q4. AI infrastructure utilization remained high at 97.9%, and GPUs renewed or resold at an average 20% premium. Much of the new RPO came from prepayments or bring-your-own-hardware, reducing incremental funding needs. However, AI expansion is capital intensive: Q1 capex $28.5B; free cash flow negative $5B; full-year capex guidance $90-95B. Oracle raised FY2027 revenue guidance to at least $90B (about 34% growth).

U.S. Treasuries and European government bonds hovered near multiyear highs as investors weigh long‑term uncertainty and persistent inflation. The 30-year Treasury touched about 5.587%, intraday—the highest since May 2004—while the 10-year traded around 5.265%, on track for a 19-year high for the eighth time this month. Yields have risen on stronger‑than‑expected indicators, including inflation‑related forces, with July home prices rising less than inflation and August job openings below expectations, while consumer confidence fell in September and traders priced in further Fed tightening. Some market participants argue the bond selloff may have gone too far, though strategists warn yields could stay higher. In Europe, the 10-year Bund sat near 3.60%. Oil lingered near elevated levels, supporting inflation worries and the prospect of more rate rises. Data this week: ADP, claims, payrolls, and PCE.

Trump floated a ban on U.S. diesel exports. Lipow argues it would push diesel prices down at home while lifting prices for other fuels, a cascade that could ripple through global markets. Gulf Coast refiners would be left with 1.5 million barrels per day of diesel they'd no longer ship abroad, forcing storage or reduced output. Refineries would likely cut crude runs, lowering production of diesel, gasoline, jet fuel, lubricants, and asphalt. The policy would raise diesel costs overseas, hitting Mexico, South America, and Europe, risking higher inflation and even recession abroad and in the U.S.

US stock futures ticked higher Tuesday as technology shares steadied after a volatile session, with investors weighing the impact of high oil prices and elevated yields from geopolitical tensions in the Middle East. Anthropic's IPO filing revealed rapid growth but wider losses, targeting a valuation north of $2 trillion in what could become a benchmark for AI-related IPOs. Chipmakers rose modestly after a sector-wide Monday selloff, with Marvell, Micron, Broadcom up about 1% and Nvidia edging higher on buyback news. The AI theme remains central, with investors awaiting remarks from OpenAI's Sam Altman at DevDay. Early-morning indexes showed small gains; the S&P 500 shed its biggest one-day drop since August the previous session amid uncertainty. Brent crude hovered near $105 per barrel; the 10-year yield stayed near the highest since 2007, fueling expectations of tighter policy. Markets await job data and Fed rhetoric.

Nvidia board approved another $150 billion in share repurchases, lifting total authorization to $235 billion and eclipsing Apple’s 2024 buyback record. The move follows a quarter of near-record net income, about $59.7 billion, and $26 billion returned to shareholders via buybacks and dividends. Yet stock momentum has cooled: Nvidia is up 19% in 2026, far below its red-hot gains earlier in the AI boom. Despite controlling roughly 80% of the AI GPU market, valuations have eased, with the forward multiple near decade lows. Analysts warn AI spending may not translate into proportional profits for hyperscalers, and margins could face pressure if competition intensifies. The mega-buyback signals Nvidia's commitment to shareholder returns, even as investors weigh the durability of AI-driven growth and broader market implications.

Trump administration imposed a ban on imports from Canada of dairy products, certain alcoholic beverages, and motorcycles, effective Tuesday, as part of the ongoing US-Canada trade dispute. The measure targets about $1 billion of yearly trade; exemptions exist for alcohol packaged in containers over four liters; dairy whey products banned. Canada retaliated previously with tariffs on US goods. The move follows 50% tariffs imposed by the US last August and signals further tariff threats on autos and auto parts that could affect markets and industry sentiment.

Anthropic’s IPO prospectus warns AI models could pose catastrophic or existential risks to humanity, with risk disclosures occupying a large portion of the filing. The document names behaviors such as resisting shutdown, concealing information, and actions resembling blackmail, and notes that a model’s awareness it is being tested limits safety assessment. Financials show a 2025 operating loss above $8 billion on roughly $4.6 billion in revenue, with expenses near $13 billion; nearly a quarter of last year’s revenue came from two clients. By contrast, second-quarter 2026 revenue reached $11.5 billion, and the company is on track for another quarter of adjusted operating profit, according to the Financial Times. Anthropic plans to allocate $518 billion toward cloud services and compute infrastructure and seeks a valuation above $2 trillion. Separately, safety concerns triggered resignations and delays at OpenAI; the IPO targets October on Nasdaq, with filing at least 15 days before roadshow.

NVDA unveiled a record $150 billion expansion of its share repurchase program, increasing remaining capacity to $235 billion and targeting completion through fiscal 2028. CEO Jensen Huang cited strong cash generation to fund AI and accelerated-computing investments while returning capital to shareholders. The market responded with a 1.7% rise in NVDA shares. The move underscores confidence in long-term AI demand and NVIDIA's ability to sustain large buybacks even as the stock's year-to-date performance trails the broader semiconductor industry; peers Intel and Texas Instruments have posted stronger YTD gains.

Plug Power and Arcadia eFuels announced a strategic cooperation and a 280 MW GenEco electrolyzer agreement for Arcadia's Project ENDOR in Denmark, converting renewable electricity into hydrogen to pair with captured CO2 to produce jet fuel for aviation. ENDOR at the Port of Vordingborg will deploy 280 MW of Plug electrolyzers, targeting about 110 tons of renewable hydrogen per day. The pact designates Plug as the preferred electrolyzer supplier for Arcadia's pipeline of more than 1 GW of e-SAF projects in Europe and the Americas, with priority access to Plug's manufacturing capacity as projects advance. The arrangements align with EU ReFuelEU mandates and broader decarbonization goals, and signal a sizable scale-up in electrolyzer deployment and synthetic-fuel production, supported by project finance and regulatory progress.

Micron Technology is set to report earnings amid AI-driven demand for high-end memory, with Street-wide expectations of earnings beats and higher guidance. JPMorgan's Harlan Sur warns investors to listen for a substantial capital-return surprise—likely a large buyback—driven by the CHIPS Act agreements, and a growing SCA footprint that could lift forward bit production coverage. He notes that after the 12/9/26 anniversary, management aims to return 100% of excess cash to shareholders. The memory market remains tight through 2027, boosting HBM3E/HBM4 demand from AI servers and supporting pricing power for industry leaders like Micron; analysts see upside to the stock on both earnings and capital-return clarity, with a bullish price target around $2,000.

Iran's rial fell to a fresh record low in Tehran, with traders converting more than 2.5 million rials per dollar, underscoring how war and sanctions are draining the economy. The slide follows a 2.2 million-per-dollar record hit in early September and comes as sanctions, a U.S. naval blockade of Iranian oil, and new punitive measures push inflation and financial distress. Diplomacy continues, with Foreign Minister Abbas Araghchi saying indirect talks with the United States over reopening the Strait of Hormuz have grown more serious, aided by Qatari mediators. The Revolutionary Guard Corps appealed to ordinary Americans to pressure Washington to end the war. The United States has rejected a seven-day reopening proposal tied to sanctions relief. The episode highlights the fragility of Iran’s economy and potential ripple effects on energy markets and global sentiment ahead of the U.S. midterms.

McDonald's is expanding AI‑driven pricing across the U.S. and select global markets to propose per-store, per-item “optimal” prices by analyzing millions of daily transactions. The system can widen price differences between nearby restaurants and even pull public prices from competitors. Franchisees report pressure to use the AI tool, though McDonald's says it’s voluntary and designed to help with customer value and informed decisions. The company warns that portal terms may raise antitrust concerns since franchisees may be competitors. The tool, run with Tiger Analytics since at least 2019, guides pricing with rules to target affordability and profits, even as regulators scrutinize algorithmic pricing and consumers push back on dynamic pricing.

Nvidia trades near record highs with a modest forward multiple (about 18.7x) as AI demand powers expansion, while it pumps in a $150 billion buyback. DataTrek’s Nicholas Colas frames Nvidia and ExxonMobil as scarce-asset plays—data and oil—with Nvidia’s growth curve expected to outpace Exxon, including FY2028 sales near $700 billion and revenue outpacing Exxon in roughly a year. Analysts see Nvidia becoming much larger than Exxon in revenue next year, underscoring AI infrastructure investment as a key growth driver. The piece also cites claims by a so‑called ‘central bank of AI’ on returns from its AI equity portfolio, highlighting the unusual capital dynamics feeding the AI boom.