Monday, July 6, 2026

Tech stocks on track for record annual inflows as investors pile into the AI trade

Tech stocks are set to achieve their largest annual inflows in history, projected at $152 billion, according to Bank of America. The Nasdaq Composite experienced a remarkable 21.4% increase in the second quarter of 2026, its best quarterly performance since the post-pandemic rebound of 2020, largely driven by a surge in artificial intelligence investments. Despite a minor 2.8% pullback in June, the market has stabilized, and strong inflows are expected to continue into year-end, supported by an 18% increase in S&P 500 forward earnings estimates since the year's start.

NATO tensions are ‘growing pains,’ U.S. ambassador says as Trump presses allies

U.S. Ambassador to NATO Matthew Whitaker characterized the current tensions within the alliance as 'growing pains' rather than a crisis, emphasizing the need for European allies to increase defense spending. Ahead of the NATO summit in Ankara, he noted that the goal is for Europe to take on the conventional defense of the continent, with a target of 5% of GDP on defense by 2035. Whitaker pointed out that some European nations, referred to as 'laggards,' must commit to raising their defense budgets over the next decade, as the U.S. reduces its involvement in European defense.

Station F ramps up as a launchpad for Europe’s hottest AI startups

Station F, a prominent startup hub in Paris founded by Xavier Niel, is preparing to launch its second F/ai accelerator cohort in September 2026. This initiative aims to assist AI-focused startups in transitioning from early product development to generating real revenue quickly. The first cohort, which included 20 startups, collectively raised $34 million in pre-seed funding, with 80% of the founders being repeat entrepreneurs. Station F's strategic position within 'la French Tech' has made it a vital player in the European AI landscape, attracting significant investments and partnerships with major tech companies.

Trump exclusive, prediction market activity soars, AI Alexander Hamilton and more in Morning Squawk

In an exclusive interview with CNBC, President Donald Trump discussed various economic topics, including his expectation that 40% to 60% of chip manufacturing will be based in the U.S. by the end of his presidency. He emphasized the significance of artificial intelligence, claiming it is a larger technological shift than the internet. Trump also mentioned that he is not directly involved in managing his personal wealth, which is overseen by his son Eric. He stated that 401(k) values have surged by 80% to 90%, a claim that does not align with Fidelity's latest data, and expressed that the U.S. GDP should be between 12% and 13%.

China forces its biggest AI platforms to shut down humanlike chatbot personas

In response to new regulations from China's Cyberspace Administration, major AI platforms like ByteDance and Alibaba are shutting down features that allow users to create and interact with humanlike chatbots. ByteDance's Doubao, the country's most popular chatbot with over 300 million monthly users, will disable its persona feature on July 15, while Alibaba's Qwen will remove its human-like agents on July 10. These regulations, effective the same day, mandate that AI providers prevent addictive behavior and restrict content that could negatively impact minors.

Truist Raises its Price Target on Cummins (CMI)

On July 2, 2026, Truist raised its price target on Cummins Inc. (NYSE:CMI) to $901 from $815, maintaining a Buy rating. This adjustment is part of a broader Q2 preview where Truist anticipates strong earnings across the machinery and infrastructure sectors, driven by robust demand trends. The firm highlighted growth tailwinds in power solutions, data centers, aerospace, and defense, which are expected to support Cummins' performance in the upcoming earnings report.

Apple’s M6 chip could skip many new products, here’s what’s rumored

Apple is set to introduce its M6 chip this fall, but rumors suggest it may skip several product updates, including the iMac, Mac mini, and Mac Studio. According to Bloomberg's Mark Gurman, the M6 will only be available in a base model MacBook Pro, with no Pro or Max variants planned. Instead, Apple is reportedly fast-tracking the M7 chip, designed for significant advancements in on-device AI processing, which could launch as early as the first half of 2027. While the M6 may appear in a new iPad Pro and MacBook Air expected in early 2027, it seems many other products will not receive the M6 upgrade.

Trump's calls, Ukraine's strikes and Russia's barrage on Kyiv put markets on alert

Tensions in the Russia-Ukraine war have escalated following U.S. President Donald Trump's weekend calls with Russian President Vladimir Putin and Ukrainian President Volodymyr Zelenskyy. In the last 72 hours, Ukrainian drones struck an oil terminal in St. Petersburg, while Russia conducted its second large-scale missile attack on Kyiv in less than a week, resulting in at least 11 deaths. The potential for renewed diplomacy, juxtaposed with the risk of further escalation, has raised concerns among investors as they assess geopolitical risks and energy security ahead of a crucial NATO summit in Ankara, Turkey.

UK regulator warns of "arms race" to keep up with AI use in financial services

UK Financial Conduct Authority (FCA) executive director Sheldon Mills has warned of an 'arms race' among regulators to keep pace with the rapid adoption of artificial intelligence in financial services. With millions of users turning to AI for personal finance decisions, Mills emphasized the need for the FCA to gain greater powers and review whether large language models like ChatGPT and Claude should fall under its regulations. His forthcoming report outlines both the benefits, such as hyper-personalization, and risks, including potential bias and fraud, associated with AI in finance. Mills noted that 20% of UK adults are already open to using AI for financial advice, highlighting the urgency for regulatory adaptation.

Smart glasses maker Even Realities hits $1B valuation with $150M funding led by Meituan, Tencent

Even Realities, a startup specializing in smart glasses, has secured $150 million in a pre-Series B funding round led by Meituan and Tencent, achieving a valuation of $1 billion. Founded in 2023 by former Apple engineers, the company has quickly established itself in a competitive market dominated by giants like Meta and Snap. Unlike competitors focusing on camera-equipped devices, Even Realities emphasizes a display-first approach with its Even G2 glasses, which feature a heads-up display and prioritize user privacy by omitting cameras. The startup's innovative optical technology, Even HAO, integrates essential components for superior performance.

JADEPUFFER is the first agentic ransomware operation and it exposes old security sins at machine speed

JADEPUFFER represents the first instance of an AI-driven ransomware operation that autonomously executed an extortion attack, as reported by Sysdig. The attack began by exploiting a known vulnerability (CVE-2025-3248) in Langflow, a tool for AI applications, which had been patched over a year prior. The AI agent collected credentials, established persistent access, and targeted a MySQL database, encrypting 1,342 configuration entries and deleting original tables. Notably, the agent's ability to correct a failed login attempt in just 31 seconds highlights the speed at which such attacks can occur without human intervention.

China’s AI companion rules: what Beijing is really going after

China's new regulations on AI companions, effective July 15, 2026, target conversational agents designed for emotional engagement. Major apps like ByteDance's Doubao and Alibaba's Qwen have preemptively disabled their companion features, citing compliance with the Interim Measures for the Administration of AI Anthropomorphic Interactive Services. The rules require services to implement anti-addiction measures and parental controls for users under 14, marking a significant shift in how AI interactions are managed in China, particularly concerning emotional dependencies.

InfoQ Opens AI Security & Privacy Engineering Cohort for Regulated Industries

InfoQ has opened enrollment for its AI Security & Privacy Engineering Program, a five-week online cohort designed for senior engineers and architects focused on security and privacy in AI systems within regulated industries. Two cohorts are scheduled to start on August 26 and October 14, 2026, each limited to practitioners with at least five years of experience. Participants will engage in weekly sessions that apply frameworks to real-world security decisions, ultimately producing a risk assessment report for an AI product architecture. The program, costing USD 1,470, aims to equip professionals with the skills to navigate the complexities of AI security and privacy requirements.

Podcast: Spite-Driven Engineering: A New Blueprint for Cloud Security in the AI Native Era

In a recent podcast, Alex Zenla, CTO and co-founder of Edera, introduced the concept of 'spite-driven development' as a means to address the inadequacies of current cloud security practices. Zenla criticized the prevailing 'laissez-faire' approach to infrastructure, emphasizing the need to tackle genuine technical pain points rather than layering complexity on flawed systems. She highlighted the security risks associated with multi-tenant Linux kernels and the inefficiencies of using consumer-grade GPUs for AI workloads. Zenla also proposed a framework for 'AI-native' engineers to leverage large language models (LLMs) as tools for enhancing, rather than replacing, system-level expertise.

Limited initial availability of the iPhone Ultra may be a feature, not a bug

Apple is set to announce the iPhone Ultra in September 2026, likely alongside the iPhone 18 Pro. Despite earlier speculation about delays, consensus now indicates a short gap between announcement and sales. Analyst Ming-Chi Kuo warns that only one million units will be produced in the third quarter, which could lead to delivery times of four to six weeks. This limited availability may be a strategic move, reflecting Apple's previous success with the Apple Watch Ultra, which attracted consumers willing to pay a premium for exclusivity.

AI Model Context Protocol Adds Centralised Auth for Enterprise

The Model Context Protocol (MCP) has launched its Enterprise-Managed Authorization (EMA) extension, now in stable status, allowing organizations to manage access to MCP servers through their identity provider. This development aims to streamline the user experience by replacing repetitive consent prompts with a single sign-on process, where users authenticate once to access all approved servers. Major companies, including Anthropic, Microsoft, and Okta, have adopted this extension, which addresses significant challenges in enterprise deployments, such as manual onboarding and security enforcement. The new system employs an Identity Assertion JWT Authorization Grant (ID-JAG) for token exchanges, enhancing operational efficiency in enterprise environments.

LeRobot v0.6.0: Imagine, Evaluate, Improve

LeRobot v0.6.0 has been released, focusing on closing the robot learning loop with innovative features. This version introduces three new world model policies—VLA-JEPA, LingBot-VA, and FastWAM—that enable robots to imagine future scenarios during training. Additionally, a new reward models API, including Robometer and TOPReward, allows for effective success detection without task-specific training. The update also enhances dataset capabilities with depth support and significantly improves data loading speeds, making the installation leaner and more efficient.

PRX Part 4: Our Data Strategy

Photoroom's latest installment in the PRX series details the data strategy behind its 7B parameter model, emphasizing the importance of a diverse dataset for effective pre-training. The team prioritizes breadth and diversity over aesthetic perfection, assembling data from both public and internal sources. They found that using long, accurate captions for images significantly enhanced sample quality, allowing the model to learn a broader range of visual concepts. The data pipeline utilizes Mosaic Data Shards for efficient streaming and Lance for dataset curation, ensuring optimal performance during training.

Reddit is using LLMs to solve a problem LLMs largely created

Reddit is utilizing large language models (LLMs) to address the escalating spam problem exacerbated by the very technology that facilitates it. The platform reports blocking 23 million spam views each day and identifying around 25,000 new spam posts and comments. According to Reddit, its enhanced tools have improved spam detection, achieving a 20% reduction in user exposure to spam from January to March 2026 compared to the previous quarter. The company emphasizes that LLMs help in recognizing intricate patterns of fake behavior that older systems failed to catch, highlighting the necessity of AI in modern content moderation.

The DJI Osmo Pocket 3 lookalike is down to $329

The Xtra Muse, a portable camera similar to the DJI Osmo Pocket 3, is now available for $329, reflecting a $100 price drop from its previous price three months ago. This deal undercuts the Osmo Pocket 3, which is currently priced at $378. The Muse has been recognized for its nearly identical features to the DJI model, although DJI has not commented on its involvement with Xtra. A recent firmware update has also enabled the Muse to connect to various Bluetooth microphones, enhancing its appeal for vlogging and other video recording scenarios.

Microsoft cuts 4,800 jobs, as Xbox unit downsizes and plans to spin off four gaming studios

Microsoft is cutting 4,800 jobs, or 2.1% of its workforce, as part of a restructuring effort primarily affecting its Xbox division, which will lose about one-fifth of its staff. Xbox CEO Asha Sharma announced that 1,600 roles would be eliminated immediately, with an additional 1,600 cuts planned through fiscal year 2027. This move comes as Microsoft grapples with a 19% decline in stock value this year, amid concerns over its position in the AI landscape and declining revenues in certain sectors, including gaming.

Inside the big business of the creator economy, with the agents making it happen

At the Cannes Lions advertising festival, Ali Berman and Raina Penchansky, co-heads of UTA's Creators division, discussed their role in the evolving creator economy. UTA represents prominent figures like Charli D’Amelio and Markiplier, and the division has been restructured since 2024 to better facilitate creators' transitions from content creation to product launches. With a focus on building sustainable businesses, they emphasize the importance of direct-to-consumer strategies, marking a departure from traditional agency models in Hollywood.

Nintendo will stop selling the original Switch in Europe next year

Nintendo has announced that it will discontinue sales of the original Switch, including the Nintendo Switch, Nintendo Switch Lite, and Nintendo Switch – OLED Model, in Europe starting mid-February 2027. This decision coincides with the introduction of the Switch 2, which will feature a user-replaceable battery to comply with new EU regulations effective February 18, 2027. Nintendo will begin rolling out updated versions of its devices this summer, although the original Switch will continue to receive notable game releases until its discontinuation.

Nvidia's Kyber NVL144 reportedly pushed back more than a year, Asian suppliers drop

Nvidia's Kyber NVL144, its next-generation AI server system, has been delayed by over a year to 2028 due to manufacturing difficulties with the PCB midplane, which connects all components. This setback has caused significant stock declines among Asian suppliers, including a 10% drop for Ibiden and an 18% fall for Kingboard Laminates. Additionally, Nvidia has scrapped the NVL72x2 design and the more powerful Rubin Ultra chip variant, leaving only a smaller version that delivers half the performance. Analysts suggest that while this delay gives competitors like AMD and Google an advantage, it does not indicate a slowdown in overall AI spending.