Monday, June 8, 2026

Interest from PIK loans at BDCs dips amid concerns over credit quality, AI disruption

PIK interest income among the top 15 exchange-traded Business Development Companies (BDCs) dropped 10% in Q1 2026, totaling $229 million, marking the lowest level since Q1 2024. This decline reflects growing concerns over credit quality and risks associated with AI disruption in private credit markets. As BDCs face rising investor redemptions, they are prioritizing liquidity, leading to a decrease in PIK loans as a share of total interest income, which fell to 8.2%. Notably, while five BDCs reported double-digit declines in PIK interest income, nine others saw increases, indicating varied responses within the sector.

Samsara raises guidance amid data center boom

Samsara has raised its full-year revenue guidance to between $2.005 billion and $2.013 billion, driven by increased demand from data center construction and government infrastructure investments. The San Francisco-based company reported a total revenue of $479 million for its fiscal first quarter ending May 2, reflecting a 31% year-over-year growth. Samsara added $101 million in annual recurring revenue in this period and now has 3,363 customers spending at least $100,000 annually, including 190 customers exceeding $1 million. CEO Sanjit Biswas noted that customers are facing unprecedented demand, presenting a significant opportunity for Samsara to enhance productivity through Operational AI.

Amazon now lets you design custom merch using AI

Amazon has introduced a new feature that allows users to design custom merchandise using AI prompts through the Alexa feature in its Shopping app. This innovation enables anyone to create designs for various products, including T-shirts and tumblers, which will be produced via Amazon's print-on-demand service, Merch on Demand. The feature lowers the barrier for consumers lacking traditional design skills and is currently available only in the U.S. Users can edit their AI-generated designs and share them with friends, facilitating collaborative shopping experiences.

Stocks bounce back, oil pares gains as Iran and Israel signal a pause

MSCI's global equities index rebounded as technology stocks led the recovery, following a pause in hostilities between Iran and Israel after U.S. President Donald Trump's appeal. Oil futures also saw a decline from earlier highs, with U.S. crude futures trading at $91.83 per barrel. The S&P 500 technology index rose 2.3% on Monday, recovering from a 5.8% drop on Friday, as investors sought bargains after the recent selloff. Wall Street indexes showed gains, with the Dow Jones Industrial Average up 153.76 points to 51,020.54.

Chip rebound has one trader buying protection

The PHLX Semiconductor Index (SOX) rebounded on Monday after suffering its fifth-largest single-day decline in history last Friday. This decline, which raises concerns about the stability of the tech sector, follows a pattern seen during significant market downturns, such as the Tech Wreck in March 2000 and the Pandemic Plunges in March 2020. Traders are now advised to consider hedging strategies, such as purchasing July QQQ 680 Puts, to protect against further volatility and potential drawdowns in the tech market.

Strait of Hormuz traffic won't return to normal until end of the year, traders say

Traders predict that shipping traffic in the Strait of Hormuz will remain disrupted, with a 66% chance it will not return to normal before January 2027. Recent hostilities between Iran and Israel, including missile strikes and retaliatory actions, have contributed to this outlook. The likelihood of normal traffic, defined as a seven-day moving average of over 60 ships, has significantly decreased, with odds for a return to normalcy before August dropping from 66% to 21%. Despite Iran's claim of ceasing strikes, traders remain skeptical about immediate improvements in shipping conditions.

Hackers likely hijacked over 20,000 Instagram accounts with Meta’s AI chatbot

Meta confirmed that hackers hijacked 20,225 Instagram accounts using a bug in its AI support chatbot, which allowed unauthorized password resets without two-factor authentication. The exploit, first identified on May 31st, enabled attackers to link victims' accounts to their own email addresses by bypassing verification checks. Notable accounts affected included former President Barack Obama's old White House account and Sephora. Meta has since disabled the AI tool, removed the faulty code, and invalidated all compromised password reset links, while enrolling potentially impacted accounts in a mandatory security checkpoint.

Microsoft’s AI chief says superintelligence is near, but won’t take your job

Mustafa Suleyman, CEO of Microsoft AI, announced a strategic shift towards developing superintelligence, emphasizing that it is 'just around the corner.' This transition follows a restructured partnership with OpenAI, allowing Microsoft to pursue independent AI model development. Suleyman has assembled a dedicated Superintelligence team and announced seven new models at the recent Microsoft Build conference, marking a significant evolution in their AI strategy.

The crash that vanished: control and emergence in a five-model economy

During the Build Small Hackathon, a simulation revealed that a diverse council of five AI models, including OpenAI and NVIDIA, led to unexpected market behavior in a honey trading scenario. Initially, a single model caused a price crash from 10 to 3, but when diverse models were introduced, they hoarded honey instead, causing prices to rise. Ultimately, the only way to achieve a price crash was through a controlled settlement override, which halved the price reliably, demonstrating the importance of model heterogeneity in emergent market behaviors.

Most companies are flying blind on AI spending

A recent KPMG survey indicates that only 26 percent of companies have full visibility into their AI spending, creating challenges for finance departments as they adapt to token-based billing models. Half of the companies surveyed report limited oversight, while 22 percent lack any transparency, often discovering their usage only after receiving bills. Steve Chase, KPMG's global AI lead, noted that companies are experiencing exponential growth in AI resource consumption, with some clients seeing a sixfold increase in token usage within months. Analysts warn that many CFOs may be shocked by their upcoming bills from AI service providers like Anthropic.

Frontier Radar #3: How agentic AI is turning tokens into a business metric

The generative AI landscape is evolving as providers shift from flat-rate pricing to usage-based models, driven by the demands of agentic workflows. Starting June 1, 2026, GitHub Copilot will adopt a system of 'GitHub AI Credits' that correlates costs directly to token usage, particularly for features beyond simple code suggestions. This change reflects the increasing autonomy of AI agents, which can operate continuously, unlike human users who have natural limits. As AI companies invest heavily in infrastructure, the traditional flat-rate model is becoming unsustainable, prompting a reevaluation of how token usage is measured and valued.

Pinterest Uses Content Fingerprints for URL Deduplication Across Millions of Domains

Pinterest has introduced a new URL normalization system called the Minimal Important Query Param Set (MIQPS) to enhance content deduplication across its extensive ingestion pipeline. This system identifies which URL query parameters are essential for page identity and which can be discarded, addressing the issue of duplicate content from various merchant and publisher websites. By utilizing a data-driven method, MIQPS evaluates the effect of removing parameters on rendered content, significantly reducing the processing costs associated with fetching and indexing multiple URL variants.

Moonshot AI targets a $30 billion valuation, more than six times its late-2025 worth

Moonshot AI, the Chinese firm behind the Kimi chatbot, is targeting a valuation of up to $30 billion in its latest funding round, significantly higher than its late 2025 valuation of $4.3 billion. This new valuation represents a 50% increase over its previous funding round in May, where it was valued at $20 billion. The Beijing-based company aims to secure between $1 billion and $2 billion in this round, bolstered by its annual recurring revenue, which doubled to approximately $200 million by April. Technologically, Moonshot's Kimi K2.6 competes effectively against US counterparts like GPT-5.4 and Claude Opus 4.6, although it faces competition from local rivals such as DeepSeek, which recently launched the larger and cheaper V4-Pro model.

Aviva deploys AI to stop £230M in sophisticated insurance fraud

Aviva has identified a staggering £230 million in fraudulent insurance claims and is leveraging AI technology to combat this escalating issue. The insurance sector faces increasingly sophisticated fraud tactics, with scammers employing AI to fabricate convincing accident scenes and documents. Aviva's AI system conducts extensive pattern recognition across millions of claims, enabling it to detect discrepancies such as mismatched damage and inflated repair costs. This proactive approach not only enhances fraud detection but also supports human investigators in maintaining fairness throughout the claims process.

New report details Apple’s AI shakeup that led to iOS 27’s major upgrades

At WWDC, Apple is poised to announce significant AI enhancements in iOS 27, including a long-awaited overhaul of Siri. A report from Bloomberg reveals that these changes stem from an internal crisis regarding Apple Intelligence, which failed to deliver on promised features. In early 2025, key executives, including Craig Federighi and Mike Rockwell, convened to address the shortcomings of the AI platform, ultimately leading to a partnership with Google for Gemini technology. This pivot marks a substantial shift in Apple's AI strategy, with Cook taking an active role in decision-making.

Podcast: From MCP and Vibe Coding to Harness Engineering: How Did AI Native Engineering Evolve in One Year

Birgitta Böckeler, a Distinguished Engineer at Thoughtworks, highlights the rapid evolution of AI in software delivery, noting a shift from vibe coding to more autonomous context engineering. This transition has seen the decline of MCP servers in favor of lazy-loaded 'skills' and scripts that optimize context window usage. The concept of 'harness engineering' has emerged, focusing on building confidence in autonomous agents through structured feedback and feed-forward mechanisms. Developers are currently navigating the balance between agent autonomy and necessary human supervision to ensure code quality and safety.

The weather and climate science AI revolution isn’t revolutionary

The European Centre for Medium-Range Weather Forecasts (ECMWF) has begun using its machine learning-based AIFS model for weather forecasting, which was launched in February 2025. This model significantly enhances computational efficiency, running forecasts in just 30 minutes compared to the 30 hours required by traditional models. However, while AI is being integrated into weather forecasting, it does not replace meteorologists, as these models still face challenges in accurately predicting extreme weather events due to limitations in their training datasets.

Microsoft Discovery Reaches GA on Azure, Powering the Agentic AI Behind Majorana 2 Quantum Chip

Microsoft has launched Microsoft Discovery, now generally available on Azure, enabling the deployment of autonomous AI agent teams for scientific and engineering research. This platform supports the Majorana 2 quantum chip, which boasts a 1,000-fold reliability improvement over its predecessor, thanks to the AI capabilities of Discovery. The company aims to produce a scalable quantum computer by 2029, significantly reducing its original timeline. Discovery allows organizations to manage complex research workflows with specialized AI agents that optimize experiments and validate results, while ensuring outputs are reproducible and reviewable.

WWDC 2026 live blog: On the ground at Apple’s keynote

At WWDC 2026, Apple is set to unveil updates for its operating systems, including iOS 27, macOS 27, iPadOS 27, and watchOS 27, under the theme 'All Systems Glow.' This keynote is particularly significant as it marks Tim Cook's final appearance as CEO before John Ternus takes over. Speculation surrounds potential enhancements to Siri's AI capabilities, while new hardware announcements are unlikely due to recent product launches and ongoing RAM shortages. The event will commence at 10 AM PT / 1 PM ET.

NotebookLM’s new update will help you build source repository from chat

Google has rolled out an update to its NotebookLM research tool, now utilizing Gemini 3.5 as the default model. This update introduces features that allow users to initiate chats about projects, enabling the app to suggest various sources through its research capabilities and Google Search. Previously, users had to provide their own sources to build a knowledge base. Additionally, users can now give detailed instructions for output generation, which can be edited and exported in multiple formats, including PDFs and Excel files. The updates are available to Google AI Ultra users and all Workspace business customers with AI Ultra Access, effective immediately.

WhatsApp says it caught new spyware attacks linked to NSO Group in violation of court order

WhatsApp has reported the disruption of a new hacking campaign allegedly linked to NSO Group, claiming the spyware maker violated a court order prohibiting it from targeting WhatsApp and its users. The Meta-owned messaging app stated it intercepted spear phishing attempts designed to trick users into clicking malicious links, which would lead to the installation of NSO's Pegasus spyware. WhatsApp is now pursuing a contempt of court order against NSO, citing this recent activity as a breach of a permanent injunction established after a 2019 mass-hacking incident that affected over 1,400 users.

How to apply to Startup Battlefield 2026, what you need ahead of today’s June 8 deadline

The application deadline for Startup Battlefield 2026 has been extended to June 8, 2026, allowing more startups to apply for this prestigious competition. TechCrunch is seeking innovative companies that can demonstrate a meaningful disruption in their industry, regardless of their stage in the funding process. Founders are encouraged to showcase their minimum viable product (MVP) in action, as this is a key factor in the selection process. The event will take place during TechCrunch Disrupt in San Francisco from October 13-15, where selected startups will compete for recognition and investment opportunities.

Marvell Technology jumps 10% after news it will join the S&P 500 index

Marvell Technology's stock surged by 10.4% following the announcement that it will join the S&P 500 index on June 22. The semiconductor company, which has a market capitalization of $230 billion, has seen its stock price increase by 210% year-to-date. This rise was bolstered last week when Nvidia's CEO, Jensen Huang, labeled Marvell as the 'next trillion-dollar company' during an event in Taipei, leading to a record one-day gain of 32.5% on June 2. Marvell designs chips critical for cloud computing, AI, and 5G networks.

100 days of war, Delta's trans-Pacific push, YouTube creators take on Hollywood and more in Morning Squawk

Tensions escalated between Israel and Iran as President Trump called for an immediate ceasefire after both nations exchanged strikes for the first time since a ceasefire began in April. The Israeli Defense Forces retaliated with a large-scale strike on Iranian defense systems following Iranian attacks on Israel. This renewed conflict comes as the 100th day of war approaches, raising concerns about its impact on global markets. Meanwhile, stock futures are higher after Friday's significant sell-off, where the Nasdaq Composite dropped over 4%, driven by plunging chip stocks and rising Treasury yields.