Anthropic's Mythos model sparks safety concerns as US government agencies partner with AI companies

The Trump administration is drafting an executive order to mandate that US government agencies partner with AI companies to bolster cybersecurity against cyber attacks. Officials David Sacks and Michael Kratsios are leading efforts to create a unified federal framework, building on actions from January and December 2025. However, the current draft notably omits mandatory approval tests for new AI models, aiming to update existing programs without adding strict regulatory hurdles for emerging technology.

Despite this focus on security partnerships, the White House faces internal pressure regarding policy direction. The release of Anthropic's Mythos model has sparked safety concerns, leading some officials like National Economic Council Director Kevin Hassett to suggest an FDA-like approval process. This proposal contradicts the administration's usual deregulatory stance, causing concern among tech industry leaders who fear a shift in messaging despite official claims of balancing innovation with safety.

Meanwhile, major tech giants are navigating rising costs and market shifts. Tencent and Alibaba are seeing growth slow to the low-teen percentage range as they double their AI investments to stay competitive. In the hardware sector, Wall Street investors are rotating capital from top firms like NVIDIA toward second-tier companies, with Micron Technology's market cap surpassing $800 billion. Analysts warn that this semiconductor rally resembles the dot-com bubble of 1999, citing comparisons from investors like Michael Burry and Paul Tudor Jones.

Consumer tech is also facing scrutiny, as Apple settled a class action lawsuit for $250 million over claims that it falsely advertised AI features for iPhone 16 and select iPhone 15 models. The settlement covers purchases made between June 2024 and March 2025, with eligible customers potentially receiving up to $95 per device. Conversely, Jim Cramer highlights Alphabet, Microsoft, and Amazon as critical players in the AI cloud infrastructure market, noting their essential role in supporting AI models for 2026.

Smaller players and academic institutions are also making significant moves. Miivo Holdings launched an AI lead finder tool to help businesses scale sales efforts, while the University of Michigan's endowment is poised to gain $2 billion from its early $20 million stake in OpenAI. In the financial sector, researchers introduced Strat-LLM, a framework allowing large language models to trade stocks using real-time data, and ZyAlpha unveiled a new AI crypto trading system to help investors navigate the 24/7 global market.

Key Takeaways

['The Trump administration plans an executive order requiring US agencies to partner with AI companies for cybersecurity, though it omits mandatory model approval tests.', "National Economic Council Director Kevin Hassett is pushing for an FDA-like approval process for AI models, contradicting the administration's typical deregulatory approach.", "Anthropic's release of the Mythos model has intensified safety concerns within the White House, complicating the push for a light-touch AI policy.", 'Tencent and Alibaba are expected to see earnings growth slow to the low-teen percentage range as they double their AI investments.', 'Wall Street investors are shifting capital from NVIDIA to second-tier hardware firms like Micron, whose market cap has surpassed $800 billion.', 'Analysts and investors including Michael Burry and Paul Tudor Jones warn that the current semiconductor rally resembles the dot-com bubble of 1999.', 'Apple agreed to a $250 million settlement regarding false advertising of AI features on iPhone 16 and certain iPhone 15 models.', 'Jim Cramer identifies Alphabet, Microsoft, and Amazon as key winners in the AI interface and cloud services market for 2026.', "The University of Michigan's endowment is expected to gain $2 billion from its early $20 million investment in OpenAI.", 'Researchers developed Strat-LLM, a framework enabling large language models to trade stocks using real-time market data and news.']

Trump Admin Plans AI Security Order for US Agencies

The Trump administration is preparing an executive order to help US government agencies fight cyber attacks using artificial intelligence. This new rule will require agencies to work closely with AI companies to improve network security. The order builds on previous actions taken in January and December 2025 to boost American leadership in AI technology. Officials like David Sacks and Michael Kratsios are tasked with creating a federal framework to manage state-level AI laws. The goal is to create a unified approach to AI safety and security across the country.

US AI Security Order Draft Omits Mandatory Model Tests

The Trump administration plans to order US agencies to partner with AI companies to protect networks from cyber threats. However, the draft executive order does not require government approval for cutting-edge AI models. People familiar with the matter say the directive aims to update existing cybersecurity programs to include AI companies. The focus is on addressing threats posed by the emerging technology without adding strict approval steps for new models.

White House Shifts AI Policy Amid Growing Safety Fears

The White House is struggling to find a clear direction for AI policy as new powerful models raise safety concerns. The Trump administration originally promoted a light-touch approach to prioritize innovation and competition. However, the release of Anthropic's Mythos model has sparked worries about security vulnerabilities. Reports suggest National Economic Council Director Kevin Hassett might push for an FDA-like approval process for AI models. This idea has caused panic in the tech industry because it contradicts the administration's usual deregulatory stance. Officials say there is no shifting messaging and that they aim to balance innovation with security.

Tencent and Alibaba Growth Slows as AI Costs Rise

Tencent's full-year earnings growth is expected to slow to the low-teen percentage range. This slowdown comes as the company doubles its investments in artificial intelligence. Rising costs associated with developing and running AI systems are putting pressure on these major tech firms. The increased spending is necessary to stay competitive but is impacting their overall financial growth rates.

Wall Street Shifts to Second-Tier Hardware Firms Amid Bubble Risks

Wall Street investors are moving capital from top AI hardware companies to second-tier firms as bubble risks rise. While NVIDIA has been the biggest winner since ChatGPT launched in late 2022, analysts see a power shift in the AI space. Memory chips and CPUs are recovering demand as data centers need a broader range of components. Micron Technology's market cap surpassed $800 billion, and Bank of America expects the data center CPU market to grow from $27 billion in 2025 to over $60 billion by 2030. However, some analysts warn the semiconductor rally resembles the dot-com bubble of 1999. Michael Burry and Paul Tudor Jones have compared the current frenzy to that era, warning of potential dramatic corrections if valuations inflate too rapidly.

ZyAlpha Launches AI Crypto Trading System for Investors

ZyAlpha has launched a new AI-powered cryptocurrency quantitative trading system to help investors navigate the global crypto market. The platform uses machine learning algorithms to analyze vast amounts of market data and identify patterns for better investment decisions. Since crypto trading operates 24/7, manual trading often suffers from emotional decisions and slow execution. This automated system helps reduce emotional bias and improves risk management for users. Investors can now monitor the market in real-time and make timely decisions with greater confidence and precision.

New Framework Helps LLMs Trade Stocks with Real-Time Signals

Researchers have proposed Strat-LLM, a new framework for using large language models in stock trading. This system uses Stratified Strategy Alignment to combine architectural reasoning with consistent trading strategies. It operates in a live-forward setting using real-time data like prices, news, and annual reports to avoid bias. Tests on A-share and US markets show that reasoning-heavy models work best in free mode, while strict rules help reduce losses during market downturns. The study found that mid-scale models perform better under strict constraints, whereas ultra-large models can suffer from alignment issues but gain performance in guided modes.

Apple Settles for $250 Million Over Overhyped iPhone AI Claims

Apple agreed to a $250 million class action settlement to resolve claims that it falsely advertised AI features for certain iPhone models. Plaintiffs say Apple marketed the iPhone 16 lineup and some iPhone 15 models as having enhanced Siri capabilities that did not exist at the time of purchase. The settlement covers consumers who bought these phones between June 10, 2024, and March 29, 2025. Eligible customers could receive about $25 per device, with the amount potentially rising to $95 depending on the number of claims filed. Apple plans to provide additional Siri features through future software updates at no extra cost. Apple denied wrongdoing, arguing its marketing was not misleading because it disclosed that features would roll out over time.

Jim Cramer Highlights Alphabet and Cloud Giants in AI Market

Jim Cramer recently highlighted Alphabet and other major companies as key players in the AI interface and cloud services market. He noted that cloud infrastructure is essential for running AI models. Cramer pointed to Google Cloud, Microsoft Azure, and Amazon Web Services as critical components for this industry. He emphasized that companies like Alphabet, Microsoft, and Amazon are all developing AI models and the necessary cloud infrastructure to support them. This focus on cloud services is seen as a major takeaway for investors looking at AI winners for 2026.

Miivo Holdings Unveils New AI Lead Finder Tool

Miivo Holdings has launched a new AI-driven lead finder tool to help businesses identify qualified prospects more efficiently. CEO Alex Damouni said the tool was created to solve the time-consuming task of finding contacts and creating outreach campaigns. The platform curates public information to provide verified emails, phone numbers, and business details. It also generates AI-assisted messaging prompts tailored to individual business profiles. This technology allows companies to scale their sales efforts without significantly increasing staffing costs. The launch is part of Miivo's strategy to expand its AI-powered business solutions for small and medium-sized enterprises.

University of Michigan Endowment Gains $2 Billion from Early OpenAI Stake

The University of Michigan's endowment is expected to gain $2 billion from its early investment in OpenAI. The university invested $20 million when the company was in its early days, and reports estimate the stake is now worth $2 billion ahead of a likely public offering. Other early investors included Khosla Ventures and the Aphorism Foundation, which also put in $50 million around the same time. Elon Musk testified that he created the idea and recruited key people for the charity. This windfall would be historic for the university, which has a total endowment of $20 billion. The news has sparked enthusiasm on social media, though the university has not yet commented officially.

Private Credit Market Faces Two Distinct AI Risks

The private credit market is facing two different risks related to artificial intelligence: loans to legacy software companies and GPU-backed infrastructure debt. Loans to legacy software companies involve lending to firms with existing products that may struggle to adapt to new market conditions. These are generally considered lower-risk investments. In contrast, GPU-backed infrastructure debt involves lending to companies building data centers that rely on graphics processing units for AI and machine learning. This type of debt is considered higher-risk because the infrastructure may not yet be generating revenue. Investors need to understand these differences to make informed decisions.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI Artificial Intelligence Trump Administration Executive Order Cybersecurity Network Security AI Safety AI Security Federal Framework State-Level AI Laws Unified Approach AI Companies Partnerships Cyber Threats Cutting-Edge AI Models Approval Steps Innovation Competition Security Vulnerabilities FDA-like Approval Process Tech Industry Deregulatory Stance Tencent Alibaba AI Costs Investments Financial Growth Rates Wall Street Hardware Firms Bubble Risks NVIDIA Micron Technology Data Center CPU Market AI Crypto Trading System ZyAlpha Machine Learning Quantitative Trading LLMs Stock Trading Strat-LLM Stratified Strategy Alignment Large Language Models Real-Time Signals Apple iPhone AI Claims Class Action Settlement Siri Capabilities Jim Cramer Alphabet Cloud Giants AI Interface Cloud Services Google Cloud Microsoft Azure Amazon Web Services Miivo Holdings AI Lead Finder Tool Business Solutions University of Michigan OpenAI Early Investment Private Credit Market AI Risks Legacy Software Companies GPU-Backed Infrastructure Debt

Comments

Loading...