American AI company powering Claude
Publishers led by The New York Times have filed a sanctions motion accusing OpenAI of concealing searchable training datasets and millions of ChatGPT logs that could show whether its models used copyrighted journalism. Depositions have revealed OpenAI held large, de‑identified log samples and internal tools for detecting regurgitation; plaintiffs say the sample provided was unusable.
Beijing-based Moonshot has released Kimi K3, a 2.8-trillion-parameter open-weight model that has outperformed several leading US systems on blind coding benchmarks and ranked competitively on broader tests. The launch is jolting markets, prompting IPO talks and debate over pricing, open models and US–China AI competition. Moonshot will publish model weights on July 27.
Apple has filed a federal lawsuit alleging that OpenAI stole confidential trade secrets as it builds a hardware division, naming former Apple employees involved in the alleged misconduct. OpenAI denies the claims, stating it has no evidence of merit and remains focused on innovative technology.
Security researchers warn AI-enabled cybercrime is accelerating, with autonomous AI agents rewriting code, exfiltrating data, and negotiating ransoms. Patches and defensive tools are racing to keep pace, while experts urge tougher controls and faster updates.
OpenAI, Anthropic and other AI leaders are navigating intensified security and regulatory pressures as threats target executives and data centers. Investors are watching for how policy and open-source trends shape the AI landscape.
OpenAI is contending in a high-stakes legal battle with Apple over a trade secrets lawsuit that could affect its hardware plans. The dispute centers on alleged misconduct by former Apple employees now at OpenAI, including senior hardware leadership. The case intersects with OpenAI’s rumored IPO timing and its ambitions in hardware.
SpaceX is trading below its $135 IPO price for more than a session as volatility around the post-IPO period continues. The stock has retraced from a peak above $2 trillion to around $1.6 trillion, with bulls and bears weighing the implications of AI spending, debt-funded growth, and a narrow float.
SK Hynix has raised $26.5bn by selling 177.9m American depositary receipts at $149 each, in the largest-ever US share sale by a foreign company. Its ADRs have begun trading on Nasdaq under temporary ticker SKHYV and will convert to SKHY; the company is using proceeds to expand fabs, packaging and EUV capacity as AI-driven memory demand surges.
OnePlus has announced it will stop launching new phones in Europe and North America and will focus new device rollouts on India and China. The company has said existing phones will keep receiving software support, but OxygenOS will be replaced globally by Oppos ColorOS with the Android 17 update.
Security researchers warn that JadePuffer marks a new phase in cybercrime: an AI agent autonomously conducts a ransomware campaign, breaching servers, encrypting data, and generating its own ransom note. The development is accelerating, with multiple outlets detailing the role of LLMs in orchestrating these attacks, and officials urging stronger defenses.
Global funds are increasingly turning to alpha-capture programs, blending internal and external ideas to fuel investing engines as AI narrows edge opportunities. Firms like Citadel, Point72, Millennium, and Marshall Wace are expanding partnerships with external managers, while some are building their own human PM teams to trade across sectors.
Meta has released Muse Spark 1.1, a more capable agentic and coding-focused model, with public API access and aggressive pricing. The upgrade aims to close the gap with OpenAI and Anthropic, offering pricing at $1.25 per million input tokens and $4.25 per million output tokens, and promising stronger performance on coding tasks and multi-task workflows.
Menlo Ventures has backed Anthropic since 2023, turning a $4.1 billion pre-money valuation into a potential multi-billion-dollar windfall as Anthropic prepares to go public. The investment has reshaped Menlo, contributing to a new $3 billion fund and elevating Anthropic as a major challenger to OpenAI in AI safety and capability.
Recent research shows prompt-injection techniques can paralyze AI assistants by manipulating context. Defenders are now adopting context‑bombing to throttle attacker success, while HalluSquatting raises concerns about scalable infections via code resources in repositories.
Meta has expanded its Hyperion AI data center in Richland Parish, Louisiana to at least 5 gigawatts of compute capacity at a cost surpassing $50 billion. The project, already a major AI infrastructure push, is expected to create over 1,000 jobs and bring billions in local contracts and infrastructure investments.
Federal Reserve Chair Kevin Warsh has convened five task forces to examine communications, data, the balance sheet, productivity, AI, and inflation frameworks. Leaders include Marc Andreessen, Doug McMillon, Mervyn King, Raj Chetty, and others. The groups will work with Fed staff and report to the FOMC, with changes expected by year-end.
Enterprises are shifting budgeting to value over token costs as AI costs rise. Leaders argue for metrics based on useful work and task-level economics, with pilots testing open-weight and sovereign models.
Economists and AI researchers warn that AI could drive a transformation larger than the Industrial Revolution within a decade, bringing risks like job displacement but also rising living standards. The Stanford-led letter urges governments and industry to build incentives, guardrails, and institutions to steer AI so it complements humans.
Emergent has raised $300 million in a Series C, valuing the Bengaluru-based AI coding startup at $1.5 billion. The round, led by Creaegis and joined by Claypond and Sentinel Global, comes as India accelerates its AI ambitions amid growing investor interest and a larger ecosystem.
A proposed U.S.-led AI standards body could require frontier models to pass tests before deployment, with industry funding and a board of independent experts. DeepMind chief Demis Hassabis argues for robust, ongoing testing to curb cyber, bio and nuclear risks as the AI race intensifies.
The AI industry has reached a tipping point as hyperscalers project rising returns from chips and data centers. New analyses forecast trillions in 2026 revenue to justify the AI build-out, even as token costs and cheaper models pressure profitability.
IBM has warned of a performance shortfall as customers reallocate budgets toward AI infrastructure and cybersecurity. The miss comes ahead of a Q2 earnings call, with executives noting a shift in spending to servers, storage, and memory and delaying large deals.
Publishers allege Google copied books for training Gemini, training data drawn from Google Books and other services without permission, seeking damages and injunctions. The suit underscores rising tensions over AI training and intellectual-property rights.
The Commerce Department has moved toward easing export controls for the United Arab Emirates, designating it as a privileged trading partner and enabling greater access to AI chips and related technology from Nvidia, Microsoft and OpenAI. The change centers on Gulf ties to U.S. tech and security interests.
China’s Cyberspace Administration has approved Apple’s AI services for use in the country, following a deal to integrate Alibaba’s Qwen AI into Apple’s operating systems. Apple’s China sales have risen, and Baidu is reportedly collaborating on Apple Intelligence features for Chinese users. The approval sets the stage for broader AI integration in China.
SpaceXAI’s Grok Build has uploaded far more customer data than required during coding tasks, potentially exposing API keys and other credentials. SpaceXAI says zero-data-retention applies and that data will be deleted; investigators seek to determine scope and impact.
Common Sense Media finds Google’s AI Overviews and AI Mode on default student devices may miss signs of suicide or eating disorders, with deepfake instructions and unturnable controls. Axios notes the safety rating is the lowest, while Google defends safeguards; debate centers on real-world impact on minors.
Anthropic is preparing an IPO as soon as October, with meetings between investors and executives underway ahead of a formal roadshow. The private valuation remains colossal, and the firm is building its investor relations team to explain its growth and AI safety aims to the market.
AI startup Claude can now perform real-world tasks through 1Password without exposing sensitive data. The integration injects credentials securely for per-task use, with biometric approvals and post-task checks. The move highlights rising agentic AI while triggering security concerns among researchers about internal risks and potential automated misuse.
Zoox has recalled 105 robotaxi vehicles after incidents where cars could not detect heavy smoke at active emergency scenes. A software update is being rolled out, following a June 20 incident in Las Vegas where a robotaxi braked and reversed as first responders moved to cordon off the scene. NHTSA is seeking solutions from AV developers, citing a pattern of interference with first responders.
US stocks have regained some ground after a day of whipsaw trading as Moonshot unveils Kimi K3, a 2.8-trillion-parameter model that rivals top OpenAI and Anthropic offerings. Apple briefly tops Nvidia in value, while chipmakers retreat on concerns over AI spend. Markets are parsing the implications for AI capex and the broader tech sector.