Latest AI (Artificial Intelligence) News
Anthropic reportedly raises $65 billion at a $965 billion valuation
Anthropic closed a massive Series H round that pushed its valuation to $965 billion, reportedly overtaking OpenAI’s $852 billion mark. The round drew heavyweight backers including Google and Amazon, signaling continued investor appetite for frontier AI labs.
Claude Opus 4.8 launches with dynamic workflow automation
Anthropic also shipped Claude Opus 4.8, adding a Dynamic Workflow feature that can plan complex tasks, spin up parallel subagents, and verify outputs. The model was highlighted for a large codebase migration benchmark, which the report says completed in 11 days with a 99.8% test pass rate.
EU reaches provisional deal on the Digital Omnibus for AI
EU negotiators reportedly struck a provisional agreement on the Digital Omnibus on AI, delaying key enforcement deadlines for high-risk AI systems to 2027 and 2028. The package also introduces targeted simplification measures and new prohibitions, indicating a shift toward staged compliance.
New Zealand issues AI guidance for regulators
New Zealand’s Ministry for Regulation released “Responsible AI in Action” guidance for the country’s 267 regulators. The guidance is framed as a practical tool for public-sector oversight as AI adoption expands across government decision-making.
Connecticut signs two AI-related bills into law
Connecticut Governor Ned Lamont signed two AI bills, including one that creates a data-broker deletion system modeled on California’s approach. The second bill extends frontier-model coverage using thresholds similar to SB 315, adding another layer of U.S. state-level AI governance.
Palo Alto Networks warns of AI-driven vulnerability escalation
New security concerns emerged after Palo Alto Networks reported that Anthropic’s Claude Mythos model could detect and exploit software flaws so effectively that the company issued two dozen security alerts in a single day. New Zealand’s National Cyber Security Centre separately warned that frontier AI models could drive a significant increase in vulnerabilities and incidents.
OpenAI model reportedly solves an 80-year-old math conjecture
The report says OpenAI’s latest reasoning model disproved Erdős problem 90, an 80-year-old conjecture in number theory. If confirmed, the result would be a notable example of frontier models contributing to advanced mathematical reasoning rather than just text generation.
Tesla completes AI5 chip design tape-out
Tesla reportedly completed the design of its AI5 chip and reached the tape-out stage, a major milestone before fabrication. That suggests continued investment in in-house silicon for AI workloads and autonomy systems.
OpenAI expands AI-driven search and advertising features
A weekly AI roundup reported that OpenAI is testing richer ad formats inside ChatGPT, including larger images, customizable calls to action, and e-commerce layouts. The company is also building audience targeting, conversion tracking, and self-serve campaign tools, which would deepen ChatGPT’s commercial role.
Google adds Preferred Sources and citation labels to AI search
Google is expanding Preferred Sources into AI Overviews and AI Mode, letting users highlight favored sites with visible “Preferred” badges. The company also introduced “Highly Cited” labels and article carousels to surface trusted journalism inside AI-generated search results.
DeepSeek cuts flagship model pricing by 75%
DeepSeek permanently reduced pricing for its V4-Pro model by 75%, continuing aggressive price competition in the AI API market. The move could lower costs for developers while intensifying pressure on other model providers to defend market share.
Google DeepMind chief warns AGI may arrive within a few years
Demis Hassabis said society may have only a few years to prepare for artificial general intelligence, with 2029 described as plausible for AGI arrival. The warning underscores growing urgency around governance, safety, and workforce adaptation.