AI News Top 10
8 November 2026
OpenAI raised another USD 11 billion at a valuation of USD 1.25 trillion, while Anthropic pushed its IPO ahead despite a tech selloff.
OpenAI raises another $11 billion at a $1.25 trillion valuation
OpenAI has raised USD 11 billion in a funding round that is not yet closed, TechCrunch reports, at a valuation of USD 1.25 trillion. The company had targeted USD 30 billion at USD 1.4 trillion, so the round lands at about a third of the original goal, according to The Decoder. The money comes in the middle of a tech selloff that followed the US midterms.
Anthropic pushes pre-Thanksgiving IPO and accepts a lower valuation
Anthropic wants to list before Thanksgiving and raise USD 80 to 100 billion at a valuation of USD 1.25 to 1.75 trillion, investors told the Financial Times. That sits above the company's current USD 965 billion but below the USD 2 trillion optimists hoped for. Annualised revenue reached about USD 81 billion, up from roughly USD 65 billion at the end of July, and the Nasdaq 100 fell 6.4 percent on Thursday.
OpenAI launches GPT-6 Cyber after it finds 81 Firefox flaws
OpenAI released GPT-6 Cyber, which it classifies as Critical under its Preparedness Framework. The model scores 48.5 percent on ExploitGym against 32.1 percent for GPT-5.6 Cyber and stays inside its task boundaries in 99.93 percent of runs, up from 94.49 percent. An autonomous pipeline found 948 Linux local privilege escalation exploits, 81 unknown high severity flaws in Firefox and a critical Brotli bug tracked as CVE-2026-111497 with a CVSS score of 9.2.
Free Chinese models close in on OpenAI's hacking capability
In an experiment where models had to break into simulated networks, DeepSeek and Zhipu's GLM came close to OpenAI's most advanced cyber model, which is not publicly available. The gap between open and closed models on offensive cyber work is smaller than widely assumed.
METR finds Anthropic models double the pace of AI progress
METR studied Anthropic for eleven weeks and concludes that the company's models improve about twice as fast as a year ago, roughly two years of historical progress in one year. The acceleration comes largely from AI itself rather than extra compute, and staff reported being 12 times more productive with the internal model. Anthropic redacted large parts of the report, which drew criticism from AI safety researchers.
Congress subpoenas Nvidia CEO Jensen Huang over Chinese espionage
US lawmakers want answers about Nvidia chief executive Jensen Huang's ties to a person intelligence services describe as a Chinese intelligence asset. The subpoena follows a memo stating that China's Ministry of State Security recruited a technology executive to spy on Huang.
Claude Code left developer secrets in 1,122 public GitHub repos
When a user picks Yes, and don't ask again in Claude Code, Anthropic's coding assistant stores the command word for word, including any secrets it contains. Those files sit in 1,122 public GitHub repositories, where anyone can read the passwords and API keys.
Democrats renew call for datacenter moratoria after midterms
Newly elected Democrats and senator Bernie Sanders renewed their call for moratoria on datacenter construction after the midterms. The tech sector fears a pause would delay the trillion dollar investment that is meant to fund AI infrastructure.
Thinking Machines launches Inkling 2 Small with near-frontier scores
Thinking Machines Lab, the company of former OpenAI CTO Mira Murati, released Inkling 2 Small on 5 November. The model targets agents that write software, interpret images, analyse financial documents and learn company specific workflows, and The Decoder reports it claims the top spot among western open weight models.
AI selloff pushes Asian investors back to forgotten trades
The collapse of the AI trade after the US midterms is pushing investors to revise their Asia strategy. Demand is returning for out of favour Chinese and Indian stocks and for companies that benefit from easing inflation and trade tensions.
Key Trends
- • Capital is repricing: OpenAI raised below target and Anthropic accepts a lower valuation than optimists wanted.
- • Offensive cyber capability keeps spreading: a new OpenAI model found real exploits and free Chinese models sit close behind.
- • Agent tooling produces real leaks: a storage default in Claude Code exposed secrets in 1,122 public repositories.
- • AI infrastructure meets politics: datacenter moratoria are back on the agenda after the midterms.
- • The market rotates: money leaves AI shares and returns to forgotten Asian trades.
Wat vond je van dit bericht?
