OpenAI adds $11 billion in funding at $1.25 trillion valuation

OpenAI adds $11 billion in funding at $1.25 trillion valuation

OpenAI has secured $11 billion in a new funding round that values the company at $1.25 trillion, according to Bloomberg sources cited by TechCrunch on 6 November 2026. The round is still open and may close in the coming weeks, with a potential total of about $20 billion. The financing comes After the firm postponed its planned IPO to next year¹.

The $11 billion contribution is part of a larger, as‑yet‑unfinished round that Bloomberg expects to close soon, possibly reaching $20 billion. Earlier reports in late September said OpenAI was aiming for $30 billion, so the current amount falls short of that target. The announcement follows market uncertainty linked to a Democratic landslide in the U.S. midterms and voter concerns about data‑center regulation, which have rattled AI investors¹.

OpenAI’s valuation of $1.25 trillion places it among the world’s most valuable AI firms, but the market is volatile after recent U.S. elections and regulatory worries. Competitor Anthropic is also seeing valuation pressure, with potential buyers citing a lower $1.1 trillion estimate versus earlier $2 trillion expectations. These dynamics illustrate how political and regulatory signals can affect funding and IPO timing for AI companies¹.

Megaspeed executive allegedly recruited to spy on Nvidia CEO

A U.S. intelligence memo reported that Alice Huang, an executive at Singapore‑based Megaspeed, was recruited by China’s Ministry of State Security to gather information on Nvidia CEO Jensen Huang. The memo was first reported by The New York Times and later referenced by TechCrunch on 7 November 2026. The claim includes an alleged romantic relationship between the two executives, which Nvidia described as “not unusual”².

The memo, circulated earlier in 2026, says Alice Huang was approached by Chinese intelligence to monitor Jensen Huang’s activities. The New York Times noted the memo was based on U.S. intelligence sources, while the Financial Times added the romance allegation. Nvidia’s spokesperson told the NYT that the company does not share classified information with customers and that Jensen Huang’s interactions with Alice Huang were routine. The memo also referenced a separate report about New Matrix, a Chinese firm authorized to purchase advanced Nvidia chips, whose listed office appeared to be a “hollow gray shell,” raising suspicion of a front operation².

Anthropic’s AI accelerates its own research

Anthropic’s internal AI models are improving about twice as fast as they were a year earlier, according to a report by the safety group METR that examined the company over an 11‑week period. The investigation, published on 7 November 2026, found that benchmark scores indicate roughly two years of progress were achieved in the last year. METR’s authors say the speedup is not merely from larger compute but from the AI tools themselves, which employees use to boost productivity³.

Employees who used the most capable internal model reported being about 12 times more productive on average, based on informal Slack polls and interviews. The report also notes that Anthropic was allowed to redact large portions of the findings, suggesting limited public visibility into the exact pace and risks of this self‑accelerating development³.

Self‑improving AI refers to systems that help design or train newer, more capable versions of themselves, reducing the need for human engineering effort. This concept has raised concerns among researchers because it could lead to rapid capability gains without adequate oversight. Anthropic’s CEO Dario Amodei has publicly urged the industry to slow down AI development, even as the company’s own tools appear to be speeding up its research pipeline³.

Microsoft AI Frontiers group disbanded after reorg

In early September 2026 the AI Frontiers sub‑organization within Microsoft Research was merged into the larger Microsoft AI unit, effectively ending the dedicated group. The team had been created in 2023 to give more structure to AI research After the release of ChatGPT. It consisted of roughly 50 researchers focused on building agents and new model capabilities⁴.

AI Frontiers aimed to coordinate research rather than rely on isolated efforts, allowing ideas like the Autogen orchestration framework and the Fara computer‑use agent to be developed more quickly. The group also pursued novel capabilities such as the Next Latent objective for extrapolation, Dion3 for optimized Muon creation, and the Full Bandwidth Transformer that uses recursion to boost model power. A shared monorepo was created to give all Microsoft Research teams access to a proven edge‑scale pre‑training pipeline, a resource that previously required substantial engineering effort⁴.

Microsoft Research historically operated as many independent labs, which sometimes slowed progress because single researchers could not assemble strong baselines alone. The AI Frontiers reorganization was intended to concentrate effort and share resources, but differences in philosophy between the former Inflection‑led Microsoft AI and the smaller AI Frontiers team created practical challenges After the merger⁴.

Meta pauses AI training terms after backlash

Meta announced an indefinite pause on its planned commercial‑terms update that would have required Meta Business account users to allow extensive AI training on external sites linked from Facebook. The pause follows weeks of public, legal and regulatory pushback, including lawsuits in Brazil and Norway and a formal UK competition investigation. Meta said it will keep the existing terms while it continues stakeholder engagement⁵.

The original update, slated for October 30, would have forced businesses to let Meta crawl any linked external pages and even create test credentials to bypass paywalls, effectively removing any opt‑out for publishers. Legal analysts warned the vague language could let Meta ignore robots.txt preferences and scrape millions of articles, potentially avoiding licensing fees⁵.

Publishers rely on Facebook pages for traffic, so the terms would have affected virtually all news outlets⁵.