OpenAI Trims Its Fundraising Goal as Apple Enters China

OpenAI is reportedly seeking a smaller amount in its latest funding round, while Apple says Apple Intelligence will reach Chinese iPhones in December. Elsewhere, a compromised Anthropic coding extension, a federal crackdown on AI-assisted student-loan fraud, and an AI agent’s mouse-brain experiment added consequential developments to the day. A proposed industry standard for opting out of AI scraping and a lawsuit over xAI data-center pollution also merit attention.

OpenAI Reportedly Cuts Its Latest Fundraising Target

OpenAI is now seeking up to $20 billion at a reported $1.25 trillion valuation, according to CNBC. The round remains in progress, and a person familiar with the deal told CNBC that the figures could still change.

The reported target is lower than the larger figure previously associated with the round. Bloomberg attributed the change, as reported by the Financial Times, to an AI-stock selloff and the approaching public listing of rival Anthropic. The company has not yet closed the financing, so its final size and valuation remain unresolved.

Apple Intelligence Is Set for a China Launch

Apple says the first Apple Intelligence services will roll out to Chinese iPhone users with iOS 27.3 in December, following an open beta for developers beginning November 4, according to the South China Morning Post. The launch follows a period of regulatory uncertainty and involves Alibaba.

The initial release is narrower than a full Siri rollout. The Verge reports that it will include features such as notification summaries, translation, writing tools, visual intelligence, and photo object removal, while the new Siri AI will not be included in China at launch.

A Claude Coding Extension Was Backdoored on Cursor

A malicious version of Anthropic’s Claude Code extension was available on Cursor’s marketplace for roughly 11 hours, according to Fast Company. The tampered extension was designed to collect information from users’ computers and attempt to establish persistent access through SSH.

Ars Technica reported that the affected version was distributed through Anthropic’s official Cursor account and could have reached at most 5,000 users. Cursor removed it, rebuilt its marketplace, and said it found no evidence that other extensions were compromised. The companies have not explained how attackers gained control of the account.

DOJ Charges Five in AI-Assisted “Ghost Student” Fraud Case

The US Department of Justice has charged five people in what Ars Technica described as the first broad federal crackdown on “ghost students.” The alleged scheme used stolen identities to create fake college students, obtain loans, and use AI in part to sustain the appearance of attendance.

The DOJ said data analysis helped identify likely fraudsters and that authorities seized more than $750,000 in cryptocurrency from a network associated with Kenyan nationals. The case remains an enforcement action and does not establish the allegations against the accused as proven.

An AI Agent Chose a Gene Therapy for a Live Mouse

A biotech startup allowed an AI agent to select a gene therapy and operate a motorized rig that injected the treatment into a live mouse’s brain, Wired reported. The experiment offers a concrete example of an agent making a treatment choice inside a laboratory procedure rather than merely analyzing results.

The accompanying account from Tom’s Hardware says Anthropic’s Sonnet controlled the surgery rig after other model variants refused or warned that they might stop during the experiment.

Major AI Companies Back a Shared Scraping Opt-Out Signal

Google, OpenAI, Anthropic, and Microsoft say they will support a common signal that publishers can place in robots.txt to express preferences about AI training scrapes, according to The Verge. The proposed “policy token” would let publishers communicate those preferences across multiple AI companies instead of configuring them separately.

Engadget described the effort as a pledge to support a new standard. The signal is proposed as part of work connected to the Internet Engineering Task Force, so adoption and practical enforcement remain to be established.

Judge Lets NAACP Pollution Suit Against xAI Proceed

A judge has allowed an NAACP lawsuit concerning pollution from an xAI data center to move forward, TechCrunch reported. The suit accuses the SpaceX-owned AI company of violating the Clean Air Act.

The ruling is a procedural development rather than a final finding on the allegations. Further proceedings will determine how the claims develop.

The Financing, Launches, and Cases Still Need Follow-Up

OpenAI’s final financing terms remain unsettled, while Apple’s December China rollout will show which features are available in practice. Investigators and users will also be watching for more disclosure about the compromised Cursor extension, and the DOJ and NAACP cases will proceed through their respective legal processes.