Chinese financial regulators have asked some banks not to classify overdue loans to China Vanke as non-performing and to extend repayment deadlines for the state-backed developer, in one of Beijing’s strongest interventions yet to prevent a default that could further unsettle the country’s financial system.

People familiar with the matter told Reuters that regulators had issued informal “window guidance” to lenders, urging them to support Vanke and avoid actions that could intensify its cash squeeze. Some banks were also asked to defer collecting interest payments owed by the developer, although the sources did not specify the size of the waivers.

The guidance was directed mainly at larger banks, many of which are state-owned, according to the people who spoke on condition of anonymity because of the sensitivity of the discussions.

The regulators have not set a fixed period for how long Vanke’s overdue loans can remain outside the non-performing category. One source said the treatment would depend on conditions in China’s property market and further discussions with Vanke.

The intervention underscores the extent to which authorities are trying to prevent Vanke from becoming another major casualty of China’s prolonged property downturn. Rather than allowing missed payments to immediately translate into formal loan defaults and potentially force banks to recognize losses, the guidance gives lenders room to extend maturities and delay some collections while Vanke works through its liquidity problems.

The approach is considered necessary especially after Evergrande’s liquidation.

The approach also highlights the importance authorities attach to Vanke’s links to the state. The developer has total assets of close to 1 trillion yuan ($149.31 billion) and is one of the few major distressed developers to have avoided an official default throughout the property crisis.

Vanke has received financing support from Shenzhen Metro, its state-owned major shareholder, and has extended some yuan bond payments to avoid defaults. Banks had also agreed earlier this year to defer interest payments owed by the developer.

Record Losses Expose Vanke’s Worsening Finances

The pressure on Vanke has intensified as weak property sales continue to erode its ability to generate cash. The developer reported a record 88.6 billion yuan loss for 2025, reflecting the severity of the downturn in its core business. Its financial position deteriorated further in the first half of this year, when net loss widened to 14.95 billion yuan.

Revenue fell 33% from a year earlier to 70.2 billion yuan during the first six months, adding pressure to a balance sheet already carrying substantial debt.

At the end of June, Vanke reported total debt of 351 billion yuan. Bank loans accounted for 72% of that amount, while bonds represented 7%. The remaining 21% consisted of other borrowings, including loans from Shenzhen Metro, according to Vanke’s interim report released in August.

Shenzhen Metro provided Vanke with a 22 billion yuan credit facility last year and extended additional loans this year, giving the developer another source of liquidity as commercial financing conditions remain difficult.

The latest regulatory guidance effectively gives banks an incentive to keep working with Vanke rather than aggressively pursuing repayment at a time when the developer’s ability to raise cash is constrained.

For the banks, however, postponing recognition of non-performing loans does not eliminate the underlying credit risk. It instead gives Vanke more time to stabilize its finances and potentially benefit from an improvement in property-market conditions.

The approach is considered necessary because the guidance leaves the duration of the loan treatment open. If property sales fail to recover and Vanke remains unable to meet its obligations, banks could eventually face the same losses they are currently being asked to defer recognizing.

Vanke Becomes A Test of Beijing’s Property Rescue Efforts

China’s property sector has been under pressure for more than five years after a government campaign to curb excessive borrowing by developers triggered a broader liquidity crisis.

The sector was once one of the country’s most important engines of economic growth, supporting construction, employment, household wealth and demand across a wide range of industries. But home prices continue to fall, property investment remains weak, and efforts to establish a sustained recovery have so far struggled to gain traction.

Beijing has recently sought to curb presales of unfinished homes in an effort to restore confidence among buyers. Those measures have yet to produce a meaningful turnaround in home prices, leaving developers dependent on increasingly extensive financial and policy support.

Vanke is at the center of the issue because it has remained closely tied to the state while avoiding the formal defaults that have defined much of the property crisis. A disorderly failure by the developer could therefore carry implications beyond its own creditors, particularly if it weakened confidence in other developers, banks or state-linked companies.

That helps explain why regulators are focusing on preventing a “risk event” at Vanke, according to one source.

The immediate objective is not necessarily to erase Vanke’s liabilities but to prevent its liquidity problems from turning into a destabilizing event. By asking banks to extend loans, delay interest collection and postpone non-performing classifications, authorities are effectively buying time for the developer and the wider property market.

The longer-term question is whether that time can translate into a genuine recovery.

Vanke’s worsening losses, declining revenue and 351 billion yuan debt burden show that the company’s problem is not simply a temporary mismatch between payments and cash flow. Its underlying business remains exposed to a property market where sales and investment have yet to recover.

Therefore, the regulatory intervention offers Vanke another financial lifeline, but it also illustrates the increasingly difficult balance Beijing faces: preventing a major developer from defaulting without allowing temporary relief to substitute indefinitely for a recovery in the property market.

Meta’s new personal AI assistant Muse has been testing a “human concierge” system in which contractors quietly place some phone calls on behalf of users, exposing a significant gap between the company’s vision of autonomous AI and the human labor still needed to make some of its most ambitious features work reliably.

The test was disclosed to Meta employees last week, shortly after the company publicly launched Muse’s phone-calling capability, according to internal company posts reviewed by Reuters.

Under the system, Muse could hand a user’s request to a trained human contractor, who would place the call and handle the interaction with the business. The employee posts said users could ask Muse to perform tasks such as booking appointments, checking product availability or obtaining quotes from contractors.

The arrangement immediately raised privacy concerns inside Meta. Employees warned that contractors working in call centers could potentially hear sensitive information during conversations that users believed were being handled by an AI system.

A vice president in Meta’s SuperIntelligence Labs acknowledged in an internal post that it “was a miss” to begin testing contractor-placed calls without proper disclosure. The company subsequently “rolled back this feature” for the time being, she said.

Meta said the experiment was designed to gather feedback and improve the system before any broader release.

A Meta spokesperson said employee feedback had been “overwhelmingly positive” and that the purpose of the test was to develop safety and privacy protections.

“We’re working with merchants to continue improving this potential calling feature, and will only roll it out when it’s ready and with the proper disclosures,” Meta spokesperson Daniel Roberts said.

But the development reveals a fundamental challenge facing the current generation of AI agents. Companies are increasingly presenting these systems as digital workers capable of carrying out tasks independently, but the reliability needed for real-world interactions can still require human intervention.

In Meta’s case, the internal testing suggested that human involvement could dramatically improve performance. The vice president said some experiments showed human-placed calls achieving success rates of between 95% and 98%, compared with a lower success rate for calls handled entirely by AI.

That difference helps explain why human contractors remain useful even as AI companies market agents as autonomous.

Phone calls are particularly difficult for AI systems because they involve unpredictable conversations, accents, interruptions, ambiguous requests, and situations in which the person on the other end may not know they are speaking to an AI system. A human can adapt immediately when a conversation moves outside the agent’s expected workflow.

The problem is that the use of humans changes the nature of the product being offered.

A consumer who asks an AI agent to negotiate a bill or make an appointment may reasonably assume that the interaction is being conducted by software. If a human contractor takes over without clear disclosure, the privacy expectations surrounding the service are different.

That issue is especially significant for Muse because Meta has positioned privacy and security as central to the product.

At launch, Meta said each Muse agent would operate inside its own secure virtual machine, essentially a dedicated computer in the cloud designed to isolate a user’s session. The company also said sensitive information such as passwords would be kept in separate secure storage.

The calling feature was designed to extend that model into the physical world. Users can instruct Muse to call U.S. businesses and handle routine tasks, after which the system provides a summary and transcript of the conversation through the app.

The human-concierge experiment shows that securing the computing environment is only part of the privacy challenge. Once an AI agent needs to interact with people outside its digital environment, protecting information also depends on how much human intervention sits behind the system.

The development also recalls Meta’s earlier attempt to build an AI assistant.

About a decade ago, Facebook tested a digital assistant called M through its Messenger service. Media reports later estimated that humans performed around 70% of the tasks assigned to the system. The difference today is the scale and sophistication of the technology. Modern AI agents can perform far more tasks autonomously, while Meta has access to billions of potential users through Facebook, Instagram and WhatsApp.

Muse has already gained significant traction. The app has accumulated more than 2.5 million downloads in the U.S. since its launch and has topped American app download charts for two weeks, according to data from Sensor Tower.

The product is central to CEO Mark Zuckerberg’s broader push to deliver what he has called “personal superintelligence” to Meta’s enormous user base. Muse can perform tasks including sending emails, shopping and booking travel, potentially shifting AI from a tool that generates information to an agent that takes actions on behalf of its user.

That transition makes reliability more important.

An inaccurate answer from a chatbot can usually be ignored or corrected. An AI agent that makes a phone call, cancels a service, negotiates a contract, or makes a purchase can create consequences outside the chatbot itself.

Meta began testing the phone-calling feature internally in August before gradually expanding access after Muse’s public launch. The human concierge system was then enabled for about half of Meta’s employees last week, according to an internal announcement.

Employees who did not want to participate could join an opt-out group.

However, the experiment highlights a difficult trade-off in building autonomous agents for Meta. Human intervention can make an AI system considerably more effective, but it can also undermine the perception that the service is genuinely autonomous and introduce new privacy and disclosure obligations.

The company has now pulled back the human-concierge feature, at least temporarily. Meta says it intends to continue improving AI calling and will only release the capability more broadly with appropriate disclosures.

The episode may prove less important for Muse’s immediate rollout than for the wider economics of agentic AI. As companies promise software that can act independently in the real world, the hidden question is how much human work is required to make those promises reliable. If agents still need people to step in whenever interactions become difficult, the technology may be autonomous in appearance while remaining partly human-operated underneath.

The Trump administration is backing a key part of Apple’s legal argument in its Supreme Court fight with “Fortnite” maker Epic Games, urging the justices to impose a stricter standard for holding companies in civil contempt when they allegedly violate court orders.

The Justice Department told the Supreme Court that civil contempt should generally be available only when an injunction clearly covers the conduct at issue, a position that could strengthen Apple’s challenge to a contempt ruling stemming from its long-running dispute with Epic. The government stressed, however, that it was supporting neither side in the case and said Apple’s conduct concerning at least one part of the lower court’s order could support a contempt finding.

The case, Apple Inc. v. Epic Games Inc., No. 25-1311, is now before the Supreme Court after the justices agreed in June to hear Apple’s challenge. The court’s docket shows that Apple filed its opening merits brief on September 14, with Epic’s response due in November.

At the center of the dispute is an injunction issued after Epic sued Apple in 2020, accusing the iPhone maker of unlawfully restricting how developers distribute apps and process payments.

Epic deliberately violated Apple’s App Store rules in 2020 by introducing its own payment system in “Fortnite,” triggering Apple’s removal of the game and the ensuing antitrust litigation. A federal judge subsequently ordered Apple to change aspects of its App Store practices, including rules governing how developers could direct consumers to payment options outside Apple’s system.

Epic later accused Apple of attempting to circumvent that injunction by imposing a 27% commission on certain purchases made outside the App Store.

In April 2025, the lower court held Apple in civil contempt, finding that its actions violated the injunction. Apple has denied that it violated the order and argues that the court effectively punished it for conduct that the injunction did not expressly prohibit.

The Supreme Court’s decision to take up the case therefore goes beyond Apple’s particular commission structure. It could clarify how courts across the country should determine when a company has violated an injunction sufficiently clearly to warrant civil contempt.

The Justice Department’s brief focuses on that broader legal question. It cited the Supreme Court’s 2019 decision in Taggart v. Lorenzen, which said civil contempt is appropriate only when there is “no fair ground of doubt” about whether an order prohibited the conduct in question. The government argues that standard requires courts to focus on the express terms of an injunction rather than imposing contempt based primarily on what a judge believes the order was intended to accomplish.

Apple is making a similar argument in its own Supreme Court filing, describing the Ninth Circuit’s approach as an impermissible “spirit-based” standard. Apple argues that parties need clear notice of what an injunction prohibits and that contempt cannot be imposed simply because conduct appears inconsistent with the broader purpose of an order.

That, analysts believe, could have consequences well beyond the App Store.

Civil injunctions are commonly used to regulate corporate conduct in antitrust, intellectual-property, consumer-protection and other cases. A ruling requiring courts to rely more strictly on the text of an order could make it harder to impose contempt sanctions when a company adopts a new business practice that was not expressly addressed in the original injunction.

For Apple, that question has become relevant because the company’s App Store business has been subjected to regulatory and judicial scrutiny in multiple jurisdictions as governments and courts examine its control over app distribution and payments.

Epic, meanwhile, has continued to argue that Apple’s approach effectively allows the company to evade restrictions by changing the form of conduct rather than its underlying effect. Epic CEO Tim Sweeney said earlier this month that Apple had been “evading court rulings and regulatory decisions for years” and that Epic wanted to bring those practices to an end.

The Supreme Court case also illustrates the unusual alignment between the administration and Apple on a legal principle, even though the Justice Department says it is not endorsing Apple’s broader position in the dispute.

The government explicitly acknowledged that Apple’s conduct relating to at least one portion of the lower court’s order could justify contempt. Its position is therefore not that Apple should automatically prevail, but that the legal test used to determine contempt should be clarified and applied according to the language of the injunction.

That makes the case potentially important for both sides of the dispute. Apple is seeking to overturn the contempt finding, while the Justice Department is asking the Supreme Court to establish a clearer boundary around the contempt power.

The court’s docket indicates that the case is moving through the merits stage, with Apple’s opening brief already filed. The Supreme Court has limited its review to the first question presented in Apple’s petition, which concerns the standard for civil contempt.

For the broader technology industry, the outcome could determine how much latitude companies have when operating under injunctions that impose behavioral restrictions but do not spell out every possible future business practice.

That issue is becoming increasingly relevant as technology companies continually redesign products, pricing models, and distribution systems in response to regulation and litigation. A court order written for one version of a business may remain in force while the underlying technology and commercial practices change.

Therefore, the Apple-Epic dispute has evolved from a fight over App Store payments into a case about the limits of judicial enforcement. The Supreme Court is being asked to decide how clearly a company must be warned before a court can punish it for violating an injunction, a question that could shape corporate litigation well beyond Apple’s dispute with Epic.