All News (Page 392)

AI company Anthropic is developing and has begun testing with early access customers a new AI model more capable than any it has released previously, the company said, following a data leak that revealed the model’s existence. 

An Anthropic spokesperson said the new model represented “a step change” in AI performance and was “the most capable we’ve built to date.” The company said the model is currently being trialed by “early access customers.”

Descriptions of the model were inadvertently stored in a publicly-accessible data cache and were reviewed by Fortune.

A draft blog post that was available in an unsecured and publicly-searchable data store prior to Thursday evening said the new model is called “Claude Mythos” and that the company believes it poses unprecedented cybersecurity risks.

The same cache of unsecured, publicly discoverable documents revealed details of a planned, invite-only CEO summit in Europe that is part of the company’s drive to sell its AI models to large corporate customers. 

The AI lab left the material, including what appeared to be a draft blog post announcing a new model, in an unsecured, public data lake, according to documents separately located and reviewed by Roy Paz, a senior AI security researcher at LayerX Security, a computer and network security company, and Alexandre Pauwels, a cybersecurity researcher at the University of Cambridge. 

In total, there appeared to be close to 3,000 assets linked to Anthropic’s blog that had not been published previously on the company’s news or research sites that were nonetheless publicly-accessible in this data cache, according to Pauwels, who Fortune asked to assess and review the material.

After being informed of the data leak by Fortune on Thursday, Anthropic removed the public’s ability to search the data store and retrieve documents from it.

In a statement provided to Fortune, Anthropic acknowledged that a “human error” in the configuration of its content management system led the draft blog post to being accessible. It described the unpublished material that was left in an unsecured and publicly-searchable data store as “early drafts of content considered for publication.”

As well as referring to Mythos, the draft blog post also discussed a new tier of AI models that it says will be called “Capybara”. In the document, Anthropic says: “’Capybara’ is a new name for a new tier of model: larger and more intelligent than our Opus models—which were, until now, our most powerful.” Capybara and Mythos appear to refer to the same underlying model.

Currently, Anthropic markets each of its models in three different sizes: the largest and most capable model versions are branded Opus, while a slightly faster and cheaper, but less capable, versions are branded Sonnet, and the smallest, cheapest, and fastest are called Haiku. However, in the blog post, Anthropic describes Capybara as a new tier of model that is even larger and more capable than Opus, but also more expensive.

“Compared to our previous best model, Claude Opus 4.6, Capybara gets dramatically higher scores on tests of software coding, academic reasoning, and cybersecurity, among others,” the company said in the blog.

The document also said the company had completed training “Claude Mythos,” which the draft blog post described as “by far the most powerful AI model we’ve ever developed.”

In response to questions about the draft blog post, the company acknowledged training and testing a new model. “We’re developing a general purpose model with meaningful advances in reasoning, coding, and cybersecurity,” an Anthropic spokesperson said. “Given the strength of its capabilities, we’re being deliberate about how we release it. As is standard practice across the industry, we’re working with a small group of early access customers to test the model. We consider this model a step change and the most capable we’ve built to date.”

The document Fortune and the cybersecurity experts reviewed consists of structured data for a webpage, complete with headings and a publication date, suggesting it forms part of a planned product launch. It outlines a cautious rollout strategy for the model, beginning with a small group of early-access users. The draft blog notes that the model is expensive to run and not yet ready for general release.

Significant new cybersecurity risks

The new AI model poses significant cybersecurity risks, according to the leaked document. 

“In preparing to release Claude Capybara, we want to act with extra caution and understand the risks it poses—even beyond what we learn in our own testing. In particular, we want to understand the model’s potential near-term risks in the realm of cybersecurity—and share the results to help cyber defenders prepare,” the document said.

Anthropic appears to be especially worried about the model’s cybersecurity implications, noting that the system is “currently far ahead of any other AI model in cyber capabilities” and “it presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders.” In other words, Anthropic is concerned that hackers could use the model to run large-scale cyberattacks.

The company said in the draft blog that because of this risk, its plan for the model’s release would focus on cyber defenders: “We’re releasing it in early access to organizations, giving them a head start in improving the robustness of their codebases against the impending wave of AI-driven exploits.”

The latest generation of frontier models from both Anthropic and OpenAI have crossed a threshold that the companies say poses new cybersecurity risks. In February, when OpenAI released GPT-5.3-Codex, the company said it was the first model it had classified as “high capability” for cybersecurity-related tasks under its Preparedness Framework—and the first it had directly trained to identify software vulnerabilities. 

Anthropic, meanwhile, navigated similar risks with its Opus 4.6, released the same week. The model demonstrated an ability to surface previously unknown vulnerabilities in production codebases, a capability that the company acknowledged was dual-use, meaning that it could both help hackers as well as help cybersecurity defenders find and close vulnerabilities in code.

The company has also reported that hacking groups, including those linked to the Chinese government, have attempted to exploit Claude in real-world cyberattacks. In one documented case, Anthropic discovered that a Chinese state-sponsored group had already been running a coordinated campaign using Claude Code to infiltrate roughly 30 organizations—including tech companies, financial institutions, and government agencies—before the company detected it. Over the following ten days, Anthropic investigated the full scope of the operation, banned the accounts involved, and notified affected organizations.

An exclusive executive retreat

The leak of not-yet-public information appears to stem from an error on the part of users of the company’s content management system (CMS), which is the software used to publish the company’s public blog, according to cybersecurity professionals. 

Digital assets created using the content management system are set to public by default and typically assigned a publicly accessible URL when uploaded—unless the user explicitly changes a setting so that these assets are kept private. As a result, a large cache of images, PDF files, and audio files seem to have been published erroneously to an unsecured and publicly-accessible URL via the off-the-shelf content management system.

Anthropic acknowledged in a statement to Fortune that “an issue with one of our external CMS tools led to draft content being accessible.” It attributed this issue to “human error.” 

Many of the documents appeared to be discarded or unused assets for past blog posts like images, banners, and logos. However, several appeared to be what were meant to be private or internal documents. For example, one asset has a title that described an employee’s “parental leave.” 

The documents also included a PDF containing information about an upcoming, invite-only retreat for the CEOs of European companies being held in the U.K., and which Anthropic CEO Dario Amodei will attend. Names of the other attendees are not listed, but are described as Europe’s most influential business leaders.

The two-day retreat is described as an “intimate gathering” to engage in “thoughtful conversation” at an 18th-century manor-turned-hotel-and-spa in the English countryside. The document says that attendees will hear from lawmakers and policymakers about how businesses are adopting AI and experience unreleased Claude capabilities.

An Anthropic spokesperson told Fortune the event “is part of an ongoing series of events we’ve hosted over the past year. We look forward to hosting European business leaders to discuss the future of AI.”

This story was originally featured on Fortune.com

AI company Anthropic has inadvertently revealed details of an upcoming model release, an exclusive CEO event, and other internal data, including images and PDFs, in what appears to be a significant security lapse. 

The not-yet-public information was made accessible via the company’s content management system (CMS), which is used by Anthropic to publish information to sections of the company’s website.

In total, there appeared to be close to 3,000 assets linked to Anthropic’s blog that had not previously been published to the company’s public-facing news or research sites that were nonetheless publicly-accessible in this data cache, according to Alexandre Pauwels, a cybersecurity researcher at the University of Cambridge, who Fortune asked to assess and review the material.

After Fortune informed Anthropic of the issue on Thursday, the company took steps to secure the data so that it was no longer publicly-accessible.

Prior to taking these measures, Anthropic stored all the content for its website—such as blog posts, images, and documents—in a central system that was accessible without a login. Anyone with technical knowledge could send requests to that public-facing system, asking it to return information about the files it contains.

While some of this content had not been published to Anthropic’s website, the underlying system would still return the digital assets it was storing to anyone who knew how to ask. This means unpublished material—including draft pages and internal assets—could be accessed directly.

The issue appears to stem from how the content management system (CMS) used by Anthropic works. All assets—such as logos, graphics, or research papers—that were uploaded to the central data store were public by default, unless explicitly set as private. The company appeared to have forgotten to restrict access to some documents that were not supposed to be public, resulting in the large cache of files being available in the company’s public data lake, cybersecurity professionals who analyzed the data told Fortune. Several of the company’s assets also had public browser addresses. 

“An issue with one of our external CMS tools led to draft content being accessible,” an Anthropic spokesperson told Fortune. The spokesperson attributed the issue to “human error in the CMS configuration.”

There have been several high-profile cases lately of technology companies experiencing technical faults and snafus due to problems with AI-generated code or with AI agents. But Anthropic, which makes the popular Claude AI models and has boasted of automating much of its own internal software development using Claude-based AI coding agents, said AI was not at fault in this case.

The issue with its CMS was “unrelated to Claude, Cowork, or any Anthropic AI tools,” the Anthropic spokesperson said.

The company also sought to downplay the significance of some of the material that had been left unsecured. “These materials were early drafts of content considered for publication and did not involve our core infrastructure, AI systems, customer data, or security architecture,” the spokesperson said.

While many of the documents appear to be discarded or unused assets for past blog posts, like images, banners, and logos, some of the data appeared to detail sensitive information. 

The documents include details of upcoming product announcements, including information about an unreleased AI model that Anthropic said in the documents is the most capable model it has yet trained.

After being contacted by Fortune, the company acknowledged that is developing and testing with early access customers a new model that it said represented a “step change” in AI capabilities, with significantly better performance in “reasoning, coding, and cybersecurity” than prior Anthropic models.

The publicly-accessible data also included information about an upcoming, invite-only retreat for the CEOs of large European companies being held in the U.K. that Anthropic CEO Dario Amodei is scheduled to attend. An Anthropic spokesperson said the retreat was “part of an ongoing series of events we’ve hosted over the past year” and the company was “developing a general-purpose model with meaningful advances in reasoning, coding, and cybersecurity.”

Among the documents were also images that appear to be for internal use, including one image with a title that describes an employee’s “parental leave.” 

It’s not the first time a tech company has inadvertently exposed internal or pre-release assets by leaving them publicly accessible before official announcements.

Apple has twice leaked information through its own website—once in 2018, when upcoming iPhone names appeared in a publicly accessible sitemap file hours before launch, and again in late 2025, when a developer discovered that Apple had shipped its redesigned App Store with debugging files left active, making the site’s entire internal code readable to anyone with a browser.

Gaming companies like Epic Games and Nintendo have also seen pre-release images, in-game assets, and other media leak via content delivery network systems (CDNs) or staging servers, similar to the data lake Anthropic used in this case. Even larger firms such as Google have accidentally exposed internal documentation at public URLs, and data associated with Tesla vehicles has been exposed through misconfigured third‑party servers.

However, the problem is likely exacerbated by AI coding tools now readily available on the market—including Anthropic’s own Claude Code.  

These tools can automate crawling, pattern detection, and correlation of publicly accessible assets, making it far easier to discover this kind of content and lower the barriers to entry for doing so. AI tools like Claude Code or Codex can also generate scripts or queries that scan entire datasets, rapidly identifying patterns or file naming conventions that a human might miss. 

This story was originally featured on Fortune.com

Global conflicts are awakening governments—and investors—to the importance of modernizing military forces, says Brandon Tseng, the cofounder and president of ShieldAI, a maker of AI-powered drones that announced Thursday that it had raised $1.5 billion in Series G funding at a $12.7 billion valuation.

The new funding more than doubles the valuation of the San Diego, Calif. defense startup, which was founded in 2015 and was previously valued at $5.6 billion by investors. Shield AI is projecting more than 80% revenue growth by the end of 2026, Shield AI cofounder Brandon Tseng and CFO Kingsley Afemikhe told Fortune. That would equate to at least $540 million in revenue this year, based on Shield AI’s 2025 revenue figures.

“We don’t expect growth to slow down,”  Tseng said in an interview. 

Shield’s Series G round—co-led by first-time investors Advent International and JPMorganChase’s Security and Resiliency Initiative—is happening in tandem with two major financial moves: the pending acquisition of the tactical simulation company Aechalon and a non-dilutive $500 million fixed return preferred equity financing deal with Blackstone. The infusion of capital will fund the acquisition, as well as help Shield scale its Hivemind autonomy platform and its V-BAT surveillance drone. The funds will also support the development of a new combat drone that is preparing for its first flight by the end of this year.

The deal underscores how global conflict is reshaping venture priorities in Silicon Valley, as a crop of young companies including Anduril and Allen Control Systems produce new tech-driven products designed for a changing battlefield. Shield AI has gained traction from the deployment of its systems in Ukraine, where its V-BAT surveillance drone has been used in active operations.

Tseng told Fortune that fundraising discussions began in November, prior to the U.S. military capturing Venezuelan President Nicolás Maduro or the recent strikes in Iran. Tseng says investor sentiment has shifted alongside “a broad observation” that the world has become less stable.

“Countries around the world are modernizing their militaries, and obviously the U.S. has pushed for an increase in defense spend among all of its allies and partners,” Tseng said. “That certainly is in the background as investors think about investing in defense.”

Tseng declined to say whether Shield AI’s V-BAT drones have been deployed in Iran, but noted the company operates “in almost every single conflict zone.”

Advent Chairman David Mussafer is joining Shield AI’s board as part of the funding deal, while investor Todd Combs, of JPMorgan Chase, will serve as a board observer. Aechelon cofounder and CEO Nacho Sanz-Pastor will continue to lead the business unit and oversee its integration with Hivemind, reporting to Shield AI CEO Gary Steele, according to Tseng.

Shield’s projection of more than 80% revenue this year does not include the acquisition of Aechelon.  

The final close of Shield AI’s funding round will also be contingent on approval of the Aechelon acquisition. If the deal fails to clear regulatory hurdles, Tseng said the company would “re-evaluate” the financing with investors.

Aechelon’s platform, which is used across the autonomy sector to simulate battlefield environments and train AI machines, will remain open to other customers following the acquisition, Tseng said.

This story was originally featured on Fortune.com

A federal judge has ruled in favor of artificial intelligence company Anthropic in temporarily blocking the Pentagon from labeling the company as a supply chain risk.

U.S. District Judge Rita Lin on Thursday said she was also blocking President Donald Trump’s directive ordering all federal agencies to stop using Anthropic and its chatbot Claude.

Lin said the “broad punitive measures” taken against the AI company by the Trump administration and Defense Secretary Pete Hegseth appeared arbitrary and capricious could “cripple Anthropic,” particularly Hegseth’s use of a rare military authority that’s typically directed at foreign adversaries.

“Nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government,” Lin wrote.

Lin’s ruling followed a 90-minute hearing in San Francisco federal court on Tuesday at which Lin questioned why the Trump administration took the extraordinary step of punishing Anthropic after negotiations over a defense contract went sour over the company’s attempt to prevent its AI technology from being deployed in fully autonomous weapons or surveillance of Americans.

Anthropic had asked Lin to issue an emergency order to remove a stigma that the company alleges was unjustifiably applied as part of an “unlawful campaign of retaliation” that provoked the San Francisco-based company to sue the Trump administration earlier this month. The Pentagon had argued that it should be able to use Claude in any way it deems lawful.

Lin said her ruling was not about that public policy debate but about the government’s actions in response to it.

“If the concern is the integrity of the operational chain of command, the Department of War could just stop using Claude. Instead, these measures appear designed to punish Anthropic,” Lin wrote.

Anthropic has also filed a separate and more narrow case that is still pending in the federal appeals court in Washington, D.C.

Lin wrote that her order is delayed for a week and doesn’t require the Pentagon to use Anthropic’s products or prevent it from transitioning to other AI providers.

This story was originally featured on Fortune.com

Bitcoin’s largest options expiry of the year is colliding with geopolitical volatility that shows no sign of letting up with make or break peace talks uncertain. 

Roughly $14 billion of Bitcoin options are set to expire Friday, as measured by the number for outstanding contracts, known as open interest. The quarterly rollover—which wipes out close to 40% of open positions on the dominant Deribit exchange—comes amid conflicting signals on the prospect of a halt to the nearly month-long war in the Middle East.

The overlap is sharpening a key question for traders: whether the expiry has been artificially muting Bitcoin’s price swings and if its removal will expose the token to a sharper move driven by geopolitics.

Bitcoin has been stuck between roughly $60,000 and $75,000 in recent weeks, drifting well below its October 2025 peak of around $126,000 after a market-wide crash on Oct. 10. The lack of direction has persisted despite geopolitical tensions and intermittent inflows into U.S. exchange-traded funds. Bitcoin fell as much as 4% to $68,122 on Thursday. 

Derivatives positioning helps explain the calm, according to market participants. Institutional investors spent much of the first quarter selling upside bets—effectively wagering that prices wouldn’t rise sharply—to generate income in a subdued market, said James Harris, chief executive officer at asset manager Tesseract. That activity shifted risk onto market makers, who have been buying on dips and selling into rallies to keep their exposure balanced.

The result has been a dampening of volatility, traders say, with price action repeatedly gravitating toward a so-called “max pain” level—the point where the largest number of options expire worthless—near $75,000. In practical terms, those hedging flows have acted like a magnet, nudging Bitcoin higher while capping gains.

“The hedging flows might pull price action toward that level as settlement approaches but effectively cap the range,” Harris said.

Once the contracts roll off, the mechanical buying and selling tied to hedging will fade, potentially leaving Bitcoin more exposed to external catalysts. And those catalysts are mounting. On Thursday, President Donald Trump pushed back his deadline for Iran to strike a deal with the U.S. or face more attacks, saying talks with the country were going “very well.” 

“Without clear direction from the Middle East, Bitcoin is likely to stay in the $70,000–$75,000 zone,” said Andreja Cobeljic, head of derivatives trading at AMINA Bank, adding that the upper bound could act as both a magnet and resistance. A credible ceasefire could push Bitcoin above $75,000, triggering further gains as bearish positions are unwound. Failure in negotiations, however, may drag the token back toward the rising trend line at $68,500, he added.

The broader backdrop offers limited support. While March has seen about $1.5 billion of net inflows into Bitcoin ETFs—a stabilization after four straight months on net outflows — those allocations have proven sensitive to macro shifts. A single day in mid-March saw $163 million pulled as interest-rate expectations changed.

That fragility underscores the central takeaway from Friday’s expiry: the calm in Bitcoin may be more structural than fundamental.

Jasper De Maere, an OTC trader at Wintermute, said that options dynamics can create a “mild upwards bias,” but conviction remains weak. Once the expiry passes, the forces suppressing volatility will recede — leaving macroeconomics and geopolitics firmly back in control.

That leaves the market exposed to sharper moves if sentiment turns.

“The risk is not that institutions are absent. The risk is that they are present but will exit rapidly if the weekend delivers an adverse outcome and the structural cushion that was there last week will not be there to slow the move. Volatility is more likely to increase from Friday than decrease as a result,” Harris said.

This story was originally featured on Fortune.com

After a major insurance company denied coverage for a medical flight that could save her daughter’s life, Alexandria McMahon took the story to social media and caught the attention of billionaire Mark Cuban.

Stella McMahon was diagnosed with T-cell leukemia at just four months old. Now, a year later to the day, she’s been fighting relentless fevers above 104 degrees for nearly a month; her liver’s overtaxed, and her body’s too immunocompromised to ward off a virus on its own. The cruel irony of Stella’s condition: because she has T-cell leukemia, her doctors at Children’s Minnesota had successfully eradicated her T-cells, the very cells she needed to fight the virus now threatening her life.

The McMahon Family

Her oncologist, Dr. Lane Miller, identified a solution: a federally funded study at Cincinnati Children’s Hospital in which genetically modified T-cells, donated and engineered in a lab, are transfused into the patient. The procedure itself was covered. The medical flight to get Stella there was not.

Her mom, Alexandria, submitted a pre-authorization request through the family’s major insurance provider on a Sunday, March 15. For five days, she heard nothing. “I called them on a Friday just to try to understand what was happening, like why it was taking so long,” she told Fortune, while the sounds of little Stella were heard in the background. “And from that phone call, I learned that we had actually been denied.”​

The silence, she said, was devastating in its timing. “To be denied on a Friday from a major business was quite a hit, because they’re closed on the weekend, and they said that they wouldn’t be getting back to us for 24 to 72 hours, business days.”​ Dr. Miller, McMahon explained, said it was imperative that little 16-month-old Stella receive treatment immediately.

When she pressed the representative for an explanation, she ran into a wall. “She couldn’t tell me the reason why they denied because I didn’t have a medical degree,” McMahon recalled. “I asked her to read [the policy] with me line by line and tell me at what point Stella became disqualified. Because when I read it with my eyes, it looked like everything should be approved.” She recorded the conversation, a decision that would change everything.​

TikTok video leads to a breakthrough

McMahon posted the recorded call on social media, and shortly after, it went viral. “I think within about 12 hours, Dr. Warris, who runs Claimable, one of Mark Cuban’s companies, he said, ‘Mark Cuban saw your video, and he wants me to take care of this for you,’” McMahon said.​

Claimable, co-founded by Dr. Warris Bokhari, a former NHS physician and healthcare strategist, uses AI to help patients and families navigate and appeal insurance denials. The company’s mission is to “amplify your voice, combining it with cutting-edge science and policy insights to help protect your rights,” according to its website.​

Cuban, who has long been a vocal critic of the American healthcare system, wrote on his blog in Jan. 2025 that “healthcare is a very simple industry made complicated,” arguing for radical transparency and the removal of insurance companies from the payment equation. His other healthcare venture, Cost Plus Drugs, has similarly sought to reduce opacity in pharmaceutical pricing.

The McMahon Family

Cuban’s big credit limit and bigger heart

Within 48 hours, the McMahons were on a chartered medical aircraft to Cincinnati, paid for by Cuban and Claimable. “The hospital case manager was able to book a medical flight with Mark Cuban and Claimable’s money,” McMahon clarified.​

The trip needed to happen in a single day. Cincinnati Children’s Hospital did not admit Stella to its campus; she had to remain based in Minneapolis. With Stella connected to five or six pumps, driving was never a realistic option. “She could not ethically be discharged,” McMahon explained, saying the family was so desperate to get Stella her treatment, no idea was too out there. “We were considering renting an RV and just trying.” Even that possibility was off the table, given Stella’s medical needs. A friend of her husband’s who works for Delta had also offered a plane, but that would have meant discharging Stella without medical support and “hoping for the best.”​

Instead, Cuban’s team moved fast. “He basically laid down a credit card right then and there, and said, ‘ Book the flight, book a medical flight,” McMahon said. “And we are going to help you fight insurance after she gets taken care of. Stella is the most important.”​

For McMahon, the speed and generosity of the response were difficult to fully absorb while simultaneously watching her daughter struggle. “We are dramatically humbled. We are so thankful that this happened,” she said. “I still feel like it’s unbelievable. I still feel like I’m kind of coming down from something that was just so exciting and so positive. And it felt weird being so incredibly happy and ecstatic and thankful while watching my daughter clearly struggle.”​

She had no prior knowledge of Cuban’s track record of intervening in cases like Stella’s until the outpouring of comments on her videos. “I had no idea that this person could do something so amazing,” she said. “He has the resources, and he has the heart, and he did it.”​

The McMahon Family

Inspiration to pay it forward

At the time of Fortune‘s interview on Thursday, Stella was showing her first signs of improvement: less jaundiced, her eyes beginning to clear. She was still running fevers above 104 degrees, and her doctors cautioned that the T-cell study can take five to seven days before showing results. But she had avoided an ICU stay. “Stella is now stable,” McMahon said. “She got everything she needs. We’re where we need to be, and it is all thanks to human connection.”​

A GoFundMe organized to help the family cover costs, including lost income for McMahon’s husband, who works in aviation and has had to take time off, had raised over $42,000 toward a $50,000 goal from more than 870 donors as of publication.

McMahon said she hopes to use the attention Stella’s story has drawn to advocate for other families facing the same walls. “If I can turn around and give back to people, I will,” she said. “I will carry this and try to pay it forward for the rest of my life.”

The McMahon Family

This story was originally featured on Fortune.com

Iran and the United States appeared at an impasse Thursday, hardening their positions over ceasefire talks and setting the stage for another potential escalation in the Middle East war as thousands more U.S. troops neared the region.

Meanwhile, President Donald Trump extended his deadline for Iran to open the Strait of Hormuz to April 6, and Tehran tightened its grip on the crucial strait while Israel poured more troops into southern Lebanon to fight the Iran-backed militant group Hezbollah.

Sirens over Israel warned of barrages of incoming Iranian missiles, and Gulf nations worked to intercept fire. Heavy strikes were reported in Iran’s capital and other cities.

In a war that appears defined by who can take the most pain, the U.S. has offered shifting objectives, including ensuring Iran’s missile and nuclear programs are no longer a threat and ending Tehran’s support for armed groups in the region. Washington at one point also pushed for the overthrow of Iran’s theocracy.

While the U.S.-Israeli campaign has hit Iran’s military and government hard, killing top leaders and striking scores of targets, Iran continues to fire missiles, and there is no sign of an uprising against the government.

Surviving could be seen as victory

for Iran

For Iran’s leadership, by contrast, merely outlasting the onslaught could be seen as victory. It may be hoping to get the U.S. to back down by roiling the world economy with its stranglehold on the the strait, which has disrupted oil and natural gas shipments and raised prices worldwide for energy and other goods.

Short of a negotiated solution, the U.S. would need a dramatic escalation to end Iran’s attacks and restore the free flow of goods through the strait, where 20% of all traded oil and natural gas is transported in peacetime. Iran rejected a ceasefire proposal put forth by the U.S., while putting forth its own demands.

Trump has vowed to strike Iran’s power plants if it does not fully reopen the strait. His new deadline pulls back on his earlier threat to bombing Iran’s energy plants if Tehran did not open the critical waterway.

Iran had threatened to retaliate against the region’s vital infrastructure, like desalination facilities, if Trump followed through. Trump said he was holding off on carrying out his threat because talks aimed at ending the conflict are going “very well.”

A Gulf Arab bloc said Thursday that Iran is now exacting tolls from ships to ensure their safe passage through the waterway.

Iran seen as operating Strait of Hormuz as ‘de facto toll booth’

Iran has been blocking ships from the strait that it perceives as linked to the U.S. and Israeli war effort, while letting through a trickle of others. Trump said during a Cabinet meeting Thursday that Iran is allowing some oil tankers through as a sign of good faith for talks.

Jasem Mohamed al-Budaiwi, secretary-general of the Gulf Cooperation Council, a bloc of six Gulf Arab nations, said Iran was charging for safe passage.

Lloyd’s List Intelligence called it a “de facto ‘toll booth’ regime,” saying that at least two vessels have paid in yuan, China’s currency.

Iran’s grip on the strait and relentless attacks on Gulf energy infrastructure have sent Brent crude, the international standard, up more than 40% since the war started.

Israel said it killed the head of the Iranian Revolutionary Guard’s navy, Commodore Alireza Tangsiri, and the country’s naval intelligence chief, Behnam Rezaei. Israeli Defense Minister Israel Katz said Tangsiri was responsible for bombing operations that have blocked ships from crossing the Strait of Hormuz. Iran did not immediately acknowledge the killings.

This story was originally featured on Fortune.com

Southeast Asia is racing to build the infrastructure powering the AI boom, but its hot, humid climate could be making that expansion more complicated. Data center demand in the region, where supply is up to 70% lower than in mature markets like the U.S. and China, is expected to grow by 20% each year through 2028, according to the U.S.-ASEAN Business Council. There are now 370 data centers in the region, with the majority in Singapore, Indonesia and Malaysia.

“The ecosystem has realized that if they don’t latch on to this next wave, they might end up being digitally colonized,” Mayank Shrivastava, the CEO of Singapore-headquartered BDx Data Centers, told Fortune. “Economic gains flow to the country that converts raw material into finished goods–and, in this case, the raw material is data.” 

Yet, Southeast Asia’s sultry, tropical climate presents a unique challenge for its data centers, which require more energy than counterparts in cooler climates to keep servers up and running. The region’s temperature ranges between 80 and 95°F throughout the year, while data centers should ideally be maintained between 64 and 81°F, according to the American Society of Heating, Refrigerating and Air-Conditioning Engineers.

“The central issue in the tropics is not heat alone, but heat and humidity together,” Lee Poh Seng, a professor specializing in thermal systems at the National University of Singapore (NUS), explains. “In tropical climates, higher ambient temperatures make heat rejection more difficult, while high humidity complicates dew-point control, increases condensation and corrosion risk, and reduces long-term reliability.”

That puts data center operators in a bind. Lots of people live in the tropics, and data centers need to be close by to ensure speedy access. “You can’t ignore the fact that 85% of the world’s population lives outside temperate regions,” Shrivastava says. 

On March 11, BDx became the first firm to implement Singapore’s Tropical Data Center Standard, a set of guidelines aiming to help data centers gradually increase operating temperatures to 26°C (or 78.8°F). The standard was launched last August, and is a core tenet of the country’s Green Data Center roadmap, which seeks to chart sustainable growth pathways for Singapore’s data centers. According to the country’s Infocomm Media Development Authority, every 1°C increase in operating temperature translates to up to 5% in energy savings.

“It’s been a mammoth effort to get different stakeholders to agree on changing the operating metrics of a data center,” Shrivastava says. “It had to be done with extreme caution, since we were tinkering with the engines while the aircraft was flying.”

‘Room to build’

Data center CEOs see Southeast Asia as filling a critical gap in the global AI ecosystem, particularly as companies in the U.S. struggle to overcome outdated power infrastructure and political opposition to new projects.

“The U.S. is still the world’s largest data center market, but it’s facing a lot of constraints, with each state having different regulations on the speed of grid build-up,” Eric Fan, the CEO of Bridge Data Centers, tells Fortune. “Many projects in the U.S. have thus been delayed—and it’s been a global play for where this gap can be plugged.” 

Malaysia plans to add as much as eight gigawatts of gas-fired power by 2030 to meet the growing needs of data centers, while Singapore has pledged over 1 billion Singapore dollars ($784 million) over the next five years for public AI research. 

The global tech sector is also pouring into the region, with giants like Amazon, Microsoft, Google, Alibaba and Tencent all investing billions of dollars in hyperscale data centers.

“Southeast Asia still offers room to build,” says Lee, from NUS. “Tech companies also increasingly see the region as an attractive deployment zone because of its population scale, fibre connectivity and position between the large North Asian and South Asian digital markets.”

BDx Data Centers was founded in 2019, and is now in four different markets: Singapore, Hong Kong, mainland China, and Indonesia. The firm’s largest operations are in Indonesia, where it has six data centers, including a 100MW campus in Jakarta. 

Fellow operator Bridge Data Centers was also founded in Singapore in 2017, and now operates data centers across India, Malaysia and Thailand. The firm is backed by Boston-headquartered Bain Capital, which sold a data center firm, Chindata, to a Chinese-led consortium for $4 billion. 

‘An energy-and-cooling challenge’

Despite the excitement about AI’s ability to scale in the digital world, companies can’t escape the real world problems of heat and electricity. 

“AI infrastructure is fundamentally an energy-and-cooling challenge wrapped inside a digital-economy opportunity,” Lee says. “The winning projects in Southeast Asia will not be the ones that simply build the fastest, but the ones to show credible performance in power usage effectiveness, water use, carbon intensity, and grid compatibility.” 

He suggests that firms need to take a “power-first, water-aware and thermally intelligent approach”, by situating projects in places with access to clean power, employing high-efficiency designs and moving beyond room-level cooling to chip-level or liquid-based heat removal.

Both BDx and Bridge Data Centres are exploring alternate energy sources to power their operations. “The big advantage of being in a tropical climate is that you get lots of sun and wind, and you have water around you,” explains Shrivastava of BDx.

Separately, Bridge Data Centers is studying hydrogen and nuclear power, with the goal of achieving carbon neutrality by 2040. One of its Malaysian data centers already sources half its energy from solar, a model Fan wants to adopt in its other locations.

And with traditional sources under strain due to conflict in the Middle East, data centers know they need alternatives. “The Iran war has caused the price of oil to skyrocket, raising concern on the reliability of traditional energies,” Fan says. “This will further push the region’s AI firms to diversify into renewable and greener forms of energy.”

This story was originally featured on Fortune.com

Most working women in the U.S. believe they are disadvantaged when it comes to earning competitive wages, but many men hold a different view, according to a new AP-NORC poll.

Equal pay emerged as a major source of concern for working women in the poll and an area where men and women are far apart in their perception of gender equity.

Most women who are employed full-time — about 6 in 10 — say men have more opportunities when it comes to earning competitive wages, according to the survey from The Associated Press-NORC Center for Public Affairs Research, while about one-third think neither gender has an advantage. About 3 in 10 employed women say they have personally experienced wage discrimination because of their gender.

Men who are employed full-time are more divided: About 4 in 10 believe men have an advantage when it comes to wages, while about half think both genders have about the same opportunities and about 1 in 10 say women have more opportunities. Just about 1 in 10 men say they have personally experienced wage discrimination because of their gender.

The survey also found that a majority of employed women say the amount of money they get paid is a “major” source of stress in their life right now, compared to about 4 in 10 employed men.

The findings come at a time when men’s earnings are rising faster than women’s, and the gender wage gap has widened for two years in a row, according to the U.S. Census Bureau.

Reflecting that shift, Equal Pay Day — which symbolizes how many more days into the year women have to work for their earnings to catch up with men — was Thursday, falling a day later than in 2025. That was still 16 days earlier than the first Equal Pay Day on April 11, 1996, when women earned about 75 cents for every dollar earned by men.

The country is deeply divided over how to confront gender pay disparity. A growing number of mostly Democratic-led states are adopting pay transparency laws aimed at making it easier to uncover unfair practices, including requiring employers to disclose pay ranges in job postings.

President Donald Trump’s second administration, for its part, has hollowed out some agencies and limited legal tools that have been key to investigating unfair pay practices, arguing they threatened meritocracy and presuppose that disparities in the workforce are the result of discrimination.

Many employed women say they’ve experienced wage discrimination

Jessica Thompson, 47, said she has seen gender bias throughout her working life. Until losing her job in January, Thompson said she earned $65,000 a year as a senior sales manager in Rockford, Illinois, while a male colleague with similar credentials had earned $87,000.

Thompson said she had to “really prove myself over four years to get the role. And you know, he just came in, just within a few months and got it.”

The poll indicates that women are particularly likely to see wages as a pain point. Fewer women, about 2 in 10, say they’ve been discriminated against in getting hired because of their gender, and men are about as likely to say the same thing.

The overrepresentation of women, especially Black and Hispanic women, in lower-paying jobs is a key driver of the gender wage gap, as is the “motherhood penalty.” Studies show that women’s earnings fall after having children while men see their wages increase after becoming fathers.

Earnings for women barely rose in 2024, while male earnings jumped 3.7%, widening the gender wage gap for the second straight year after two decades of slight narrowing, according to the latest annual report from the U.S. Census Bureau, which analyzes earnings for full-time workers. Women working full-time on average earned 80.9% of what men earned in 2024, down from 82.7% in 2023.

Most employed women say their pay is a ‘major’ source of stress

Women aren’t just likelier than men to be worried about pay equity — the poll also found that employed women are more economically stressed on a range of measures.

About 6 in 10 working women say the cost of groceries and the cost of housing are a “major” source of stress in their lives, and about half, 56%, say this about the amount of money they get paid. By contrast, about 4 in 10 employed men say the same.

Economists attribute the widening pay gap in part to the post-pandemic return to work of many low-wage women, which brought down the average female earnings. But the past two years have also seen a drop-off in the labor force participation rate of mothers with young children, in part because return-to-office mandates have reduced pandemic-era flexibility.

Few men think they are disadvantaged

Democratic lawmakers have criticized the Trump administration for making it more difficult to investigate wage discrimination as part of its campaign to stamp out diversity and inclusion practices.

Trump has ordered federal agencies to stop enforcing ” disparate impact liability,” a concept in civil rights law that has been used in wage discrimination cases against top companies. The Labor Department has also gutted the Office of Federal Contract Compliance Programs, an agency that has audited the pay practices of major companies and obtained hundreds of millions of dollars in compensation for women and minorities who have suffered from unfair policies.

The Equal Employment Opportunity Commission, meanwhile, has pivoted to prioritizing anti-DEI investigations under the premise that men, especially white men, have been discriminated against by practices aimed at advancing women and minorities in the workplace.

The poll suggests that few men see themselves as disadvantaged compared to women in the workplace. Only about 1 in 10 employed men said women had more opportunities when it comes to competitive wages or job advancement.

Michael Bettger, a 51-year-old mechanic who earns $26 an hour in rural Arkansas, said he has seen his wages fall as a result of layoffs and a decade-long struggle with opioid addiction that started after he hurt his back in a worksite accident. But he still believes women struggle more to get ahead in his male-dominated field because of the misogyny he sees, saying other mechanics make jokes about being prone to accidents because female colleagues are a distraction.

“Men do have an advantage and more opportunities for wages. I’ve seen that first hand,” Bettger said. “I have a daughter who wants to be a mechanic, and I’m scared to death of what kind of work she’s going to get.”

___

Savage reported from Chicago and Sanders reported from Washington.

___

The AP-NORC poll of 1,156 adults was conducted Feb. 5-8 using a sample drawn from NORC’s probability-based AmeriSpeak Panel, which is designed to be representative of the U.S. population. The margin of sampling error for adults overall is plus or minus 3.9 percentage points.

___

The Associated Press’ women in the workforce coverage receives financial support from Pivotal Ventures. AP is solely responsible for all content. Find AP’s standards for working with philanthropies, a list of supporters and funded coverage areas at AP.org.

This story was originally featured on Fortune.com

Josh D’Amaro stepped into that corner CEO office at Walt Disney with a clear goal, pledging the company would focus “on coming together as one Disney to deliver a more connected, personalized, and immersive experience to our consumers.”

D’Amaro described a world of Disney entertainment that consolidates its vast intellectual property of movies, games, experiences, and more under one roof. It took a week for some of the biggest bets underpinning that vision to collapse.

The vision of a sweeping field of Disney content relied on multiple external partnerships worth billions of dollars. Several of them unraveled entirely. 

D’Amaro, a veteran leader in Disney’s theme park division, was brought in to steady the ship after a period of uncertain leadership. But three major developments, mostly stemming from decisions made far away from the Magic Kingdom, have made his debut at the top of the company memorable for all the wrong reasons.

OpenAI pulls out the rug

A watershed deal Disney struck with OpenAI late last year dissolved suddenly on Tuesday when the tech company announced it was closing down its Sora video generator app, part of OpenAI’s wider efforts to contain spending ahead of a possible IPO later this year. That ended what was supposed to be a three-year $1 billion partnership, under which some 200 Disney characters from Star Wars, Marvel, and other brands would populate short-form AI-generated videos on Disney+. 

OpenAI’s decision came as a shock to Disney executives, who learned that Sora would be shut down just 30 minutes after they had been meeting with OpenAI about the video generator’s future, according to Reuters. One anonymous source called OpenAI’s decision a “big rug-pull.” 

OpenAI’s CEO Sam Altman is reportedly planning a strategy shift to refocus on business fundamentals and a more streamlined product lineup. Sora was wildly popular in terms of downloads and engagement, but proved difficult to monetize despite high operating costs, making it an obvious target if OpenAI was looking to cut costs. Disney may choose to pursue deals with other AI-powered video platforms, but at least for now, its ambitions of fully integrated AI video populated with Disney characters have effectively become collateral damage in another company’s pivot.

Fortnite isn’t so fun anymore

Also on Tuesday, Epic Games—the videogame developer of Fortnite fame—announced it was laying off 1,000 employees after updates to its hit signature product failed to translate to higher engagement. That’s bad news in general. For D’Amaro, it’s personal.

D’Amaro was the chief architect of Disney’s $1.5 billion investment in Epic, announced in 2024. The deal gave Disney a large equity stake and called for the creation of an entirely new digital universe built around Disney characters and stories, where users could engage in immersive entertainment and shopping. As part of the deal, D’Amaro also joined Epic’s board as an observer. The partnership was the cornerstone of his fan-engagement mission: a Fortnite-powered Disney metaverse where Marvel heroes and Star Wars villains lived alongside players.

“He sees the digital realm—and Epic is a manifestation of that—as a very important place for fans to interact with their favorite characters, franchises and brands in a comprehensive way that you can monetize,” Kevin Mayer, a former head of strategy, told the Hollywood Reporter about D’Amaro’s ambitions in February.

In a memo to staff, Epic founder Tim Sweeney said a downturn in Fortnite engagement had left the company in a financial rut, though he added that $500 million in cost cuts should position Epic for major launch plans toward the end of the year. Whether those plans still include Disney’s digital universe remains to be seen.

A Bachelorette scandal

If the tech wreckage weren’t enough, D’Amaro also inherited a reputational fire at ABC, the network company owned by Disney. Last week, ABC cancelled the already-filmed 22nd season of The Bachelorette amid domestic violence allegations directed at Taylor Frankie Paul, the planned star this season.

It was a messy and widely covered distraction that arrived precisely when the new CEO least needed one. It’s the latest in a string of controversies surrounding The Bachelorette and its companion show, The Bachelor, both of which have long been criticized for underrepresenting people of color in their lineups and promoting sexist stereotypes. But ABC’s move last week is the first time it has cancelled a season of one of its signature franchises after having already filmed it, a call that could cost the company millions.

Disney stock has dipped more than 4% over the past week and underlines the challenge behind D’Amaro’s vision of technology as a growth engine. The developments at OpenAI and Epic may have been out of his control, but they have regardless undermined the universe D’Amaro described so keenly just a week ago.

This story was originally featured on Fortune.com