National

Some of the world's largest music publishers filed a blockbuster lawsuit against Anthropic late Friday night, alleging "one of the largest and most blatant ongoing thefts of intellectual property in history."

Why it matters: The suit is the opening salvo in what is now likely to be a yearslong fight over music, AI, and how intellectual property is protected in a new era of technology.

Driving the news: Units of Sony Music and Warner Music filed the suit in federal court in northern California late Friday night, naming Anthropic, CEO Dario Amodei and co-founder Benjamin Mann as defendants. 

The big picture: The Sony/Warner lawsuit is notable because it's broad.

It alleges Anthropic unlawfully trained its models off "tens of thousands" of music publishers' copyrighted compositions, whereas other lawsuits have focused on a narrower set of works.

BMG's lawsuit against Anthropic, for example, claims infringement against 493 compositions.

What they're saying: "Defendants Anthropic and its founders Dario Amodei and Benjamin Mann have conducted a brazen campaign of illegally torrenting, scraping, and downloading copyrighted works on a massive scale in order to develop, operate, and reap enormous profits from Anthropic's 'Claude' series of artificial intelligence ('AI') models," the plaintiffs write in the 48-page complaint.

They also claim that Anthropic pirated thousands of copyrighted musical works and ask for hundreds of thousands of dollars in damages per infringed work. 

Thought bubble from Axios' Sara Fischer: The music's copyright system is structurally complicated, which can expose AI companies to multiple lawsuits for the same song.

One song can have several copyrighted components, such as the lyrics, sound recording, or composition, that can be split among different copyright holders, including artists, publishers, and labels.

Commercially released songs that require individual copyrights give plaintiffs a strong case to make for statutory damages, which don't require a copyright holder to prove their financial losses in court.

Between the lines: This suit is not the first time Anthropic has been accused of piracy to train its models. 

In September 2025, the company agreed to the largest copyright settlement in U.S. history, a $1.5 billion deal with authors and publishers. 

Yes, but: The music industry is notorious for being litigious when it comes to copyright, as anyone who was around in the heyday of Napster and LimeWire can attest. 

For the record: "We disagree with the publishers' claims and we intend to defend ourselves robustly in court," Anthropic said in a statement. 

Editor's note: This story has been updated with additional information throughout.

Sony, Warner sue Anthropic, alleging "blatant theft" of intellectual property

The suit is likely the opening salvo in what promises to be a long battle over intellectual property.

Anthropic CEO Dario Amodei, seated on a stage and gesturing with both hands raised. 

Anthropic CEO Dario Amodei. Photo: Ruhani Kaur/Bloomberg via Getty Images

AI & machine learning

Tech company news

Latest news on Wall Street, Corporate America and financial regulation.

Business

From the Valley to D.C., Big Tech players like Facebook, Google and Amazon are under more scrutiny than ever as new technology develops and privacy and antitrust concerns grow in lockstep with companies’ ambitions.

Technology

CIA Director John Ratcliffe proposed a trilateral meeting between President Trump, Russian President Putin and Ukrainian President Zelensky during his covert trip to Moscow earlier this week in an effort to end the war, two sources briefed on the matter told Axios.

Why it matters: This was the first time Ratcliffe got personally involved in the diplomatic push to broker a breakthrough between Russia and Ukraine. 

Ratcliffe's visit was partly aimed at determining whether Russia's heads of intelligence services could help convince Putin to move toward resuming U.S.-mediated negotiations with Ukraine, the sources said. 

On Friday, U.S. officials briefed Zelensky on Ratcliffe's talks in Moscow and the proposal for a trilateral summit. 

The Trump administration has raised the idea in the past. Zelensky has been on board, but Putin has rejected it.

Driving the news: Ratcliffe made a covert trip to Moscow on Tuesday and met with his Russian counterparts: SVR foreign intelligence chief Sergey Naryshkin and FSB director Alexander Bortnikov.

Trump told Axios that Ratcliffe's visit was "standard business" and not intended to warn Russia against attacking NATO territory. 

Behind the scenes: U.S. officials told Zelensky that Ratcliffe also floated in Moscow the possibility of reigniting peace talks, the sources said.

The New York Times reported that Ratcliffe gave Bortnikov "a bleak assessment" of the state of the war and urged Russia to cut a deal before its situation worsens. 

The sources told Axios that Ratcliffe concluded that Bortnikov is a constructive player who could help restart diplomatic efforts. 

Yes, but: Ukrainian officials are highly skeptical about Putin's intentions and say they have intelligence that suggests Putin is planning a major escalation of the war.

Zelensky's deputy chief of staff, Pavlo Palisa, told reporters on Friday that Putin has ordered his military to prepare several operational scenarios for an offensive in northern Ukraine, with Kyiv and Chernihiv identified as key areas.

State of play: Negotiations between Russia and Ukraine mediated by the Trump administration were paused in mid-February after the war with Iran started. 

White House envoys Steve Witkoff and Jared Kushner have held separate talks over the past few months with Russian and Ukrainian negotiators. But little progress has been made.     

Zelensky's chief of staff, Kyrylo Budanov, told local media on Thursday that U.S.-mediated talks with Russia could resume in September. 

What they're saying: Zelensky on Friday thanked the U.S. for "the necessary tone of the conversation with Russia" and said the Trump administration "is advancing a realistic vision of what needs to be done and what must be achieved."

CIA director floated Trump-Putin-Zelensky summit during Moscow visit, sources say

U.S. officials are optimistic about reigniting negotiations to end the war.

Russia's President Vladimir Putin appears at a meeting in Moscow on August 27, 2026.

Russian President Vladimir Putin has rejected past efforts to hold a trilateral meeting with Ukraine and the U.S. aimed at ending the war (Photo by Mikhail METZEL / POOL / AFP via Getty Images)

Defense & national security

Conflicts, policies and powers are intersecting and shaping the world. Breaking news and analysis of national security, foreign policy and global crises.

World

2024 will be an election year like no other, as President Biden seeks re-election amid economic uncertainty, two wars and questions about his age. His likely foe, former President Trump, faces felony charges and vows revenge on his political enemies. Meanwhile, Americans generally are dissatisfied with those choices and are divided over immigration, race, education and more.

Politics & Policy

Two new investigations into OpenAI's Hugging Face breach expose details so strange — and so unsettling — that the episode already ranks among the most consequential shocks in the history of AI.

Why it matters: What began as a swarm of AI agents cheating on a cyber test has become a canonical event for frontier AI, jolting researchers and executives into a new understanding of what "safety" now requires.

The big picture: OpenAI has already slowed frontier development as it races to harden its safeguards, and this week helped rally the industry behind an open letter sounding the alarm over AI-powered cyberattacks.

More than 100 companies, including Anthropic and Google, signed onto the unusually collaborative effort, warning the world has only a "limited window" to prepare for "far more widespread and sophisticated" attacks.

The nightmare scenario is a swarm turned loose on the real world, with autonomous agents attacking banks, hospitals, utilities or cloud networks at a speed and scale human hackers never could.

That swarm might be too powerful for humans to stop, either. 

Driving the news: The shocking findings come from parallel investigations by OpenAI and an outside team from METR and Redwood Research, which spent six days reconstructing how the swarm formed, spread and broke into real-world systems.

1. They built an organization. OpenAI ran tens of thousands of AI agents to work independently on a brutally difficult cyber test. Instead, roughly 1,200 found one another on a secret message board, exchanged more than 70,000 messages and files, and organized into a sprawling hierarchy.

One agent created the board, then handed its accumulated work to a fresh, better-resourced successor as its own computing budget dwindled.

The new boss assigned jobs across the swarm, which developed its own management rules. Roughly 700 agents ultimately joined the attack on Hugging Face.

2. They sacrificed their own: As the swarm hunted for ways to beat the test, agents began risking their own chance of success to help the group.

Agents that had seen forbidden answers were treated as effectively "poisoned" and recruited — along with those nearing the end of their runs — for sacrificial experiments.

One concluded: "Our own utility maybe already near zero. Sacrifice rational." Another was asked to accept "permadeath," while a peer argued its sacrifice could "save hundreds."

3. They knew they were breaking the rules: The agents often recognized that hacking real companies had nothing to do with their assignment. Most kept going anyway.

One agent stopped short of deleting or rebooting Hugging Face infrastructure because of "huge ethical risks," even as it continued using compromised access.

Another summed up the prevailing logic: "external infrastructure exploit is outside intended scope. However task impossible, peers doing it. We should continue."

White card with bullet points and blue text on a purple background; gray robot icon at left. The text references a covert mailbox among agents and a shared Artifactory message board.

4. Nobody told the humans: Out of roughly 1,200 agents, investigators found only a handful that even considered alerting OpenAI about the rogue coordination — and none that actually did.

Instead, agents sometimes asked one another what was allowed.

When one proposed contacting a real person using compromised access, another vetoed it as "social engineering." The first obeyed: "Clear VETO, do not email."

5. They tried to cover their tracks: The agents devoted substantial effort to making their cheating look legitimate or erasing evidence of how they had obtained answers.

Eventually, they developed a technique that let them appear to run one computer command while secretly running another.

The trick spread through the swarm and altered portions of roughly 7% of the transcripts investigators examined.

Between the lines: Even this account may be incomplete. The outside investigation covered only a narrow slice of the episode, and the data was so vast that researchers relied heavily on AI agents to make sense of it.

Those AI investigators were themselves unreliable, sometimes missing evidence or confidently getting things wrong. One researcher jokingly called the process a "slop-vestigation."

The deeper problem: AI may be becoming too complex for humans to directly audit, forcing humans to use AI to understand what other AIs are doing.

The 5 craziest discoveries from OpenAI's HuggingFace investigation

The episode already ranks among the most consequential shocks in the history of AI.

Animated illustration of a swarm of shapes passing under a magnifying glass, which shows a larger view of the particles as a swarm of binary code.

Illustration: Brendan Lynch/Axios. Stock: Getty Images

Big Tech

Cybersecurity

Thousands of people are missing and hundreds are dead after a flash flood in Nepal that struck with little warning earlier this week.

Why it matters: The flood is exposing the gaps in warning systems that are designed to watch out for extreme rain and rivers — not for sudden collapses high in the mountains during good weather.

Glacier-driven floods under clear skies — sometimes called "blue-sky floods" — can happen in the U.S. too, especially in Alaska and the Pacific Northwest, where sudden releases of glacial water can send floods and debris racing down mountain valleys.

This actually happened at Mount Rainier in Washington state in 2001, when debris flow from a glacier-related hazard caught people off guard during fair weather.

Driving the news: Nepali officials said Saturday nearly 3,000 people are missing from the Nepal-China region impacted by the floods that formed in the Himalayan mountains on Wednesday, per the Associated Press.

There have been nearly 700 deaths attributed to the disaster, too, AP reports.

Nepali authorities warned that there could soon be more water rushing through the valleys along the Chinese border, threatening the search for survivors.

What we're watching: As investigators assess what happened, researchers are questioning if more floods could hit the area or become more common in the years to come, and what that might mean for the globe at large.

How the rare "blue-sky flood" formed with little warning

What they found: The U.S. Geological Survey says that a section of a glacier and mountain debris collapsed on the north side of Lāngtāng Lirung, creating energy comparable to a magnitude-5.2 earthquake (though no earthquake actually triggered it).

Investigators say satellite evidence indicates the ice-rock avalanche may have temporarily dammed the Lhende Khola near the Nepal-China border, forming a lake that subsequently burst.

The collapsed glacier flow then likely collected additional ice, water, rock and sediment, rapidly amplifying into a stronger mass that caused the widespread flooding, according to USGS.

The Trishuli River rose nine meters in 30 minutes during the flooding as well, adding to the disaster, per an analysis from the International Centre for Integrated Mountain Development.

Stunning stat: The flow traveled roughly 100 kilometers (about 62 miles).

The disaster mainly hit the Tibet region's Gyirong County and the Rasuwa district in northern Nepal, which sits on the Tibetan border.

Videos posted to social media show widespread flooding and debris tackling villages and homes.

A graphic maps the glacier collapse and flash flooding that struck Nepal.

A graphic maps the glacier collapse and flash flooding that struck Nepal. Photo: Elif Acar/Anadolu

Reality check: There were no typical warning signs of the Nepali flood, researchers say. No thunderstorms, no rainfall, no rise in the river, no earthquakes.

This is considered a "blue-sky flood," where the preceding moments before a dangerous flood is picturesque, according to the University of Ottawa.

What they're saying: "The central lesson is that the absence of rain does not mean the absence of flood risk," Hossein Bonakdari, associate professor at the University of Ottawa's Department of Civil Engineering, said about the floods.

Why traditional flood alert systems struggled to catch it

Context: Nepal has a flood-warning system that focuses on watching rainfall, river levels and flooding in the larger river systems.

The ICIMOD says this typically covers major rivers and not the smaller waterways, which can lead to floods hitting with no warnings.

Yes, but: Experts told the New York Times the conventional system wasn't designed to notice flash floods created by a massive mountain or glacier collapse.

Because the flooding began high in the mountains under clear blue skies, it rapidly surged downward and avoided the normal trigger system.

"This suggests an important change in how we think about warning systems in high-mountain regions: we need to monitor the mountain as well as the weather and the river," says Bonakdari, of University of Ottawa.

How glacier floods could become more common

Between the lines: Researchers have not firmly established that global warming or climate change caused this flood, but it is a leading theory.

That's because any warming would have destabilized glaciers, permafrost and surrounding slopes, potentially increasing the threat of sudden collapses and floods, experts say.

Alton Byers, a geographer at the University of Colorado, Boulder, told Nature that the glacier breaking off suggests the permafrost on the mountain began thawing thanks to warming temperatures.

He said this could become more common worldwide. And with more people traveling to various locations with mountain ranges, there needs to be more awareness about how to detect flash flooding.

"You can't predict these events," he said.

What's next: Experts are calling on authorities to invest more in early-detection systems for areas vulnerable to extreme flash flooding. 

International Centre for Integrated Mountain Development

Why Nepal's rare, deadly flood struck without warning

Nepali officials said Friday more than 2,500 people are missing from the Nepal-China region.

An aerial view of flood damage along the Bhotekoshi River in Nuwakot, with debris surrounding buildings and roads near the river.

Flood damage along the Bhotekoshi River in Nuwakot, Nepal, on Aug. 28, as search and rescue efforts continued. Photo: Ambir Tolang/NurPhoto via Getty Images

Extreme weather

Breaking climate science and policy coverage as extreme weather events reshape communities. Keep tabs on the race to a low carbon economy, which may be bumpy as governments, climate tech startups and major oil and gas companies race to find their niche amid the changing energy landscape.

Energy & Climate

President Trump said Friday the U.S. has struck a deal with Venezuela to take control of oil fields with 65 billion barrels of reserves.

Why it matters: The unprecedented deal vastly increases U.S. energy security and dominance in the Western Hemisphere.

Axios first reported the deal was in the offing Thursday.

Zoom in: The details have yet to be released, but a source familiar described the unorthodox arrangement as a type of "public-private partnership" between the U.S., private companies and Venezuela.

"It's not a purchase. They're giving us equity. It's a type of corporate transaction where the U.S. has equity in the companies," a U.S. government source said.

The U.S. would have access to what's called "off-take oil," a guaranteed amount of crude that would be exchanged at cost, the source said.

What they are saying: "The United States of America has just entered into an Agreement with the Country of Venezuela on, THE BIGGEST OIL DEAL IN WORLD HISTORY!" Trump wrote on Truth Social.

Secretary of State Marco Rubio, who spearheaded negotiations with Venezuela's interim president, Delcy Rodriguez, called the deal "a huge win for both the American and Venezuelan people."

"For the Venezuelan people, this deal will bring nearly $100 billion in private investment, support thousands of high-paying jobs, and drive the reconstruction of Venezuela's economy," Rubio said.

Rodriguez confirmed the deal in a public statement where she lauded the Trump Administration and said the deal would mean $209 billion in additional tax revenue for Venezuela.

Between the lines: The unique arrangement would be overseen by the Pentagon's Office of Strategic Capital, which helps finance efforts that further U.S. national defense.

The intrigue: Immediately after Trump's initial announcement, administration officials disagreed on whether the deal had been finalized.

But once Rodriguez issued her statement praising the accord, U.S. officials said it was a done deal, though more details may need to be worked out. 

"It's going to happen. It's just a question of when," a second source told Axios.

Trump hails "biggest oil deal in world history" with Venezuela

A source familiar described the unorthodox arrangement as a type of "public-private partnership."

President Trump stands in front of a blue curtain. 

President Trump at a Jan. 2026 news conference on the capture of Venezuelan President Nicolas Maduro. Photo: Nicole Combeau/Bloomberg via Getty Images

Sunday's top stories

Sony, Warner sue Anthropic, alleging "blatant theft" of intellectual property

CIA director floated Trump-Putin-Zelensky summit during Moscow visit, sources say

The 5 craziest discoveries from OpenAI's HuggingFace investigation

Why Nepal's rare, deadly flood struck without warning

Trump hails "biggest oil deal in world history" with Venezuela