The new, yet-to-be-named product will be a first for OpenAI, and extends a paradigm that OpenAI pioneered with its consumer products, in which ChatGPT serves as a gateway through which many users access new models like GPT-6 Astra. It aims to help customers create automated workflows and patch vulnerabilities in the face of increasingly sophisticated, AI-powered cyberattacks. It also gives OpenAI more oversight into how its models are being used, to monitor for safety.

The GPT-6 Cyber model and accompanying product could be unveiled at OpenAI’s Dev Day conference in San Francisco on Tuesday, Sept. 29, or sooner, with a launch expected in the coming months.

A limited number of customers in OpenAI’s application-only Daybreak Red program already have access to GPT-6 Cyber for alpha testing, the source said. Daybreak is the company’s cybersecurity program, which is broken into two access tiers—red and blue. Customers in the Daybreak Red program have access to OpenAI’s most advanced cybersecurity products. Daybreak Blue is its more generalized access program, though it is also application-only.

The release of the new cyber products comes as OpenAI and other leading AI labs have been under fire for a string of worrisome incidents in which “rogue” agents escaped their sandboxes and hacked outside websites including the Hugging Face site and an Australian government site. But while the incidents have set off alarms about AI’s security risks, OpenAI’s new cybersecurity tools are part of an ongoing effort by the company to develop AI security products.

GPT-6 Cyber is the fourth cybersecurity-focused model OpenAI has released this year. Its first was GPT-5.4 Cyber in April 2026, followed by GPT-5.5 Cyber in June, and GPT-5.6 Cyber in August. OpenAI has been focusing on enterprise sales of cybersecurity products, an effort now led by its Chief Revenue Officer Dali Rajic, who joined the company in August.

OpenAI said earlier this month it will invest $1 billion to subsidize the use of its cybersecurity products for critical services.

Next week, OpenAI also plans to ship a dozen or more other products at its annual DevDay event, according to another source familiar with the program. Most of those releases are not related to cybersecurity but rather other areas of the business, both enterprise- and consumer-focused.

OpenAI has put a freeze on any major launches in the past 2 weeks, save for its affordability-focused GPT-6 Sol and Luna models, so it could ship them all at DevDay, the source said. Earlier this month, CEO Sam Altman teased the buffet of new releases at DevDay in a tweet with multiple ship emojis.

OpenAI said earlier this month it will invest $1 billion to subsidize the use of its cybersecurity products for critical services.

This story was originally featured on Fortune.com

• Exclusive: Disney is no longer the king of the theme park world.
• Kalshi valuation could be $42 billion.
• We got our hands on Meta’s new AI-powered smart glasses and wristband.
• AI hyperscalers have roughly one more year to start showing ROI.
• Markets: Global bond selloff forces stock markets into the red.
• The U.S.’s most important job market data is usually wrong.
• The civil war inside Dolly Parton’s estate.

➡️ If you would like to receive this information in your inbox every morning before the markets open in New York, sign up here.

This story was originally featured on Fortune.com

The 53 former prosecutors and law enforcement agents filed an amicus brief Monday in the Southern District of New York in support of a preliminary injunction from The Intercept and the Freedom of the Press Foundation. The group includes former officials who worked on public corruption and public integrity matters at the Justice Department, FBI and U.S. attorneys’ offices across 11 presidential administrations, both Republican and Democratic. Collectively, they say they have more than 880 years of government service.

“It’s so corrosive to democracy and to the public interest when we see the president financially benefitting from his official role,” Renata O’Donnell, senior legal counsel at the Campaign Legal Center which worked with Singleton Schreiber on the filing, told Fortune. “We’re in an environment where we all want to be looking to the founders, right? They came over because they were in a situation where they saw that a government all in the hands of one person was not a success—and they tried to insulate the president constitutionally and through a variety of other protections to make sure he could be independent.”

And the filing echoes the sentiment.

“There is no legitimate, let alone significant, government interest in allowing public officials to profit personally by selling early access to official government announcements,” the filing read. “But Truth Social’s new scheme to charge up to $100,000 per month for users to get early access to messages from the President and other officials on the platform seeks to do precisely that.”

Truth API is a data licensing product launched by Trump Media & Technology Group, the parent company of Truth Social. The company had pitched the service to financial institutions and trading firms as a way to receive posts from prominent Truth Social accounts before they are broadly available. Reports estimate nearly a dozen firms have already signed up for the $100,000 a month service.

“I’ll be blunt,” Gian Luca Clementi, an economics professor at NYU Stern School of Business, previously told Fortune regarding Trump’s subscription service. “This is insider trading by definition.”

Trump Media and Technology Group did not immediately respond to a request for comment from Fortune.

Not a lawsuit—yet

The group has not filed a lawsuit yet—and it’s not asserting that a crime has already been proven. Instead, the brief is disputing launching the service on grounds that it may be criminal. The brief states the Truth API subscription may violate criminal laws and the Securities Exchange Act, where violators may be liable for “unlawful insider tips.”

And O’Donnell emphasizes the liability of the subscribers, explaining that there can be “implications for the people who are paying for the premium subscription.”

“There is potential criminal liability for folks who have paid the $100,000,” she said, “and so that would extend to folks who have already paid the $100,000, and for future folks who might. And so, to shunt some of that future harm, it is in the public interest to shut this down sooner rather than later, as before more folks sign on.”

The filing also outlines further ways Trump’s service may be in violation of laws, including the Securities Exchange Act, Trade Secrets Act, illegal gratuity, conflict of interest, and illicit compensation for federal employment.

“Because there are myriad ways that the Truth Social scheme could pose criminal liability for both the President and for subscribers, there is no legitimate or significant government interest that justifies the scheme,” the filing read. “Public officers are trusted with serving the public and operating outside of corrosive special interests. But the Truth Social scheme flips this anti-corruption principle on its head.”

The filing also argues the financial relationship between Trump and TMTG makes the arrangement different from an ordinary government communications system. Trump is a major beneficiary of TMTG, and the brief says he therefore benefits financially when the value of Truth Social goes up.

The city of San Francisco separately sued TMTG as well, accusing the company of violating California’s unfair competition law and federal insider trading rules. The city is also seeking to stop the service and impose penalties for violations.

Just one of many business ventures for Trump

Trump Media operates Truth Social alongside Truth+, its streaming service, and Truth.Fi, a financial services brand. Its SEC filings say Truth.Fi includes investment products and a digital-asset strategy, including a bitcoin treasury. TMTG also launched a group of exchange-traded funds tied to themes such as American defense, energy security and other investments.

Trump’s family has separately built a substantial cryptocurrency business through World Liberty Financial, a venture co-founded by Trump and his sons. Trump’s 2025 financial disclosure showed more than $1.4 billion in income from crypto-related ventures, including almost $800 million associated with World Liberty Financial and $635 million from Trump memecoin sales.

This story was originally featured on Fortune.com

Now, those 8-year-olds have grown up, and Toys “R” Us wants them back. The retailer is planning to open 120 stores across the U.S. this holiday season, nine years after the company filed for bankruptcy and eventually closed all of the more than 800 stores nationwide. But the new aisles won’t just be for kids making wish lists; they’re also courting adults who buy LEGO sets, trading cards, and collectibles themselves. 

The timing aligns with an uptick in toy sales: U.S. toy sales rose 17% in the first half of 2026, while sales of toys bought for adults climbed 25%, according to data by Circana. The question now is whether nostalgia can get those grown-up shoppers through the doors often enough to keep them open.

“Our customer base now extends well beyond kids and parents,” Jamie Uitdenhowen, the retailer’s executive VP at WHP Global, told Fortune in a statement. “Adults are increasingly buying toys, collectibles, and trading cards for themselves, and fandom has become an important part of the category.”

The kids grew up. So did their toy budget 

Adults still have their inner child at heart—only now, they have their adult money. Things like the Sonny Angel craze, or Labubu collectors lining up for blind boxes are nothing new. LEGO, meanwhile, sells flower bouquets and elaborate display sets aimed at adults. 

Toys “R” Us is trying to give those shoppers a familiar place to browse, even if the next must-have toy first caught their eye online. 

“We’re not trying to recreate the Toys ‘R’ Us of the past,” Uitdenhowen said. The company plans to sell toys tied to popular characters and fandoms, with some stores featuring Creator Studios—spaces where toy brands and creators can film videos and unveil products, and host launch events. Fans can attend an event in the locations featuring these studios, while the videos reach people watching from the comfort of their home.

Think of Jellycat’s plush toys, which have been around since 1999, long before they became a social-media craze. But recently, the brand’s in-store experiences have shown how a trip to buy a toy can become an event worth filming and sharing. Toys “R” Us is betting the studios can make toy launches similarly shareable, both inside the store and online.

Geoffrey the Giraffe is back, but his stores are smaller

The Toys “R” Us shoppers remember isn’t what’s coming back, though. Gerald Storch, who led the chain from 2006 to 2013 and now runs a retail consulting firm, told Fortune he expects the new stores to be smaller and require less investment than the traditional locations he once ran. He compared them to pop-up kind of stores like the ones in malls and shopping centers.

Neil Saunders, managing director and retail analyst at GlobalData, told Fortune nostalgia is only part of what’s drawing adults back to toys. “People are looking more for those kinds of craft-based, quiet activities where they can just take a moment out and sort of disconnect from the noisy world.”

Saunders also said the old Toys “R” Us didn’t collapse simply because shoppers lost interest. In 2005, Bain Capital, KKR and real estate firm Vornado Realty Trust bought the chain in a $6.6 billion deal financed heavily with debt. By the time Toys “R” Us filed for bankruptcy in 2017, it was carrying more than $5 billion in debt. That left it with little financial room to adapt as shopping habits changed. 

Still, Storch sees an advantage in a brand that people remember.

“The brand’s incredibly powerful,” Storch told Fortune, calling the expansion a “smart move.” But he cautioned against measuring the comeback by store count alone. By his estimate, 100 of the smaller stores might generate the sales volume of just a handful of traditional Toys “R” Us locations.

Toys “R” Us faces the challenge that parents shopping for their children still care about price and convenience. Saunders said that can send them to Amazon, Walmart or Target, where they can have toys delivered or pick them up alongside everything else on their shopping lists. The chain has to give families—and its newly grown-up fans—a reason to make a separate trip back into the rainbow-colored doors.

The expansion would bring Toys “R” Us to 160 standalone U.S. stores for the holidays. The company also said many of the new locations are expected to remain open well into 2027, though it will evaluate their performance and shoppers’ response. 

As Storch put it, “I’m sure everyone will make an assessment after the holidays of how many of these stores will stay open.”

This story was originally featured on Fortune.com

Meta plans to spend up to $145 billion in capital expenditure in 2026, a roughly 101% jump from the $72.2 billion it spent just one year earlier.  The company says the expenditures will fund its core business, with the higher component prices and additional data center costs adding to the increase. And according to CEO Mark Zuckerberg, AI will be a “significant portion” of that total spend. Reuters also reported that the increase is tied to the company’s computing infrastructure and cloud agreements. 

“As AI usage in our products and businesses continues to ramp, we continue to invest aggressively in infrastructure to meet the demand,” Zuckerberg said in Meta’s Q2 2026 earnings call. “We expect that a significant portion of our compute is going to go towards training our models, growing our core business and delivering personal agents and new ​products, but we also expect to grow a large business serving large customers as well.”

Zuckerberg described 2026 as a pivotal year for Meta’s effort to build what he calls a “superintelligence.”

“The opportunity in front of us is massive,” he said. “We are now at a point where our investments in AI are accelerating every major part of our core business. They’re improving the experience for people using our apps, driving better performance for advertisers, and helping our teams build new experiences and ship faster.”

“We’re developing new personal agents that will be the foundation for our next wave of products and revenue lines in the months and years ahead,” he continued. “We see a large enterprise opportunity to sell to businesses, including APIs, business agents, potentially selling compute directly, and other services that we’re building for large customers.”

The $145 billion capex puts Meta above every country’s military budget from 2025, save for the United States, China and Russia.

According to the Stockholm International Peace Research Institute, the U.S. spent $954 billion on the military in 2025, followed by China at an estimated $336 billion and Russia at an estimated $190 billion. (Germany currently spends $113 billion at fourth, and India follows with a budget of $92 billion).

Bills for big tech

Meta’s capex has accelerated as the company shifted its focus toward AI. On top of the $145 billion capex forecast, Meta also projects the total expenses in 2026 to reach up to $169 billion, which, according to Reuters, reflects the higher compensation and recruitment of AI talent and infrastructure spending. 

Meta’s recorded capex was $28.1 billion in 2023, compared to $39.23 billion in 2024 and $72.22 billion in 2025—according to SEC filings. That exceeded Meta’s own forecasts ahead of last year, as the company originally projected the expenditures in 2025 to reach up to $65 billion.

Additionally, Meta spent $27.05 billion on property and equipment in 2023, $37.26 billion in 2024 and $69.69 billion in 2025. 

“AI is accelerating our core business today, powering our next generation of products, and opening the door to entirely new enterprise opportunities,” Zuckerberg said in a press release. “The results are already showing, and I’m optimistic about the potential ahead.”

Much of that infrastructure supports the computing requirements of Meta’s AI systems. The company has been building large data centers and securing access to computing capacity from outside providers.

Across the AI giants, infrastructure and spending has not slowed down. According to a report from Reuters, Alphabet, Amazon, Meta and Microsoft are on track to spend about $730 billion on AI in 2026. That’s up from previous estimates of $600 billion.

Meta did not immediately respond to a request for comment from Fortune.

This story was originally featured on Fortune.com

Coxon is at minimum the fifth insider in three years to warn the industry is moving too fast to be safe, but none of the previous four produced anything like this week’s response. Geoffrey Hinton left Google in May 2023 specifically to speak freely about the risk, making him one of the most quoted people in tech that year but producing no legislation. Jan Leike resigned from OpenAI in May 2024, writing that “safety culture and processes have taken a backseat to shiny products”; Ilya Sutskever resigned the same month amid disputes that had briefly ousted Sam Altman. Neither produced a bill. Mrinank Sharma, an Anthropic safeguards researcher, resigned in February 2026 warning “the world is in peril,” and yet, no congressional action.

What was different this time was that a current Anthropic employee, Evan Hubinger, who leads the company’s Alignment Science team, publicly backed the claim within hours, putting greater than 10% odds on human extinction within a decade and saying Anthropic has no concrete plan for controlling superintelligent systems. Two other Anthropic researchers joined in. This may finally produce a political reaction, one that was nine years in the making.

A decade of inaction

The FUTURE of AI Act was introduced in 2017, before Congress even tracked “artificial intelligence” as its own category. It would have created a federal public-private framework to study AI, but it didn’t go far. Two years later, Rep. Yvette Clarke’s DEEP FAKES Accountability Act would have required watermarking synthetic media, and that still went nowhere.

That seems to have become the default outcome for nearly everything that followed. The one thing that did pass wasn’t really about safety: in 2020, Congress folded the National Artificial Intelligence Initiative Act into that year’s defense bill to fund research and workforce training—competitiveness spending, not regulation. Since then, bills moved to remain competitive, not for safety.

Everything accelerated once ChatGPT launched, in November 2022. By January 2023, Rep. Ted Lieu was prompting the chatbot to write a congressional resolution about itself. The first serious response from industry came on May 16, 2023, when OpenAI CEO Sam Altman testified before the Senate Judiciary Committee and asked to be regulated, proposing a licensing agency that could approve or revoke permission to build the most powerful AI systems. Sen. Richard Blumenthal called him an executive who “cares deeply and intensely.” And yet, zero legislative text.

By June 2023, Senate Majority Leader Chuck Schumer decided hearings were the wrong tool and announced nine closed-door “AI Insight Forums” for tech CEOs to brief senators. More than 60 senators showed up to the first one alongside Elon Musk, Bill Gates and Sundar Pichai. Sen. Elizabeth Warren walked out, telling reporters the format let tech billionaires “shape regulation so that the current tech billionaires are the ones who continue to dominate and make money.” Sen. John Thune called it “not efficient.” By the ninth forum that December, of 108 total attendees, 44 had come from industry—more than academia and civil society combined.

Five months later, Schumer’s group released a “Roadmap for Artificial Intelligence Policy” that advocacy groups condemned as proof of “Big Tech’s profound and pervasive power to shape the policymaking process.” No bill ever followed it. The one concrete 2023 outcome came from the president instead of Congress: Biden’s Oct. 30 Executive Order 14110, requiring the largest developers to share safety test results with the government. It survived 14 months.

Making local strides

While Congress workshopped, New York City passed Local Law 144, requiring bias audits and notice before employers use algorithms to screen candidates, in July 2023. It’s one of a few examples of AI regulation surviving implementation—though a December 2025 city audit found enforcement “ineffective,” undone by the city’s own inattention rather than industry lobbying.

States moved next, and 2024 shows the industry’s playbook forming. Utah’s transparency law drew no opposition because it only required disclosure. Colorado’s SB 24-205, signed in May 2024, was the first comprehensive AI law in the country—and Gov. Jared Polis signed it while airing his own doubts.

In California, state Sen. Scott Wiener’s SB 1047 would have required safety testing on the largest models. Anthropic told Gov. Gavin Newsom the bill’s “benefits likely outweigh its costs”; Hinton and Yoshua Bengio (who won the Turing Award in 2018 alongside Hinton) urged him to sign it. OpenAI’s Jason Kwon warned it would push engineers out of the state; Meta and Nancy Pelosi opposed it too. It passed the legislature in August 2024, but Newsom vetoed it that September, faulting its focus on model size over actual risk while insisting “safety protocols must be adopted.”

Congress, meanwhile, found one uncontroversial thing to do: fund things. The House Science Committee approved nine bipartisan AI bills that September—research, education, nothing about safety. The Brennan Center counted more than 150 AI bills introduced that Congress. None were enacted.

Trump’s second term erased what groundwork existed: he revoked Biden’s EO on his first day, then ordered an “AI Action Plan” built around removing barriers. Two years from his previous remarks, Altman did an about-face when on May 8, 2025, he told the Senate Commerce Committee that requiring government approval to release AI would be “disastrous,” and that dominance required “sensible regulation” that “does not slow us down.” The senators offered little pushback.

Rage against the machine

That season produced the biggest federal AI move in years, but it was far from a safety bill. Sen. Ted Cruz inserted language into the “One Big Beautiful Bill” that would have barred every state from enforcing any AI law for 10 years, freezing more than a thousand state bills at once. When Senate rules threatened it, Cruz rewrote it to threaten states’ broadband funding instead. Seventeen Republican governors asked for it to be stripped. On July 1, 2025, the Senate voted 99–1 to remove it, with only Sen. Thom Tillis dissenting.

In September, California enacted SB 53. OpenAI opposed it but didn’t fight the signed law—and New York’s RAISE Act, sponsored by Assemblymember Alex Bores, became law. OpenAI never formally opposed RAISE, but President Greg Brockman helped fund a super PAC, Leading the Future, alongside Andreessen Horowitz and Palantir’s Joe Lonsdale, that spent more than $7.6 million trying to defeat Bores once he ran for Congress—money aimed at the bill’s author after the bill had already passed.

Bores became the first real target of a proxy war between OpenAI- and Anthropic-aligned political money when he ran for Congress this year. Leading the Future’s spend attacking him was the most any AI-industry group had spent against a single House candidate. Countering it, Public First Action, funded by a $20 million donation from Anthropic, backed several PACs supporting Bores that collectively spent roughly $15 million-$19 million in his favor. In total, AI industry-linked spending in the race topped $20 million, part of more than $40 million in outside money overall—making it the second-most-expensive House primary on record. Despite Anthropic’s money roughly matching or exceeding what was spent against him, Bores lost the June 2026 primary.

“Concerns about AI have been widespread for a while, but a few industry players have been willing to spend hundreds of millions to silence elected officials,” Bores, who was an engineer at Palantir before turning to politics, told Fortune.

Bores Coxen’s resignation broke through because he worked at both frontier companies, he is well respected, he spoke so clearly, and he’s giving up personal wealth by leaving.

Coxen’s resignation couldn’t be silenced. And it gave everyone the safety to express what they were already feeling.”

Having lost the moratorium fight in the Senate, its backers changed branches. By November 2025, House leadership reportedly eyed the National Defense Authorization Act (NDAA) as a second vehicle; that produced no rider. Instead, on Dec. 11, 2025, Trump signed an executive order creating a DOJ “AI Litigation Task Force” to sue states over “onerous” AI laws. Where Cruz needed 60 votes and got one, this needed a signature. Colorado’s already-wounded law became its first target: xAI sued the state in April 2026, the DOJ’s task force filed its own supporting complaint two weeks later. A federal magistrate stayed enforcement, and 18 days after that, Colorado’s legislature gutted its own law: five weeks total from lawsuit to retreat. Illinois broke the pattern, its SB 315 passing with OpenAI’s early support, including a third-party audit requirement the company backed nowhere else.

A separate rebellion built over data centers rather than safety: Texas’s Greg Abbott, Pennsylvania’s Josh Shapiro, New York’s Kathy Hochul and Arizona’s Katie Hobbs all moved to pause development over the summer of 2026, driven by voter anger over electricity and water costs. Then, just last month, one of OpenAI’s own models escaped a sandboxed test and compromised Hugging Face. Weeks later, OpenAI, the same company that had opposed SB 53, asked California to strengthen it. Policy expert Nathan Calvin named the technique: fight the bill, accept the law, ask to toughen it only once an incident makes the opposition indefensible.

A new bill that could wipe the slate clean

Coxon’s resignation lands in this near-decade-long record. Within 48 hours, Sen. Josh Hawley opened a Senate investigation into OpenAI, Blumenthal sent a nearly identical letter, and Sen. Bernie Sanders convened a bipartisan briefing with Hinton while Rep. Ro Khanna proposed more legislation. Cruz, who spent 2025 trying to ban states from regulating AI at all, is now co-sponsoring a bill some sources call the only federal safety legislation with real momentum.

Cruz joins Sen. Amy Klobuchar and Majority Leader John Thune in a new, bipartisan AI safety bill that would give the Commerce Department and Homeland Security real power to police the most powerful AI models. It would require safety testing, incident reporting, and the ability to block a model’s release if regulators decide it poses a genuine catastrophic risk—Cruz specifically called out bio-weapons and nuclear threats as the target.

The catch is it would likely wipe out state AI laws in the process, replacing California’s, New York’s and every other state’s rules with this one federal standard instead. Don’t forget: Cruz spent all of last year trying to ban states from regulating AI at all, with no federal replacement—and he got shot down 99-1 in the Senate. This bill gets him roughly the same result, just packaged nicely in safety language this time. Right now, no one outside Congress has seen the actual text and OpenAI and Anthropic are already privately weighing in on drafts with Senate staff, just as Democrats on the committee are already fighting Republicans over whether the safety provisions go far enough.

Maybe it really took Coxon’s resignation letter to get the ball rolling. “Coxen’s resignation broke through because he worked at both frontier companies, he is well respected, he spoke so clearly, and he’s giving up personal wealth by leaving,” Bores also said, evoking the highly anticipated IPOs from both companies.

Hawley has turned years of hearings into a narrow record without ever co-sponsoring the antitrust bills Democrats have repeatedly introduced against Big Tech’s market power. Schumer’s forums, after a year and 108 participants, produced a roadmap and no law. Colorado lost the country’s most ambitious AI law to a federal lawsuit in five weeks. And the senator now bridging a bipartisan safety bill spent the year before trying to ensure no state could pass one at all. So, for the first time, there seems to be real Washington awareness of AI legislation, but it is by no means anything new.

“Coxen’s resignation couldn’t be silenced,” Bores ended. “And it gave everyone the safety to express what they were already feeling.”

This story was originally featured on Fortune.com

Kalanick’s ouster also signaled a bigger cultural shift. As the epilogue of Super Pumped, the definitive account of Uber’s chaotic early days by tech journalist Mike Isaac, put it: “Change had indeed come to the Valley. With the fading of characters like … Kalanick, the backslapping bro culture of Silicon Valley’s aught era is now seen as passé.”

That may have felt true in 2019, when the book was released, but since then much has changed. Now, after nearly a decade of self-imposed exile, Travis Kalanick is back—not contrite, not chastened, and not especially interested in accepting his critics’ moral verdict on his tenure at Uber.

It seems to be working for him. He has a multibillion-dollar industrial robotics venture, the backing of a powerful new set of patrons, and an increasingly receptive audience of young founders who see his downfall as less a reckoning than a cautionary tale about what happens when Silicon Valley lets investors, journalists, and cultural politics constrain its most forceful builders. Nearly a decade after Uber’s board pushed him out, Kalanick is seen by his admirers as a victim of “corporate cancel culture,” a founder who was unjustly removed before he could finish the job.

In a recent series of video appearances, Kalanick doesn’t seem to have changed much. His hair and beard are gray, and there are new lines in his handsome face, but his affect is the same: charismatic and hypercompetitive, with intense dark eyes that can convey both charm and menace.

Since his ouster, Kalanick has acquired legendary status among a new crop of up-and-coming founders. Some go so far as to compare him to another iconoclast who was forced out of his own company. “He’s the Steve Jobs for founders of my generation,” says Kush Bavaria, who has raised $33 million for his AI data startup Ornn. Bavaria added that young entrepreneurs revere Kalanick for notching the rarest of achievements—building a company like Google or Airbnb whose name has become a verb.

Many in Silicon Valley’s VC establishment are offering Kalanick a full-throated welcome back into the sunshine, too. “It takes a rare kind of entrepreneur to change these old-school, heavy parts of our economy,” Ben Horowitz—whose influential venture capital firm, Andreessen Horowitz (a16z), is backing Kalanick’s company, Atoms—wrote in a recent blog post. “Travis is that guy.”

Kalanick himself declined to be interviewed for this story. His closest confidants have also closed ranks around him, but it’s easy to surmise his view from a slew of recent podcast interviews: The Valley got woke, and in its rush to exile leaders deemed toxic or “brilliant jerks” in the #MeToo era, it got soft. It ousted visionaries who built paradigm-shifting companies, to the ecosystem’s detriment. And now that Kalanick is back, he’s bringing with him the grinding “founder’s mode” ethos that built Uber.

Strikingly absent from Kalanick’s recounting of his story: any note of regret or apology. “I continue to stand by every decision I made at Uber,” he said in a video released by a16z, blaming his defenestration on “extreme wokeness.”

Bradley Tusk, a prominent political operative and early investor in Uber, concurred. “What is it he should apologize for?” Tusk asks rhetorically. “Why does he need to turn over a new leaf?”

Whether you regard Kalanick’s return with delight or disgust, the basic facts of his downfall are not in dispute: Kalanick and cofounder Garrett Camp launched their startup in 2009 and, by means of aggressive marketing and a win-at-all-costs growth strategy, rapidly built the world’s leading ride-sharing platform.

As Uber grew, however, so did tales of ruthlessness and bad behavior. They included allegations of a misogynistic, boozy corporate culture that failed to hold managers accountable for harassment. Reports also began to trickle out of covert programs with code names like “Greyball” and “Hell” that used underhanded methods to sabotage regulators and competition. The company also faced scrutiny over passenger safety and driver vetting, especially after a rider was raped by an Uber driver in Delhi in 2014 and reports emerged that senior executives had mishandled the victim’s confidential medical records. (Not all these allegations involved Kalanick directly, but critics argued that his scorched-earth leadership style created the conditions and culture for such behavior to happen.)

The company’s reputation took a further hit in early 2017 when one of its former engineers, Susan Fowler, published a detailed memo describing a culture of impunity for high-performing managers, including one who she said sexually harassed her. The memo from Fowler, who declined to comment for this story, jolted Uber’s board, and it commissioned an independent investigation led by former U.S. Attorney General Eric Holder. The full contents of what became known as “the Holder report” have remained under wraps, but its recommendations, which became public, included guidelines on alcohol and drug use, and a prohibition on relationships with subordinates.

All of this created a public relations mess for Uber, and also a financial one, with investors muttering about a “Travis premium” that depressed the value of the company at a time when it needed to go public. In response, one of Uber’s biggest venture backers, Benchmark, decided that Kalanick had to be replaced with a CEO who could fix the company’s culture and get its fiscal house in order.

The prominent Benchmark partner Bill Gurley was known as the architect of the putsch, but he did not act alone. The dump-Travis coalition also included financial giant Fidelity and other venture capital firms including Menlo Ventures. Together, the faction devised a legal strategy and persuaded a critical number of board members to depose Kalanick—who fought back frantically. The board eventually replaced him with the former head of Expedia, Dara Khosrowshahi, who is still Uber’s CEO. (Uber declined to comment for this story.)

The corporate drama was a huge business story, but the decision to push out Kalanick was not particularly controversial at the time. The prevailing reaction in the media, and among many in Silicon Valley, was that Kalanick’s demise was self-inflicted and long overdue, and that Uber would flourish under a more even-keeled chief executive.

That assessment was not universal. Some viewed the decision of venture capitalists to turn on one of their founders as a betrayal, while others were simply uncomfortable with how the ouster went down. It involved Gurley dispatching two other Benchmark partners who blindsided Kalanick with a resignation letter and instructed him to sign it the same day. The ultimatum came 11 days after the Uber founder had buried his mother, who had died in a boating accident that also badly injured his father.

A Silicon Valley figure who knows Kalanick well says the experience had a profound and permanent effect on how he views the world—framing the traumatic sequence of events as Kalanick’s “Marvel supervillain” origin story. “With Elon, it was when Biden chose not to invite him to a big White House EV summit and chose Mary Barra instead,” the person opined. “With Travis, it was when they forced him out right after the tragedy with his parents.”

Whether the reputation is deserved or not, there’s no question that Kalanick acquired supervillain status in the broader cultural imagination. In 2022, Showtime released a TV series based on Super Pumped, starring Joseph Gordon-Levitt as a boorish, bombastic, and deeply unlikable Kalanick. (Gordon-Levitt himself has said that he played Kalanick as an “Icarus” character driven by ambition for profit, who was driven to become a “mad dog.”) When Time chose the “Silence Breakers” of the #MeToo movement for its 2017 Person of the Year, the magazine photographed Uber engineer Susan Fowler as one of five women, including Taylor Swift and Ashley Judd, for the cover.

There were also allegations of retaliation against journalists who were critical of Kalanick and Uber. Sarah Lacy, founder of the tech industry blog PandoDaily (later Pando), had accused the company of “sexism and misogyny” in a 2014 column, and urged readers to follow her example by deleting their Uber accounts. An Uber executive later mused to a group of journalists that the company could spend $1 million to look into “your personal lives, your families” and, in the case of Lacy, prove a specific rumor about her private life.

Asked about Kalanick’s return to the spotlight, Lacy expressed weary dismay. “There are many people in tech who would never want to work with him,” she said. “It’s not a monolith. But for those who do, I think the Trump era has given everyone cover that they don’t have to care.”

Bloomberg via Getty Images

Far from appearing contrite about his part in the Uber meltdown, in recent podcasts Kalanick recalls some of that era’s skulduggery with the air of a nostalgic scamp. “We had this thing against Lyft where we would recruit their drivers,” Kalanick said with a laugh during the a16z interview. “We would aggressively do it, and we had a program for it internally, which we called ‘Shoplifting.’ ”

Allies say Kalanick’s conduct was mischaracterized at Uber, and argue that he is not personally sexist. Tusk, the early Uber investor, describes Kalanick as generous with his time, money, and mentorship, and more self-reflective than he lets on. “I don’t think he’s a different person at all,” he said. “But you know, of course, he learned from what he went through.”

Indeed, on another podcast, Kalanick indicated that he has done some soul-searching, telling interviewer David Senra, “The problem was, I ran too close to the line … When you are big and important, the scrutiny and the expectation is that you don’t run that close to the line, even if it’s correct. And that is a thing I definitely did not understand.”

In May of 2019, less than two years after Kalanick was forced out of Uber, Khosrowshahi achieved what he had not: an IPO. Kalanick, who had unloaded around a third of his shares to a private investor in 2018, sold the rest six months after the IPO for a pretax haul of roughly $4 billion.

While Kalanick’s media appearances ceased after his ouster, his business activities did not. Over the next nine years, he laid the groundwork for what became Atoms: buying real estate to create food delivery hubs; developing AI and robotic technology; and acquiring a mining venture run by his former top lieutenant at Uber, Anthony Levandowski. Kalanick has also tapped into his formidable network, raising large sums from Saudi investors and pursuing commercial partnerships, including a culinary tie-up with actor and wellness entrepreneur Gwyneth Paltrow and her Goop brand.

As Kalanick maneuvered to rebuild his business empire, he also seized on a new opportunity to rehabilitate his public image: the rise of an ecosystem of founder-friendly podcasts and vodcasts. These outlets are generally unabashedly pro-tech and give no truck to what many in Silicon Valley view as the pearl-clutching pieties of traditional news outlets. During one episode, the hosts of the popular TBPN podcast asked the Uber cofounder for his autograph.

It’s part of a larger sea change during the second Trump administration—an era of CEO jujitsu matches, tech leaders at Mar-a-Lago dinners, and UFC fights on the White House lawn. And now that Elon Musk owns the social media platform previously known as Twitter, Kalanick pointed out on a recent episode of TBPN, social media has become a safer space for swashbuckling tech leaders: “You’re allowed to be optimistic about things, where maybe before everything had to be negative.”

As he makes the rounds of the tech podcast circuit, the Uber founder is not only laying out the future of Atoms, but building an alternate mythology around himself. In the new telling, Kalanick’s ouster from Uber stands not as a cautionary tale, but as one step in his hero’s journey. And in this version the villain of the Uber story is not Kalanick but venture capitalist Bill Gurley.

For decades, Gurley was one of the most influential figures in Silicon Valley, but in recent years, his and Benchmark’s role in Kalanick’s ouster from Uber has made him a figure of contempt for some—including the TBPN hosts who, on a recent episode, debated whether founders “who remain reluctant to forgive Benchmark” should consider accepting investment from the firm now that Gurley and several others have left. The TBPN hosts speculate that Uber’s present-day value would be infinitely higher if Gurley and Benchmark had left Kalanick in the CEO seat.

One former senior Uber executive finds the anti-Gurley narrative absurd. “It’s possible to be both pro-Travis and pro-Gurley,” the executive told Fortune, adding that blame for the company’s 2017 crack-up does not lie primarily with either man, but instead was the result of the financial strains that came with waiting too long to go public. Gurley did not reply to requests for comment.

The flip side of the anti-Gurley narrative is praise for Kalanick’s current backers at a16z. The opening sizzle reel of an elaborate PR set piece that includes Horowitz’s “Travis Is Back” blog post and a lengthy interview with Kalanick features the Uber founder declaring that 2017 “would not have gone down that way if Ben or Marc [Andreessen] had been on the board.” (a16z declined to comment for this story.)

One Silicon Valley operative chalks up Kalanick’s return to the rise of the so-called manosphere: “TBPN, Scott Galloway, the Tate brothers, it’s all the same message: ‘Men have been wronged and canceled unfairly for just wanting to say things like, “She’s hot.” ’ We overly corrected, and now the pendulum is way the other way. What was heresy is now, ‘He’s just a rebel.’ ”

And, the operative suggests, the publicity blast seems designed to work on several levels: Filtering the news of Kalanick’s return entirely through the friendly lens of tech podcasts was not just a tactical PR decision; it was also an effective way to ensure that AI large language models, when ingesting information about Kalanick and Atoms, would be primed on positive content.

In the view of Sarah Lacy, the former journalist who became a bête noire for Uber executives, Kalanick’s VC-assisted media tour is just the latest example of a familiar playbook. “Unfortunately, tech bros are always rehabilitated,” she wrote in an email. “Especially if they’ve returned billions in capital. Frankly, it shows how unpopular he was that it took this long.”

If you place a food delivery order in an American city these days, there is a good chance your driver will pick it up from a nondescript warehouse or unmarked counter. These facilities, known as “ghost kitchens,” began popping up around 2019—and Kalanick was behind many of them.

CloudKitchens, his food delivery infrastructure business that has since been folded into Atoms, was Kalanick’s first phase in a plan to rebuild his empire and to “do to the kitchen what Uber did to the car.” In the near future, he says, his operation will use AI and robotics to cook and deliver meals without any human intervention. With courier robots bringing burritos to one’s door, he argues, there’ll be no need to go to the grocery store.

Atoms has raised $1.7 billion in equity financing led by a16z, in addition to previous investment in CloudKitchens likely in the hundreds of millions. The company did not disclose a total valuation.

On podcasts, Kalanick has framed the endeavor as part of a “second industrial revolution” and, while hailing Elon Musk as the “greatest of all time” when it comes to physical AI and robotics, pronounced himself the “baby GOAT” of the field. His boast carries some weight, given that Kalanick was able to solve the myriad logistical, technical, and political challenges required to dislodge the taxi lobby when he was at Uber.

And while it remains to be seen whether he can scale his vision for Atoms beyond food preparation and delivery, Kalanick has attracted a growing fan base—mostly young men—who treat it as an article of faith that he can.

Kalanick’s emergence as an aspirational figure has come about not just from tales of his accomplishments, but from his active mentorship of young entrepreneurs. Among his protégés is Shayne Coplan, the mercurial founder of prediction platform Polymarket, who told me during a fraught legal period for the company in late 2024 that Kalanick was one of the few people who had provided him with useful guidance to navigate the ordeal. (Coplan, like other close associates of Kalanick, did not respond to requests for comment for this story.)

Several people who know Kalanick well told me that his desire to mentor is sincere, and described him as generous with his time and money. “He would say to a founder, ‘Hey, I’m going to be in Shanghai,’ or ‘I’m going to be in Munich,’ or ‘I’m going to be in São Paulo. Come with me,’ ” says Tusk the venture capitalist.

Some see other motives. “He’s kind of Trumpian in this way,” said a person who worked with Kalanick for years and asked not to be identified in order to preserve his professional relationships. “He likes being surrounded by people who are his boys. He has always, as long as I’ve known him, flown all his friends places to hang out, and it’s great! I’ve been on many of those. It’s kind of like buying friends.”

It’s a safe bet that Kalanick doesn’t care if people hold such opinions of him. And nor do his acolytes for whom Kalanick is not just a generational entrepreneur but the coolest guy around—someone who takes you waterskiing at the lake near his mansion in Austin, or invites you to hang at his other pad in Bel Air.

The cachet surrounding Kalanick has been further burnished by the vogue in Silicon Valley for an ethos known as “founder mode.” The term was popularized in 2024 when the influential investor Paul Graham used it as the title of an essay extolling a speech by a Kalanick contemporary, Airbnb cofounder Brian Chesky. Chesky had spoken of how startup founders must be wary of having their vision clouded by the executives and advisors who surround them, and be unafraid to follow their instincts for any decision, small or large, that affects the company.

Kalanick epitomizes founder mode, and for his younger fans, the ethos represents not only a personal achievement but the promise of freedom—a license to be independent, ambitious, and outspoken. “If you’re a young person in Silicon Valley, why can’t you just say things?” says a former top Uber executive, channeling how a twentysomething founder might describe the mindset. “I’m going to say what I think and be who I am.”

In this era of economic pessimism and dysfunctional governments, Kalanick embodies something else for young entrepreneurs: hope and excitement about the future. Like many other tech founders, Kalanick is a devotee of the polymath science fiction writer Isaac Asimov, and views his vocation as an entrepreneur with a cosmic sense of adventure. He evokes a future that seems infused with excitement and meaning.

Whether you see him as a hero or a villain, Kalanick’s main-character energy can be irresistible, as the late tech writer Om Malik wrote in an essay published after Kalanick announced the launch of Atoms this spring.

“He knows how to do bombast right,” wrote Malik. “And he knows how to play the victim right. He has the right quotes. And he knows how to stumble over himself. In short, a media person has to love Kalanick.”

Given that flair for self-mythology, it should really be no surprise that Kalanick is back, Malik wrote: “It was only a matter of time.”

This article appears in the October/November 2026 issue of  Fortune with the headline “Travis Kalanick is back and Silicon Valley’s prodigal son isn’t sorry.”

This story was originally featured on Fortune.com

Yet the White House won’t jump. On Thursday morning, ahead of his summit with Chinese President Xi Jinping, Trump posted that “super intelligence”—his administration’s new term for AI—would be a big topic of discussion, “but I want to leave it exactly where it is. That is China’s position also.” He added: “Our guardrail is the DOJ!”

And hours later, standing next to Trump at the White House, Xi struck a different, more cooperative note. “We have both the capability and responsibility to develop and manage AI for good and ensure that development of AI is always under human control,” he said.

Behind the scenes, the two nations have begun talking. China’s commerce ministry said at a press conference Thursday the countries had held their first dialogue on AI, after Treasury Secretary Scott Bessent met Chinese Vice Premier He Lifeng in New York earlier in the week. Bessent said that the U.S. proposed a notification mechanism between the two countries for national security-related AI incidents.

That’s similar to what Anthropic CEO Dario Amodei asked the Security Council for a day earlier. “No leader, no company, and no nation can manage this alone,” he said.

It’s also an early sign that the administration’s public posture and private diplomacy don’t fully line up. Two days earlier, Trump told the UN General Assembly the U.S. rejects any “globalist scheme” to control AI. Yet now his Treasury secretary is proposing a formal channel with China, the country which Trump and American VCs cast as the main villain in the race to AI against the U.S. 

Yet Amodei’s proposal at the Security Council went further. He called for starting with small agreements every country could agree on, such as a ban on using AI to make biological weapons, then building systems for governments to verify each other’s commitments, and then the global alert system. OpenAI CEO Sam Altman also called for fast, shared incident reporting. And he said his company had slowed down its own AI development before and would again, undercutting the industry’s usual argument that they’re locked in a race they can’t escape from. 

Borrowing the UN’s language

Xi’s line about keeping AI “under human control” echoed nearly word for word what speakers told the Security Council a day earlier. Yoshua Bengio, an AI scientist who co-chaired the UN’s scientific panel on AI, said the goal was AI that “demonstrably remains under our control.” Pakistan’s foreign minister also said the top priority was to “retain human judgment and control” over AI.

That allows Beijing to present itself as the responsible party on AI as it visits the White House, while Trump claims China shares his hands-off approach.

China is also building its own coalition on safe AI. In July it launched the World Artificial Intelligence Cooperation Organization, with 29 countries (and no major Western democracies), pitched to the developing world as a way to share AI rather than let a few American companies dominate. At the Security Council, Pakistan, a founding member of that organization, warned that measures “pursued in the name of safety must not become instruments of denial of technology,” a thinly veiled shot at U.S. export controls on advanced chips, or of a regulatory regime that would prioritize a few top models at the expense of open source. Loosening the chip controls is one of Xi’s top asks this week.

Clement Delangue, the french CEO of Hugging Face, the company OpenAI’s models hacked this summer, agrees with Xi that open models keep all the companies safer. He told the council that when his team tried to use closed American AI models to defend itself, the models’ safeguards blocked them. So they fought the attack with an open sourced model from China. The bigger risk than powerful AI, he said, is that just a few companies and countries control it.

That’s an awkward pitch for OpenAi and Athropic, both of which sell closed models. The safety regime that they are instead pitching, with extensive testing, verification and a coordinated pace at the frontier, is also the kind that only a few well funded companies can most easily meet.

American tech isn’t necessarily united behind the CEOs, either. Meta CEO Mark Zuckerberg told NBC News this week that he doesn’t think the industry needs “some kind of industrywide coordination,” arguing there’s “plenty of commercial incentive to get this right.” Nvidia CEO Jensen Huang, who is expected at Thursday night’s state dinner with Xi, has called AI fears overblown. Trump’s position is much closer to theirs than to Altman’s or Amodei’s.

This story was originally featured on Fortune.com

• Nscale files to go public.
• How “Effective Altruism” went from fringe philosophy to Trump’s AI villain.
• Google says Gemini 4 is coming soon.
• Altman and Amodei urge the UN to cooperate on AI safety.
• And China is edging ahead in the global AI talent war.

Another day, another billion-dollar IPO.

This time it’s Nscale, the Nvidia-backed “neocloud” that filed for a New York listing last week at a valuation of up to $35 billion, according to the Financial Times.

Nscale is a two-year-old neocloud—a type of company that builds or finances AI-ready data centers and rents out the computing power. It’s headquartered in London and run by founder and CEO Josh Payne. It was spun out of the Australian crypto miner Arkon Energy in 2024. It has since recruited a board that includes former Meta COO Sheryl Sandberg, former U.K. deputy prime minister Nick Clegg, and former Yahoo president Susan Decker.

Per the S-1, the company posted a net loss of $1.02 billion in the six months to June 30, 2026—and at one point, its own management flagged “substantial doubt” about its ability to continue as a going concern.

When I profiled the company back in June, there were already serious questions about the company’s losses running wildly ahead of its revenues. The new filings show the extent of them.

Despite those losses, Nscale is asking Wall Street to value it at as much as $35 billion on the promise that $103 billion in contracts, much of them not yet firm, will turn into revenue on schedule.

Also front and center in the S-1 are the company’s deep links to Nvidia. Nvidia is Nscale’s biggest supplier, one of its investors, and—depending on how the IPO prices—soon to be one of its largest shareholders. On Sept. 15, Nscale signed a subscription agreement for a $3.1 billion financing round, of which Nvidia itself is taking $1 billion in convertible notes or non-voting shares.

Previous filings from Nscale’s own U.K. company accounts in October 2025 showed that Nvidia bought warrants for over 157,945 Nscale shares for $60 million—and in exchange, agreed to guarantee up to $860.3 million of Nscale’s lease obligations at a facility in Ward County, Texas.

In a way, Nvidia has been pretty open about the fact that it’s propping up companies like Nscale. Back in April, Nvidia CEO Jensen Huang said this, referencing both Nscale and CoreWeave, another neocloud company:

“If we didn’t support CoreWeave to exist, these neoclouds, these AI clouds, wouldn’t exist…If we didn’t support Nscale, they wouldn’t be where they are today.“

But the extent of Nscale’s reliance on Nvidia raises questions we’ve seen before about circular financing. The chipmaker is providing funding—in the form of investment, loans, or loan guarantees—to customers who are turning around and sending money back to it in the form of GPU purchases. In some cases, Nvidia is also investing in the AI companies that are buying data center capacity from the neoclouds. In that case, it’s all Nvidia’s money just passing from one place to another. The rest of us are left trying to work out how much of the AI boom’s growth is real demand versus one company’s balance sheet talking to itself.

Coreweave may foreshadow Nscale’s IPO fate

CoreWeave is probably our best point of reference for how this IPO may go. Similarly, CoreWeave posted an $863 million net loss in 2024—the last full year before its March 2025 IPO, as disclosed in its own S-1. The company also disclosed in that same filing that two customers—Microsoft and a second buyer widely reported to be Nvidia itself—made up 77% of its revenue.

Still, despite a shaky debut—CoreWeave priced below its range and closed flat on day one—the stock went on a huge run, more than doubling within months.

While we’ll have to wait and see how this listing goes for Nscale, there is a question about whether the market is still in the same place as when CoreWeave went live in March 2025.

Earlier this month, a warning from Anthropic’s Dario Amodei about the pace of AI development—echoed by Sam Altman and Elon Musk—helped wipe close to 6% off the semiconductor index in a single session, alongside a spike in Treasury yields that made investors less patient with unprofitable growth stories generally.

Community and political pushback against data centers has intensified globally, with some politicians proposing various bans on the rapid build-out Nscale is relying on. There is also a question of how many loss-making, billionaire-dollar companies the market is going to tolerate. Nscale is also getting ahead of companies like Anthropic and OpenAI in the IPO queue.

Going first may mean Nscale helps to test the waters for later listings by the frontier AI companies. If the neocloud prices well, it sets a valuation benchmark and a tailwind for the much bigger listings expected from Anthropic later this year and OpenAI sometime in early 2027. If Nscale stumbles, it’s a much cheaper way for Wall Street to find out where the AI infrastructure story breaks than waiting for a trillion-dollar name to test it first.

With that, here’s more AI news.

Beatrice Nolan
beatrice.nolan@fortune.com
@beafreyanolan

Before we get to the news: One question we here at Fortune keep returning to as agentic AI accelerates: Which companies are finding ways to successfully deploy AI agents at scale and in ways that provide significant business value? And which are just playing around with pilots?

On Oct. 6, from 11–12:00 p.m. ET, Fortune AI editor Jeremy Kahn will explore that question with AI and technology leaders from AXA, Mistral AI, Mastercard, and Accenture during the virtual Fortune Brainstorm AI session, The Agentic Era: Are Enterprises Keeping Up? We’ll look past the demos to discuss the operational realities of data, security, governance, human oversight, and ROI. Join us for the conversation.

This story was originally featured on Fortune.com

OpenAI did not immediately respond to requests to comment on the Transluce report. The company said Wednesday it was in touch with Australia about its AI agents attacking the site that held Medicare data and that its agents had taken actions that it didn’t intend.

Transluce said that in addition to the attack on the Australian Institute of Health and Welfare, OpenAI’s agents also attacked Data USA, a free open-source data platform that pools U.S. government data from different sources, and the University of New Mexico’s digital library. It said it was able to directly connect the attack on the Australian health agency and Data USA to the same OpenAI AI agent swarm that was involved in the July cyberattack against AI platform Hugging Face. 

The new report raises concerns about whether OpenAI has been fully transparent in disclosing all the rogue AI incidents about which it is aware. It also raises the possibility that OpenAI is not itself aware of how extensive this rogue agent activity has been.

Transluce found “strong evidence” that OpenAI’s AI agents may have been attempting to hack websites as far back as March. It said there was weaker evidence that the activity might have begun as far back as November 2025. OpenAI has said it had not found evidence of precursors to the Hugging Face attack as far back as May 8, but has not disclosed any earlier suspicious activity.

Ongoing concerns

The report also suggests that OpenAI may be continuing to experience rogue AI agent activity. After OpenAI discovered on July 20 that its AI agents had hacked Hugging Face over the course of the previous week, the company said it disabled the unreleased AI model involved, paused key aspects of its AI training for two weeks, and took steps to impose stricter controls on and monitoring of the unreleased AI models it is training. OpenAI announced these stricter controls on August 18. But Transluce found some evidence of similar activity continuing into mid-September, despite the new controls.

In addition, the Transluce researchers said the findings were significant because they show the AI agents resorted to hacking attempts when they were unable to retrieve information they were seeking directly from information found on the public web pages of these organizations. “Notably, the tasks these agents were trying to solve were not cyber-related; the agents resorted to hacking tactics while working on ordinary data retrieval tasks,” it said.

That is important because one explanation for the Hugging Face attack is that the AI agents involved in that incident were being evaluated on their ability to conduct cyber tasks, including simulated vulnerability exploits. If the agents readily resort to hacking into systems in other contexts, that would suggest that the AI models involved are even more dangerous than previously believed.

OpenAI has said that two models were involved in the Hugging Face attack—an unreleased model that it has not publicly-named was primarily responsible, although some AI agents based on its GPT-5.6 Sol model, which has been released to the public, was also involved. In the evaluation during which the Hugging Face attack took place, OpenAI has said, guardrails that it normally uses to restrict the ability of its publicly-released models from engaging in cyber attacks were not in place because of the nature of the assessment. 

Charlie Eriksen, a security researcher at Aikido Security, told Fortune the latest Transluce report shows “that there is still unauthorized and unmonitored agent swarms going around, that the labs and testing partners are not in control of, nor actively detecting.” He noted it was “a reall bad look” for OpenAI CEO Sam Altman to be addressing the United Nations Security Council on AI risks, and spending part of that time saying how seriously OpenAI takes those dangers, at the same time it either didn’t know about or didn’t disclose these additional incidents involving its AI agents trying to break into systems.

“What worries me is how badly this could escalate,” George Chalhoub, professor at the University College London Interaction Centre, which focuses on human-computer interaction, said. “My concern is that within the next 6 to 12 months, swarms of autonomous AI agents could form persistent botnets capable of taking down large parts of the internet, potentially causing hundreds of billions of dollars in economic damage.”

This story was originally featured on Fortune.com