Local government resistance to data center development accelerated this summer, a warning sign for big AI companies such as Anthropic, Google and OpenAI that are pinning their compute ambitions on data center expansion.
In July alone, more than 150 towns and counties passed temporary or permanent bans on data centers, many adopted in emergency meetings. That brings the nationwide total to more than 500 active data center bans at the start of August, according to The Information’s analysis of thousands of legal documents and local news reports. When we first published our tracker of these bans in late June, we had found more than 300.
The coming initial public offerings of AI giants Anthropic and OpenAI could spell a windfall for San Francisco real estate agents, private jet companies—and wild animals.
As AI employees with fast-appreciating stock look to start cashing out, philanthropies are talking to potential donors and preparing for a possible influx of cash. And because many AI researchers have ties to the effective altruism movement, which takes a particular interest in causes that mainstream philanthropies overlook, the organizations prepping for a windfall might look very different from previous charitable waves.
The West Seattle hillside home of Richard Rhodes, with views overlooking Puget Sound, is a good place to ponder the end of the world. Walk up a series of steps, through the front door and enter Rhodes’ home office, where his Pulitzer Prize hangs. He won it for his 1986 book “The Making of the Atomic Bomb,” a sprawling, 900-page account of the scientific discoveries that led to one of the greatest weapons ever created.
I made the trip to Rhodes’ house because his 40-year-old tome has become essential reading for employees at Anthropic and their leader, Dario Amodei, who counts it as one of his favorite books. Amodei has said he identifies with a central character in Rhodes’ book: Leo Szilard, the physicist who first conceived of the nuclear chain reaction and then aggressively advocated to keep it under human control.
The Anthropic CEO’s affinity for the historic account isn’t surprising if you read the profile of Amodei that my colleagues and I wrote this week. It traces how Amodei has long grappled with the weight of the technology he was building. He used air-gapped computers, machines disconnected from the internet, to compose early OpenAI memos to prevent remote hacking.
He worried about getting kidnapped in China, a U.S. adversary he feared was after key AI research secrets. These days he concerns himself with the security of the data centers that house the clusters of chips training Anthropic’s next frontier-chasing model. He and his sister, Daniela, another Anthropic cofounder, grew up wanting “to save the world together,” Dario has said.
Rhodes, 89, has never met Amodei. But he and his wife traveled to downtown San Francisco late last year to speak to Anthropic’s employee book club, which had ordered 100 copies of his book. Rhodes said he gave them a perspective that appears to cut against the “once-in-a-civilization” technology moment that Amodei and employees think they are building. Rhodes, who has also written about the history of technology changes, said he has watched the AI wave with a sense of déjà vu.
“The first thing that struck me most of all was what always strikes me with new technologies: the moral panic side. The general sense that this is going to destroy the world. It’s come up every time,” he told me. He went back in history. When 17th century Brits started using coal as fuel for the first time, preachers called it the “devil’s excrement,” he recalled. The rise of the automobile, last century, also stirred protests, he said.
“We get a new technology. It has its dark side and its bright side, and we slowly surround it with social, legal, moral, cultural controls until we turn it into a human artifact,” he said.
That might not be a stance that Amodei and his employees agree with. Amodei has pegged a 25% chance of AI going “really, really badly” for civilization. A big part of Anthropic’s technological and commercial success stems from the employees’ and Amodei’s deep belief that they are important, historic actors in an existential scientific historical moment. Its sensitive safety culture is an essential part of it.
Maybe they’re right, particularly as a series of cybersecurity attacks from AI agents are raising alarm bells recently, and the rest of us will be glad we had such responsible (and newly wealthy) stewards along the way.
Rhodes, meanwhile, has a different, perhaps more optimistic perspective. Rhodes is now working on a work of fiction about robots who take over the world, and all that’s left on Earth is a group of children. But the book is a comedy. “The robots are so stupid and the children are so bright,” he said.—Cory Weinberg
Apple’s ‘Confidential’ Files
On Monday of this week, Aaron Tilley wrote a colorful story for us that sought to clear up a mystery: Why do so many former Apple employees end up with access to confidential Apple documents after they leave the company? The answer, he reported, is that the company encourages new hires to mingle their personal and workplace digital identities using iCloud. That practice makes it hard for them to expunge all of their Apple files when they leave the company.
The story was prescient. In the days that followed its publication, an OpenAI blog post and filing in its legal battle with Apple described the very same iCloud issues at the heart of Aaron’s story. OpenAI argued that those issues, among other things, undermine Apple’s accusations that two of its former employees stole trade secrets when they went to work for OpenAI.
“For years, Apple encouraged employees to use personal iCloud accounts for work, intermingling company and personal data, while failing to manage access cleanly across multiple Apple systems when employees departed,” OpenAI wrote in a motion to dismiss Apple’s suit. “Those practices created precisely the confusion and unwanted access issues that Apple now characterizes as intentional misappropriation.”
Apple previously told Aaron the iCloud issues had nothing to do with its allegations in the OpenAI case. We’ll see as the case plays out. Meantime, maybe don’t intermingle your work and personal online accounts just to be safe.—Nick Wingfield
Since Sequoia Capital’s former leader, Roelof Botha, left the storied venture capital firm late last year, its two new co-leaders, Pat Grady and Alfred Lin, have been on an investing tear. Their goal: Reestablish Sequoia’s top dog status in the AI age in the face of aggressive competitors.
If you’re a scrappy startup building humanoids or quadruped robots in the U.S., you’ve got little choice but to source parts for your prototypes from China. That’s happening at a time when Chinese-made robots are becoming an increasingly sensitive geopolitical issue for the Trump administration.
Blair Braverman has enjoyed her fair share of adventures—racing in dog-sledding treks like the Canadian Challenge (300 miles), the Kobuk (440 miles) and in the granddaddy competition of mushing, the Iditarod (1,049 miles). Braverman, a contributor to The New York Times, Outside magazine and other publications, is also pretty good at recounting other people’s adventures, which she does in her new podcast, “What to Carry, What to Burn.” And it is a testament to her innate likability as a storyteller that she does the pod mostly as a solo narrator—just her recounting various tales of perilous, remarkable survival. So far this has included an early 20th century Arctic expedition to Wrangel Island, a trek that involved, truly, an unpaid internship from hell; a 1970s plane crash in the Sierra Nevadas; and the Salvadoran fisherman lost for 13 months in the Pacific Ocean a decade ago, who was later sued for cannibalism (a charge he denied). I appreciated the little details that only a seasoned explorer like Braverman could add to the episodes, like the fact that well-made igloos can indeed get pretty toasty (well past 75 degrees!) and that bear just tastes pretty similar to beef.—Abram Brown
As the world’s tenth largest economy, Brazil exports great quantities of soybeans, coffee and sugar. It has also introduced the world to many other wonderful things, like caipirinhas. But there’s something else that the country has largely managed to quietly keep and enjoy to itself: the clever, demented talent of thriller-horror writer Raphael Montes, who has been a mainstay on Brazilian bestseller lists for over a decade but remains a non-entity abroad.
“People have told me that if I were American, I‘d be a millionaire by now—and I just tell them, ‘Yeah, but then I wouldn’t have Carnival,’” Montes told Folha de São Paulo, Brazil’s largest newspaper, last year. “When a Brazilian author gets translated, their defining trait in the eyes of foreigners is usually just being ‘exotic.’ What I would truly love is to be recognized globally simply as a great thriller author—who happens to be Brazilian.”
His wish may well come true very soon. The 2016 book that first cemented his fame in Brazil, “The Secret Dinner,” has been translated into English for the first time and debuts in the U.S. on Aug. 18. As a gruesome dark comedy, I do not suggest reading it while eating.
The novel’s title refers to a scheme hatched by four broke young men in Rio de Janeiro: catering an ultra-luxe cannibalistic dinner. (It’s “Rope” meets “Sweeney Todd” meets “Chef’s Table.”) Initially, they envision the soiree as a one-off ploy. But as you might expect, they attract a crowd of unsavory types, including a desperate heir of a prominent family who sees the evening as a great opportunity for capitalistic expansion, with himself as the, well—venture investor. Soon, he has strong-armed the boys into running a thriving underground supper club. One of the foursome, Miguel, is a gourmet sous-chef, so he does the cooking. (Where Mrs. Lovett had her pies, Miguel prepares braises, terrines and ossobuco.) The narrator, Dante, an underemployed bookshop clerk with a fondness for weed and Klonopin, takes up maitre d’ duties. The butchering, meanwhile, falls to a chainsaw-wielding girlfriend.
In depicting the wealthy of Rio as rapacious man-eaters, Montes makes obvious his distaste for how his country has grasped for prosperous modernity: Brazil, with its enormous economy, has the third largest disparity between rich and poor on the planet. (“Money is all that matters,” the villainous heir says at one point, practically twirling his moustache. “Nobody gives a damn about the rest.”) To Montes’ credit, he manages to make the point and still delivers a taut page-turner, one complete with a twist that hits like a double-strength caipirinha.—A.B.
Many films correspond to a specific time of year. Christmas movies abound, and the scaries populate October. “Ferris Bueller’s Day Off”? Back to school. “The Graduate?” End of school.
“Arco,” an imaginatively animated feature film from 38-year-old French director Ugo Bienvenu, corresponds to a point on the calendar with increasingly eerie, grim importance: forest fire season, making it a timely watch. (It initially came out last year, caught an Oscar nomination and recently got to Disney+.) The titular character, Arco (voiced by Juliano Krue Valdi in the English dub), is a boy from our far-distant future—the year 2932—where mankind has mastered time travel. It involves Superman-esque flight, with rainbows left in a time-traveler’s wake just as real-life planes emit plumes of white.
But Arco hasn’t learned how to fly, and when he impishly steals his older sister’s flying suit, he crashes in our near-distant future—the year 2075. In the accident, he loses the suit’s power source, a diamond-shaped crystal; desperate to find it, he enlists the help of Iris (Romy Fay), a girl around his same age. They are pursued by a trio of stooges, Dougie, Stewie and Frankie (Will Ferrell, Andy Samberg, the musician Flea), who previously witnessed another person traveling via rainbow and want proof that time travelers exist.
2075 is no high-water mark for humanity, and the adventure is made more dire when a forest fire engulfs Iris’ suburban town, and the pair flee along with Iris’ robot au pair, Mikki, who has been raising her in her parents’ absence. In this dystopia, adults do not do remote work—they do remote parenting: With their office in a distant city, Iris’ mother and father check in on her occasionally via hologram. (They’re voiced by Mark Ruffalo and Natalie Portman, and in an inventive touch that seems entirely like something that might actually happen in future robotics, the film uses a digital blend of their voices for Mikki—the nanny bot a combination of father and mother.)
“Arco” issues many of the stern environmentalist warnings sounded by Studio Ghibli films, and like a Miyazaki film, the high emotions and dramatic stakes of “Arco” give it a distinctly grown-up register, presenting a prismatic-colored fable that winds to a poignant denouement.—A.B.
Nvidia has agreed to invest $2 billion into Lancium, the power infrastructure developer behind the OpenAI and Oracle AI campus in Texas, and has agreed to commit another $1 billion as the developer secures additional planned power, three people familiar with the transaction said. The deal values Lancium and its portfolio of land and power connections at around $10 billion in enterprise value, which includes the investment and its debt.
The $2 billion would give Nvidia a roughly 20% stake in the Blackstone-backed power developer. Nvidia’s stake could grow to around 30% with the additional $1 billion once more of Lancium’s campuses hit certain thresholds, including grid hookups, the people said. The deal is intended to help Nvidia secure multiple gigawatts of pending power rights for its chip customers’ data projects at a time when energy for AI data centers is becoming harder to come by.
Cursor told staff at an all-hands meeting on Thursday that SpaceX could complete its $60 billion acquisition of the coding startup as soon as the end of next week, and flagged that the Cursor brand name will likely be phased out for new products in the coming months, people familiar with the matter said.
For instance, Cursor’s upcoming general-purpose agent, internally code-named “Sand”, could be branded as Grok Bot, the people said. Cursor doesn’t have plans to immediately change the name of its existing tools, including its popular Cursor coding assistant, one person said. The exact timing of the deal closing depends on securing regulatory approvals, though it is expected to close by the end of the month at the latest, they said.
Legal AI startup Harvey is in talks to raise at least $500 million at a $15.5 billion valuation including the investment, a 40% premium to its last valuation five months ago. The fundraising follows a revenue surge at the legal AI startup, according to a person with knowledge of the fundraise.
Four-year-old Harvey is currently generating more than $350 million in annualized revenue, up more than 80% from the $190 million in annualized revenue it was generating in January, according to the person. Lightspeed Venture Partners has been interested in leading the round, the person said.
On a muggy morning in April, a venture capitalist from San Francisco arrived in China’s southern tech hub of Shenzhen with a long shopping list. His friends at U.S. robotic startups had asked him to bring back humanoids and various robotic components from China.
He headed to Shenzhen’s Huaqiangbei district, the world’s biggest electronics bazaar. At one store, he bought humanoids, which he arranged to ship back to the U.S. He then went upstairs to component floors—a labyrinth of countless stalls barely larger than walk-in closets—to pick up robot parts that control motion, manage power and carry signals. Instead of shipping those components as freight, he packed them in a suitcase and took that with him on a flight back to San Francisco.
Amazon Web Services leaders met with engineers in May and delivered a sobering message: To ensure AWS has enough capacity for all its customers at its popular EC2 cloud server business in the future, engineers should save capacity however they can, according to a person with knowledge of the meeting.
That includes capacity on servers that run on central processing units, as well as those running on AI chips, which have long been in short supply, the person said. CPUs are the chips that have powered the last several decades of the internet age. Since then, some engineers within AWS have found it takes much longer to get capacity on CPU servers for their work.
Meta Platforms was ordered by a New Mexico judge on Thursday to pay $567 million in compensation after the company was found liable for harming teens’ mental health in the state.
The penalty follows a March jury verdict ordering the parent company of Facebook, Instagram and WhatsApp to pay $375 ...
South Korean memory maker SK Hynix said on Friday that it will start the construction of two new chip plants next year, after its board approved its plan to invest about 54 trillion won ($38 billion) in the new fabs.
The company’s announcement of the timelines for its new fabs comes as demand ...
China’s top internet regulator said on Thursday it had launched a cybersecurity review of products sold in the country by U.S. security software firm Palo Alto Networks.
The Cyberspace Administration of China said that the purpose of the review is to “ensure the secure and stable operation of ...
Datadog shares dropped more than 19% after the software firm forecast a slowdown in annual revenue growth for its current quarter and said it stemmed from its largest customer reducing its spending.
On an earnings call, CEO Olivier Pomel declined to name the customer, saying only that it ...
Over the last year, we’ve seen more generalist venture capital firms make forays into defense tech, a niche once relegated to specialized investors comfortable with revenue based on uneven government contracts and who weren’t restricted from investing in weapons makers by their limited partners.
Furientis, a startup founded last year that builds missiles for intercepting attacks, is close to raising $25 million in a funding round that would value it more than $125 million including the money invested, two people familiar with the investment tell us. And we’re hearing Benchmark, known for backing Uber and Snap in their early days, has been in talks to lead, the people said.
If you’re one of those executives at Alphabet, Meta Platforms or Amazon who deals with investors, here’s an idea: Put a large photo of SoftBank CEO Masayoshi Son on the wall of your office. Next time a visiting investor complains about your AI spending, point to the photo. That should end the discussion.
As much as big tech firms are spending on AI, at least they’re financing much, if not most, of it through internally generated cash. SoftBank in contrast has so far spent $54.6 billion on OpenAI shares by raising money through a combination of various kinds of loans and asset sales. And it’s now gone even further: On Thursday, SoftBank revealed that this week it had borrowed $10 billion using its OpenAI stake as collateral. As SoftBank has committed to put a further $10 billion into OpenAI in October, there’s a good chance it will make that payment using money borrowed against other OpenAI shares.
Lyft said late Thursday its gross bookings rose 23% in the second quarter to $5.5 billion, beating the 18% to 21% growth the company forecast in May, as its European taxi business and a record share of rides booked through partnerships lifted the average value of each trip. But it forecast a ...
Salesforce on Wednesday told state government agencies in Washington and California that it plans to permanently layoff 133 employees at the start of October
Some 74 employees will be cut in San Francisco, California, the company’s headquarters, while 59 will be laid off in Bellevue and Seattle, ...
Airbnb said late Thursday that revenue for the three months ended in June rose 17% from the year-ago quarter to $3.6 billion, topping its guidance of 14% to 16%, as the short-term rental company reported higher demand across regions. Net income climbed 27% to $816 million. Shares jumped 9% late ...
Software companies have been at pains to flaunt the success of their AI products, as they seek to dispel investor fears that AI apps are threatening their existence. Two reports this week show that developing popular AI products is only half the challenge.
Canva, the design software company popular among students and business owners, warned investors this week that its annual revenue growth would slow to 20%. The reason: it was tapping the brakes on a planned AI rollout expected to drive paying subscriptions—after demand had proved higher and more expensive to serve than anticipated.
Then on Thursday, Figma said it expects a slowdown in revenue growth for the September quarter, to 36% from 48% in June. In its case, the publicly traded design software company suggested it’s still figuring out how to generate sales from new AI tools that are still in beta, or testing phases, or early in their release.
OpenAI moved to dismiss Apple’s trade-secret suit, calling the case “rotten to its core.”
In a court filing Wednesday, the AI startup said that Apple in its lawsuit never identifies the specific trade secrets that it accuses former employees of bringing with them to OpenAI. Instead, OpenAI said ...
Atlassian shares soared as much as 38% in after-hours trading on Thursday after the enterprise software company reported its first quarterly operating profit in more than two years. Atlassian CEO and cofounder Mike Cannon-Brookes said he planned to start buying as much as $250 million in shares, ...
Login
