Justin Yek, a former VP of finance at Airwallex, will become Binance’s new CFO, succeeding Kaiser Ng, who has held the position since 2022 and is stepping down to pursue a new venture, the company said Saturday. Binance co-CEO Yi He said in a statement that a transition plan had been “implemented over the last half year.”

On Oct. 29, hackers drained an estimated $6.3 billion from the platform, Fortune’s Jeff John Roberts reported. Binance later said North Korean hackers orchestrated the attack, using social engineering and malicious software to compromise engineers and gain access to systems that enabled unauthorized withdrawals.

The damage was partly contained. Tether and Circle froze about $2.5 billion in stolen stablecoins. Binance says it has replenished 98% of the stolen coins through reserve funds and other sources, and that users won’t bear the losses. Read more here.

But, the incident brings questions that go beyond cybersecurity. How strong are Binance’s internal controls? And how can the exchange convince customers and regulators that their assets are safe?

Those are questions that will inevitably land on the new CFO’s desk.

Yek brings a résumé suited to a growing global fintech. At Airwallex, he progressed from head of strategic finance and corporate development to VP of finance, leading the company’s global finance team. Before that, he spent about nine years in investment banking at Citi and Morgan Stanley, working on more than $30 billion in equity, M&A and debt transactions, according to his LinkedIn profile.

That’s substantial experience in corporate finance and dealmaking. But the CFO title is new to Yek, and he lacks the crypto-exchange experience of Ng, who joined Binance from Kraken. 

Cybersecurity is increasingly a finance issue, not just an IT concern. The financial consequences of a breach, including lost assets, customer remediation, regulatory scrutiny and reputational damage, can ultimately fall on the CFO. For example, Deloitte’s Q3 2026 CFO Signals survey found cybersecurity was the top external concern for finance chiefs, cited by 50%, ahead of the economy and inflation. 

At Binance, Yek’s first test as CFO will be helping restore the trust that money alone can’t replace.

Sheryl Estrada
Sheryl.Estrada@fortune.com

This story was originally featured on Fortune.com

• Nvidia’s history of questionable statements about Megaspeed executive’s affair with Jensen Huang.
• AI companies have $3.2 trillion in off-balance-sheet commitments.
• Markets: Resilience in the face of controversy.
• There’s an 80% chance the Fed will raise rates in December.
• The AI investment boom is about to become bigger than the dotcom bubble.
• Canva cut 1,200 jobs, and the internet loved it.

This story was originally featured on Fortune.com

Huang’s net worth, tied to the computer processor and AI company he co-founded, stood at just over $171 billion early last week. It dropped to $150.9 billion by Thursday, and briefly rebounded to $154.9 billion by Friday—the latest data point on Bloomberg’s Billionaires Index.

The trajectory follows the path of Nvidia’s share price, now down 18.6% in the past month.

But even as signs of a small reversal bloomed Friday, the weekend brought another blow. “I was unfaithful to my wife. My actions hurt her and our children. There is no excuse for what I did. I am deeply sorry,” Huang said in a statement shared overnight.

“To my colleagues at NVIDIA … you deserve better from me. Leading NVIDIA is the privilege of my life. My responsibility is to serve you and be worthy of your trust. I have fallen short of that responsibility,” he added, saying he would work to “earn back” the confidence of the board, partners and company shareholders.

The woman he had an affair with has been identified in a series of reports as a suspected Chinese spy. Nvidia’s board of directors confirmed in an accompanying statement overnight that it had launched an independent investigation “in light of recent reporting that founder and CEO Jensen Huang has been the target of a foreign intelligence effort.”

The Chinese government has denied involvement in the alleged spy plot involving Huang, calling the debacle “nothing but smears.”

The board added that while it has no reason to believe sensitive national security or NVIDIA “secrets” were compromised, it will “examine all available information.”

Analysts are digesting the news this morning ahead of markets opening in New York. Equities in Asia rallied this morning, noted Deutsche Bank’s Jim Reid, “although concerns surrounding an internal investigation involving Nvidia CEO Jensen Huang are capping gains.”

“The company is down around -3.5% on platforms that provide weekend prices,” he added.

Trump backing

The importance of Nvidia among the companies at the forefront of AI development cannot be overstated. It is the premier supplier of chips to AI companies and one of the largest investors and guarantors of companies in the AI supply chain—to the tune of hundreds of billions.

Yet President Trump was unruffled when asked about one of the country’s most famous CEOs being potentially implicated in a national security threat.

“I have not spoken to [Huang] about it,” Trump told reporters last night. “I just know he’s a fine person, but you never know what is going on.”

Betting on the prediction market Kalshi had the odds of Huang leaving the company before July 27 at 25%. Professor Jeffrey Sonnenfeld, senior associate dean for leadership studies at the Yale School of Management, put the odds of his resignation at 50-50. Huang owns the largest single stake in Nvidia, and his children are among his employees. If he were to leave, there is no obvious successor.

Nvidia’s board said it has “confidence in Jensen and his leadership team.”

President Trump was also confident Huang would continue to lead the company, saying: “He’s done an amazing job for our country.”

This story was originally featured on Fortune.com

So what does this mean for what the industry has been calling an unstoppable technology? My personal read is an AI slowdown, at least for now.

Increased regulation

When the new Congress is seated in January, expect a wave of regulation. Some of it will be needed, some of it will be onerous on the industry for no good reason. Mandatory testing of frontier models against government-set standards is worth defining and achieving. We should also expect stronger whistleblower protections for the people inside these labs who see something dangerous and bring it to the public’s attention.

I predict there will also be an attempt to write today’s Executive Branch review of frontier models into statute. We need transparent incident reporting and independent verification organizations. We may also see federal liability, possibly strict liability, with product liability framing and CFAA additions aimed at AI agents. There will be attempts at export controls widening from chips to the equipment that makes them, plus funded chip geolocation.

The change most apparent to the ordinary citizen will be action on surveillance pricing. Surveillance pricing is the use of AI and individual consumer data like browsing history, purchase habits, or location to set a personalized price for each shopper. The FTC moved in August with a proposed enforcement policy stating that undisclosed personalized pricing likely constitutes an unfair or deceptive practice, and states have begun moving on their own. Expect Congress to try to write that into law. Reining this in would bring real relief to the bank accounts of Americans.

We also need a serious debate on pacing. At least two bills already filed, Ro Khanna’s Human Control Over AI Act and the Sanders and Casar superintelligence bill, would prohibit recursive self-improvement outright. Congress will pass a good chunk of this. For the next two years, Trump and his veto pen remain the wild card.

Are we handing the AI race to China?

This view is too alarmist. The United States still leads in chips, capital markets, talent, and open research. A slower, costlier buildout narrows the lead. It does not surrender it. The 10 to 15 gigawatts that analysts say “may never come online” by 2027 is a pacing problem, and a real one. But the cost of the fix is minuscule next to the cost of leaving it unaddressed.

There are terms that even data center opponents will accept, and they are easy wins: pay your own grid upgrade costs, sign longer contracts, publish water and energy use, and flex load at peak hours. If these centers want approval from a public that will not be fooled, they have to pay their own way. Duke University’s Nicholas Institute estimates the existing grid can absorb roughly 76 gigawatts of new flexible load at a 0.25 percent annual curtailment rate. The capacity exists. It will go to the companies willing to accept these terms.

Texas as a data center test lab

States and local governments weighing data centers need to study Texas. Greg Abbott courted the industry harder than any governor in America, then reversed course in August with a pause on new grid connections once the polling turned. It was not enough. Gina Hinojosa, who campaigned on a data center moratorium and a seat at the table for local communities, unseated a three-term incumbent who had never won the office by less than double digits. The pattern held elsewhere. Sherrod Brown won back a Senate seat in Ohio after casting his opponent as a cheerleader for data centers, and Keisha Lance Bottoms won the Georgia governorship while backing a pause on new construction. The lesson for every other state is that once the public turns, a late conversion does not save you. The open question is whether the industry will accept terms that put the costs it expected to share with the public squarely on its own shoulders.

Who wins the slowdown

Another way to lessen the load is shifting AI to distributed hosting, meaning AI processing moves out of a few massive centralized data centers and is split across a coordinated network of smaller regional facilities, edge devices, and on-premises hardware. Training frontier models still requires giant clusters, but inference can live closer to the user at lower peak load.

“My prediction: a more distributed AI map, with many smaller data centers built where spare power already sits, wired together by fiber like a nervous system,” says Andrej Danis, head of digital infrastructure at AlixPartners. “Fewer superclusters may actually give us a more resilient network, and a better one for physical AI. This summer, our survey had sites of 50 megawatts or less as the players most likely to be left behind. A permitting squeeze on the biggest campuses may flip that.”

“Every fully permitted data center under construction just got more valuable,” Danis says. “But many advanced projects do not have their ducks in a row on power, water, and operating permits. Those deserve a hard second look, because after Tuesday a missing permit can kill or delay a project. Tight real estate will also speed up the chip refresh. If you cannot build new halls, you swap in newer chips faster, and the older ones move to other countries where export rules allow. Part of the US slowdown gets exported.”

The part worth arguing about

The industry spent two years telling the public that AI is an existential risk to humanity, and people believed it. That fear campaign now puts the entire technology at risk in the United States. Sam Altman signed a 2023 statement declaring that mitigating the risk of extinction from AI should be a global priority alongside pandemics and nuclear war. OpenAI’s own researchers wrote that superintelligence “could lead to the disempowerment of humanity or even human extinction.” Dario Amodei told the Axios AI+ DC Summit: “I think there’s a 25% chance that things go really, really badly and a 75% chance that things go really, really well, with not much space in between.” If the pilot told you there was a 25% chance your plane would crash, would you get on it?

These AI companies made this mess. It is their job to clean it up. In my estimation, that looks like more payouts in the form of community benefit agreements, revenue shares, tax commitments, and grid upgrades paid in full by the company doing the building. Money alone will not solve this. Communities need the data and resources to make their own decisions, and they will no longer allow their governments to take a check from big tech and shove a data center down their throats. That was made very clear at the ballot box, and it was already visible in last year’s Virginia, New Jersey, and Georgia results.

What compute, AI, and data storage need now is a social contract. The terms are simple. The company pays the full cost of its grid connection, publishes its water and power use, flexes load when the grid is stressed, and shares revenue with the host community. In return, the community gives it a permit, a timeline, and a fair hearing. Companies that treat energy, siting, water, and community economics as strategy will build, because they will work with a community’s citizens, not only its leadership. Those that don’t will have to find a patch of dirt in the middle of absolutely nowhere, or not build at all.

There will be an AI and data-center slowdown. There is no doubt about that. But I am betting on the resilience of the greatest minds in this country to solve the problem and find a way to keep innovating and leading the world. It’s what we’ve always done. It’s what we must do now.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

This story was originally featured on Fortune.com

Good morning. Zendesk CEO Tom Eggemeier says he’s been telling his kids since 2023 that AI will be “bigger than the printing press, the internet, the Industrial Revolution,” and that they’d fall behind without it. Earlier this year, his son Will, a Boston College senior and basketball player who’s majoring in history with a minor in business, called his bluff. Will suggested they work together on a predictive analytics app to help college athletics programs find players in the NCAA transfer portal before richer schools do. “We wanted to create something that might be able to level that playing field a little bit,” Will told me during a recent interview alongside his dad. Tom helped design the interface, and they named it Underdog AI.

At first, everything was easy. “We were vibe coding, and it was exciting because of the cool features that were coming out,” Tom said. But Will knew something was off when the AI made recruitment suggestions that didn’t sound right. Speedy development created technical debt in areas from security to API protocols, and hours of debugging that reinforced that there are no shortcuts to making software seamless and secure. Two college basketball general managers told them the app matched about 70% of what’s already for sale; Will also can’t sell to rival schools while he’s on BC’s roster, so he’s tweaking the project until he graduates.

Will says he loved seeing his father work through problems in real time. He’s hoping to build a career in the sports business and believes knowing how to deploy AI is essential to success. He credits his dad with helping him develop grit, citing “suck it up” as a favorite line.

Tom gained as much or more from the experience, saying it made him “a better leader for our R&D organization, our product, and AI.” The experience changed how Tom thinks about AI, from investing heavily in employees’ AI skills to focusing on how AI agents help people cut expenses, time, complexity, or solve other real-world problems. 

Some lessons he’d share with other CEOs?

Use the technology yourself. Tom used a personal agent to renegotiate his insurance coverage, internet bill, mobile plan, and subscriptions, and says he’s “a little embarrassed to say it saved me $3,150 a year.” Having trained ChatGPT Work to write in his voice, he noticed that it kept ignoring his instructions to avoid contractions, which he thinks are harder for non-native English speakers to understand. “That recurring mistake has led me to think more deeply about how memory and skills affect our products and internal use of AI.”

Also, keep hiring young employees; AI natives bring a fluency that creates an interesting dynamic alongside experienced workers. Understand the anxiety around AI while helping people gain skills. Create projects that put AI to work on real problems. And if your kids want to build something with you, do it.

Contact CEO Daily via Diane Brady at diane.brady@fortune.com

This story was originally featured on Fortune.com

Its door was wide open, and anyone on staff could simply walk in and take what they wanted off the shelves. When products weren’t there, they were written on a sticky note—or simply remembered by staff—and orders worth tens of thousands of dollars were made on that basis.

I sat down with the owner and asked her: How much have you sold over the last three months? She said she had no idea: current stock wasn’t written down, and if she wanted to do an audit, staff had to go through paper receipts. Later, in the corner of a supply room, I found a box of expired filler worth $50,000 retail. It had never even been opened.

I come from four generations of doctors: My great-grandfather ran a hospital, my grandfather ran a radiology department, and both my parents are physicians—I am the first in the line not to go to medical school. Instead, I did a master’s in public health at Columbia, and after graduating joined a family office in New York.

In 2023, we started looking at rolling up med spas, buying small practices and building them into something bigger. We’d looked at 50 to 60 health care verticals and decided that aesthetics and wellness was the best bet. It had grown massively, from a couple of thousand practices countrywide to close to 20,000 in less than a decade.

But then I started visiting them.

It is standard practice when investing to ask to look at the books, but in many cases there were none. The more practices I visited, the more I realized that this wasn’t a couple of bad apples—this was the industry at large. It is, for the most part, run by clinicians or first-time entrepreneurs with little business background: they weren’t willfully negligent, they had simply never been given the tools.

In a med spa, inventory is the single largest expense, even bigger than payroll.

A vial of Botox, for example, costs several hundred dollars—and it is small enough to fit into a jacket pocket. It would have been easy for me to walk out of any one of those med spas with three vials—worth $2,000—with very little effort. Given the razor-thin margins of many practices in the U.S., that can be the difference between a profit-making month and a loss.

And then there was the safety aspect. God forbid that a state pharmacy board inspector had dropped in at any point because many of these spas could not have produced documentation.

I still visit clinics where there is no Hylenex (a drug used to dissolve hyaluronic acid when a filler injection goes wrong) on the shelf, a serious issue. If filler is accidentally injected into a blood vessel, a patient can suffer permanent tissue damage: Hylenex is the standard emergency treatment, and needs to be on the shelf. I saw clinics that were using unauthorized numbing creams: one bad reaction and a practice would be finished.

So that day in New York, I went back to the family office and told them I didn’t want to help them buy med spas—I wanted to fix them. Their answer was: good luck. If it works, we’ll invest.

I went back on the road, knocking on doors across New York and persuading med spas to let me take over their purchasing: I visited 40 or 50 outlets and came away with seven yeses. Then we signed our first enterprise client and word started to spread. Today we work with the largest med spa groups in the U.S., serving more than 1,500 locations in all.

All that has taught me a lot, but the best lesson I ever received about building a product was from my mother, a physician. When I first described the company to her, everything ordered automatically, no phone calls, no fuss, she said: “I would never use you.”

She said that her sales reps weren’t a nuisance that could be automated away, they were how she learned what was new, what worked and what to watch for.

So we pivoted: instead of trying to replace human reps, we started streamlining the paperwork around them. The vendors noticed their best-run customers were growing and buying more, and started sending new practices our way.

Three years in, I am convinced that the messy supply closet is a microcosm of a much bigger story: wellness businesses are two decades behind the technological curve and it is the simplest problems that are holding them back. Insurance reform and new drugs will play a seismic role in American health care reform, for sure. But equally important are the simplest issues—procurement, purchasing, admin—that help businesses keep the lights on.

Transforming health care doesn’t always start in the operating room or the lab. Sometimes it starts in the supply closet.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

This story was originally featured on Fortune.com

“You read these stats about how 48 million Americans or 50 million Americans live in food-insecure households. That’s the reported number!” said Rao, CEO and co-founder of Sharebite, a New York–based employee meal-benefits platform. He argues the real figure is roughly three times higher.

“So you’re talking about close to 150 million people. Nearly half the country’s population is struggling; they’re living from paycheck-to-paycheck, and they don’t know where their next meals are coming from,” he said.

That number is Rao’s estimate. The official baseline is 47.9 million people who experienced food insecurity in 2024, according to Feeding America’s analysis of federal data. It may be the last reliable count for some time, because the USDA has stopped collecting the household survey behind it.

The pressure behind those numbers is showing up at work. Annual inflation reached 3.7% in September, according to the Bureau of Labor Statistics, and food prices are up 33% since 2019, per an Associated Press analysis of federal data cited by NPR. A Government Accountability Office report found that Uber, Lyft and DoorDash now lead major employers in workers receiving SNAP benefits, ahead of Walmart and McDonald’s.

Rao’s pitch is that companies can close part of that gap through something many already pay for: lunch. “We just have to start thinking about hunger, not only as a societal problem but more as an economic problem that businesses can help solve,” he said.

How Sharebite turns office lunch into donations

Founded in 2015, Sharebite is built for businesses, not consumers. Companies use it to manage employee food ordering, meal allowances, and catering for in-office, remote, and hybrid teams, setting stipends and spending limits from one dashboard instead of processing expense reports.

For every meal ordered on the platform, Sharebite donates an equivalent meal through nonprofit partners including Feeding America and City Harvest. The company says it has donated more than 18 million meals to date.

Its investors include Contour Venture Partners, Essential Capital, Lafayette Square, Liberty City Ventures, and Fiserv, the payments company on the 2026 Fortune 500. Sharebite raised a $39 million Series B round in 2022.

Rao, a former Credit Suisse investment banker, says the idea came from watching how other countries tackle food insecurity by “mobilizing private sectors to step in and help solve it.” He felt it was “high time for us to be able to implement something like that.”

He describes the company as his form of civic service, arguing that serving society, building a successful business, and drawing on the private sector are “all interwoven together.”

Why Rao points to Brazil

Rao’s favorite model is Brazil’s Workers’ Food Program, created in 1976. Companies that join receive tax incentives for feeding their employees, and the program now reaches about 22 million workers, according to Brazil’s Ministry of Labor and Employment.

“Meal benefits or meal vouchers are a standard practice in Brazil as a result,” Rao said.

He argues the U.S. should add food to the employer benefits framework that already covers health care and retirement, a change he says would support workers and stimulate small-business activity. Rao, who has written commentary for Fortune, made a similar case in a 2023 essay on hunger among American workers and again in a 2025 piece written as the record government shutdown ended.

The productivity case for feeding workers

The business argument rests partly on cost. Healthcare spending and lost productivity tied to poor diets and food insecurity exceed $1.1 trillion a year, according to a fact sheet from the Food is Medicine Institute at Tufts University.

Lunch has also become a lever in the return-to-office. Workers surveyed by Owl Labs said they would come back if employers covered commuting, childcare, and lunch and coffee, Fortune reported in 2023.

Rao says Sharebite’s job is aligning “a real business need with a much broader problem,” and credits his team with developing “solutions to solve those problems at that scale” as it works with larger customers.

Record months on a small team

Rao says demand is accelerating. “Last week was our best week ever in the company’s history. October was a record month for us. And it’s the same people who’ve been here for the last few years. Our headcount is small, but our business continues to grow,” he said.

He added that the business is more than 30 times larger than it was four-plus years ago.

After roughly a decade “building inside this category,” Rao doesn’t think the market has matured. In his view, it is “still at the early beginnings of this entire category.”

This story was originally featured on Fortune.com

The phenomenon, which SHL calls “AI rebounding,” shows an early indication that some companies may be discovering the limitations of replacing human employees with AI.

“Organizations are making fast, sometimes myopic, decisions about their people based on an incomplete picture of what their employees can do with or without AI,” Lucy Beaumont, global senior vice president of product and design at SHL, said in a press release. “AI rebounding is what happens when a company automates a role before it has genuinely assessed the human capability inside it. The AI doesn’t deliver, the gap becomes obvious, and the business ends up rehiring the same experts they just let go of, at a far greater cost.”

The survey, conducted in June and July, found that 19.1% of American respondents said they had personally been replaced by AI and were subsequently asked to return to their jobs. Across both the U.S. and the U.K., the figure was 16%.

And among the returnees, 51% wanted a substantial pay increase before agreeing to return, SHL found. Additionally, 42% said they would require written guarantees about future job security, and 40% wanted transparency about how AI would be used in their positions.

Companies are finding out AI can’t do everything

Similarly, a Robert Half survey from July found that 32% of U.S. hiring managers whose organizations eliminated positions because of AI later had to reinstate the same or similar roles. Among that group, 40% said the positions required institutional knowledge or context that AI couldn’t replace. Another 39% cited relationship management, 39% pointed to increased business demand, and 38% said the positions required more human oversight and quality control than expected.

And some prominent companies have already encountered those limitations. Swedish buy-now, pay-later company Klarna began recruiting human customer service representatives after relying heavily on AI chatbots. CEO Sebastian Siemiatkowski acknowledged the company had gone too far in its reliance on automation, according to Bloomberg.

A Klarna spokesperson told Fortune human agents were needed “for complex, sensitive cases.” The company began directly hiring a “small number” of high-skilled humans into the customer service process to identify where the “human touch” can bring most value to customers.

Ford also hired and promoted 350 experienced engineers in June after AI-powered quality control tools failed to perform as expected. The automaker said AI had contributed to quality improvements but emphasized the importance of combining the tech with experienced engineering teams.

“These engineers carry the hard-earned wisdom of decades of design,” a Ford representative previously told Fortune. The company said AI is very important to quality gains, “and that, in tandem with deep technical expertise, is what’s needed.”

AI layoffs have been looming, but jobs are growing

Despite the rise of AI rebounding, the number of companies that cited AI as a reason for workforce reductions has grown substantially.

According to outplacement firm Challenger, Gray & Christmas, employers cited AI in 121,401 announced job cuts through October, more than double the 54,836 announced layoff plans for all of 2025.

And younger workers are especially vulnerable to AI-led job displacement. A Stanford Digital Economy Lab study found that employment among workers aged 22 to 25 in occupations highly exposed to AI was 19% below where it would have been if it had grown at the same pace as employment among their less-exposed peers.

SHL’s research shows some promise for the younger workers, however. The firm found that younger workers were more likely to report experiencing the “AI rebounding.” Among respondents aged 18 to 24, 24% said they had personally been replaced by AI and later asked to return, compared to 19.1% of U.S. respondents overall.

This story was originally featured on Fortune.com

Nvidia’s board released a statement Sunday that a committee of independent directors and an outside counsel will oversee the probe, adding that Huang supports the review and plans to cooperate.

“The Board prioritizes U.S. national and economic security, which is paramount to the interests of NVIDIA and its shareholders,” it said. “While the Board has no reason to believe that either national security or NVIDIA secrets were compromised, the Board’s inquiry will examine all available information.” 

The announcement followed reports that Huang had a romantic relationship with Megaspeed executive Alice Huang, whom U.S. intelligence suspects was working as a spy for China.

According to the New York Times and Financial Times, the Nvidia boss admitted to the affair and apologized in an internal message to company employees.

“I was unfaithful to my wife. My actions hurt her and our children. There is no excuse for what I did. I am deeply sorry,” he said in the message.

Representatives for Nvidia and Megaspeed didn’t immediately respond to requests for comment.

U.S. officials believe Alice Huang was recruited by China’s Ministry of State Security, sources told the FT, and surveillance revealed that she was feeding information to China, according to the New York Times.

Jensen Huang, who isn’t related to Alice, is a central figure in the U.S. AI industry, leading a company that develops world-class chips powering large language models.

Its technology is so sensitive that the Trump administration has restricted exports of Nvidia’s most cutting-edge products to China while President Donald Trump has often cited the U.S.-China AI rivalry as a reason for not slowing down development.

Meanwhile, Megaspeed is a Singapore-based data center company that quickly became one of Nvidia’s top customers in Asia, purchasing more than $4.8 billion in chips over the past two years.

But the U.S. also investigated Megaspeed last year for allegedly helping smuggle restricted AI chips into China, which the company has denied.

News of the Nvidia’s probe into Huang comes weeks after reports indicated the U.S. defense and intelligence communities are getting more aggressive about industrial espionage.

Sources told the Washington Post that the Pentagon is planning a spying program that deploys “nonofficial cover officers” who pose as executives—a covert effort in a larger economic warfare campaign.

Politico also reported that the Defense Department is quietly building an elite agency to covertly sabotage foreign technology programs that could threaten U.S. military superiority.

The Technology Support Office was established in April by reorganizing two classified intelligence programs that have been secretly carrying out technology sabotage at a smaller scale since 2023.

The Pentagon has denied the reports.

 “This query appears to be built on information that is simply not true. Whoever is feeding you these claims are either badly misinformed or deliberately leading you astray to peddle their fantasies,” it told both outlets.

But other signs point to the U.S. taking economic warfare more seriously. In September, CIA Deputy Director Michael Ellis said the U.S. is expanding espionage against China by targeting companies in sensitive areas like AI, biotechnology and chips. 

Also this year, the Pentagon formed the Economic Defense Unit for “deterring, denying, disrupting, and helping to defeat adversaries through economic means,” sources told DefenseScoop.

AI in particular has emerged as such an important technology that the Defense Department has been more heavy-handed with U.S. companies as well.

Anthropic’s AI models were used in the military, but a dispute between the San Francisco-based startup and top brass led to a schism just as the Iran war started.

CEO Dario Amodei has said the company is patriotic and seeks to defend the U.S., but wouldn’t allow its AI to be used in mass domestic surveillance or autonomous weapons.

The Pentagon insisted it would use the AI in lawful scenarios and refused to abide by any limits from the company that would go beyond those constraints.

After failing to reach a compromise earlier this year, Trump ordered the Pentagon to phase out Anthropic’s technology. Defense Secretary Pete Hegseth also designated the company a supply-chain risk, meaning contractors can’t use it for military work.

Last month, an appeals court ruled the supply-chain risk designation is illegal, but the Pentagon confirmed it’s no longer using any Anthropic tools.

This story was originally featured on Fortune.com

Within days, $3.2 trillion of AI-related stock market wealth was wiped out as investors bet the election winners would slow the AI buildout. Texas holds one-fifth of the planned data-center pipeline, and its governor-elect has promised to ban new approvals until the state legislature acts.

I had been privately skeptical, because data centers are the bedrock of the massive industrial boom powering most of America’s GDP growth. From an economic perspective, being anti-data center is like being pro-recession, I assumed, but Tocqueville’s Democracy in America explains a lot of what has really gone on in America for centuries.

Tocqueville was a young French aristocrat who toured the United States in 1831 and published Democracy in America in two volumes, dated 1835 and 1840. The Library of America calls it arguably “the most perceptive and influential book ever written about American politics and society.” One big idea he offered was what he called the “general equality of condition” among Americans, meaning how widely property and status were spread around. And this, he proposed, is the fact that the rest of American life comes out of. He argued that Americans hold tight to their property, care more about equality than liberty, and organize locally to get what they want. The data-center industry made its plans without accounting for any of that, and it’s an open question whether the Democrats who just won a landslide really understand it, themselves.

Most Gen Xers can barely remember a wave like this one. The closest historical comp is the 1974 blue wave that swept aside the corruption of the Nixon Watergate era. The 2026 midterms suggest a scrambling of the calculus of red states and blue states much like Donald Trump’s shocking election did a decade ago, when the “blue wall” of Pennsylvania, Wisconsin and Michigan crumbled.

Nationally, Democrats have so far flipped seven Senate seats and are on track for at least 258 House seats. They took a majority of governors’ mansions for the first time since 2010, and Trump acknowledged that Democrats had secured both chambers. Down the ballot, voters in 10 states decided 41 local measures on data centers, and in 40 of the 41 they favored limits on development. In NBC’s exit poll, 76% said they would oppose a new data center in their community. There were many stunners, like Kansas electing its first Democratic senator since 1932, but Texas stands out as the signature shock.

The Lone Star state went blue across its statewide offices. Beyond the Senate and governor’s races, Democrats won agriculture commissioner, attorney general, land commissioner, comptroller, lieutenant governor and railroad commissioner. They have taken the state House for the first time since 2002. Gina Hinojosa unseated Gov. Greg Abbott and will be the first Democrat elected governor of Texas since Ann Richards in 1990, and the first ever Hispanic person to hold the office. The Texas House looks likely to have its first female and first Black speaker in the form of 87-year-old Senfronia Thompson, who took office in 1973.

Reading Tocqueville in this climate is downright spooky. He wrote in 1835 that American political bickering at first looks to a stranger “so incomprehensible and so puerile” that he can’t decide whether to pity the people or envy the happiness that lets them argue. I personally thought as much for months about the issue of data centers — a massively overstated environmental risk but a massively understated political and emotional issue — until voters proved they were not messing around.

‘Popular authority’

Look closer, Tocqueville wrote, and nearly every faction of American political life comes down to one side trying “to limit,” and the other “to extend, the popular authority.” Five years later, in his second volume, he explained what made Americans different from the Europeans watching them. “The great advantage of the Americans is that they have arrived at a state of democracy without having to endure a democratic revolution; and that they are born equal, instead of becoming so.” Two centuries on, that birthright came out on Nov. 3 as a mass of voters told the Republican Party, and the data-center builders, something like “No, I don’t think so.”

Property, unlike aristocratic Europe where it is passed down from along family lines for centuries, is what keeps the whole thing from flying apart for most Americans, who are “possessed of sufficient property to desire the maintenance of order, yet not enough to excite envy.” Tocqueville observed that “the passions engendered by property are keenest and most tenacious amongst the middle classes.” And because equality leaves each person weak on his own, Americans band together to form what he called “associations,” without which they “become powerless if they do not learn voluntarily to help one another.”

Tocqueville also prophesied that Texas would hold the key to the American future. “The inhabitants of the United States are perpetually migrating to Texas, where they purchase land; and although they conform to the laws of the country, they are gradually founding the empire of their own language and their own manners,” he wrote, a decade before Texas joined the Union.

Texas has spent the 2020s proving him right. It added 391,243 residents in the year to July 2025, more than any other state for the third year in a row, reaching 31.7 million. It produced $2.9 trillion in GDP in 2025, and holds one-fifth of the planned 253-gigawatt data center pipeline nationwide. The five fastest-growing American cities, the Census Bureau revealed in May, were all in Texas, some for the second year running.

This year, Texas put every one of Tocqueville’s arguments on the ballot at once. Landowners fought transmission lines after a 2023 law cut the time regulators have to consider them from 365 days to 180. Groups like Save Rural Texas organized against data centers, and while Abbott ordered an audit of every data center in the grid queue, Hinojosa called for a statewide moratorium, vowing days before the vote to “stop the data center takeover.” Then the voters made clear in resounding fashion how they felt about property, equality and popular authority.

Equality, property and associations in the 21st century

For a lot of voters this year, equality came down to the cost of living. In Gallup’s final preelection survey, 44% of voters called income and wealth distribution extremely important, the highest reading in the three midterms Gallup has measured it, up from 32% in 2018 and 31% in 2014. On PJM, the grid running from Illinois to New Jersey, the independent market monitor puts data centers at $6.3 billion of the latest auction’s $16.4 billion in capacity charges, and at $29.4 billion, or 46%, across the last four auctions. In the October ABC News/Washington Post/Ipsos poll, 57% of Americans called home energy and utility costs unaffordable, up from 45% in February.

On property, Tocqueville wrote that “the heart of man is not so much caught by the undisturbed possession of anything valuable as by the desire, as yet imperfectly satisfied, of possessing it, and by the incessant dread of losing it.” The data-center fears, environmental concerns aside, cut straight to the American dread of losing one’s property. In August, a University of Texas/Texas Politics Project poll found 57% of Texas voters opposed a data center in their community, including 43% of Republicans and 60% of rural voters. The other side claims property, too. Texas Policy Research, a free-market think tank, argues Texas should protect private property rights and free enterprise while making large electricity users pay the costs they impose on the grid.

The “associations” that Tocqueville observed were so obvious in this election season they hardly need mentioning. Data Center Watch, a tracker maintained by AI research firm 10a Labs, counted 833 opposition groups across 49 states by March 2026, up from 396 at the end of 2025. In June, Seattle’s city council voted unanimously for a one-year ban, making it the largest U.S. city to pass one. None of these ran through Washington.

250 years of American surprises

Tocqueville spelled out the lesson in 1840: Americans describe their civic life in terms of self-interest. “The Americans… are fond of explaining almost all the actions of their lives by the principle of interest rightly understood,” he wrote. “They show with complacency how an enlightened regard for themselves constantly prompts them to assist each other.”

When they do something public-spirited, in other words, they insist it is in their own interest. They “seldom allow that they yield to emotions of this kind; they are more anxious to do honor to their philosophy than to themselves.” He didn’t mean that as an insult, calling the principle “the best suited of all philosophical theories to the wants of the men of our time,” and “their chief remaining security against themselves.”

That’s what most readings of the data-center fight, including mine, missed. From outside it looks like people guarding their power bills and their land. Others suspected the opposition was manufactured, even from China. The experts Fortune spoke to in June found it organic, with little evidence to support claims of China’s interference. The people in the fight talk about bills and property, as Tocqueville said they would, and they organized and voted in 41 places.

Americans misread themselves on the bigger question, too. Before the vote, fewer than half expected free and fair midterms, according to a University of California, San Diego, survey. Experts weren’t much sunnier. In Bright Line Watch’s October survey, their median forecast of a free and fair election was 75%, and one article in Political Science Quarterly argued the country had descended into soft competitive authoritarianism. But Tocqueville reasoned in the chapter “Why Great Revolutions Will Become More Rare” that a country of small owners has too much to lose to blow itself up.

“Nations are less disposed to make revolutions in proportion as personal property is augmented and distributed amongst them,” he wrote. Americans’ revolutions come through the ballot box when equality feels threatened. The contrast with Germany and France, where disaffected Gen Zers’ protests have turned violent in the streets, looks straight out of the late 18th century.

Democrats would be making a Tocqueville-sized mistake to read Nov. 3 as a conversion. The border counties that swung toward them had swung the other way two years earlier. In 2024, Trump won all four counties in the Rio Grande Valley, a region Democrats had swept in almost every presidential election since at least the 1920s. In Texas, where Democrats won their largest House majority since the 1993 session, several Democratic flips were decided by mere hundreds of votes. In Kansas, Hamilton won as an “independent-minded Democrat.” He avoided association with national Democrats and said he wouldn’t support Sen. Chuck Schumer for another leadership position, even though his campaign took $5 million from a PAC associated with Schumer.

The voters who elected them were following the bills, the land and the town meeting and will surely follow them again.

This story was originally featured on Fortune.com