This is exactly the plan, it turns out — Argentina wants more people like him. The country has been preparing for the last year to launch a full citizenship-by-investment scheme. In July 2025, Decree 524/2025 established an Investment Citizenship Programs Agency within its Ministry of Economy that would allow, for the first time, foreign investors to apply for citizenship without needing to reside in Argentina first. The exact parameters are still being worked out, with the Financial Times reporting that wealthy foreigners may be able to obtain Argentine citizenship in exchange for a non-refundable donation of about $500,000 or buying $1 million in zero-coupon government bonds, citing people familiar with the government’s plans.

The sheer scale of the plan is what sets it apart from anything the citizenship-by-investment industry has tried before, according to “This is a country of over 40 million people, and the opportunities, the business opportunities that are available in Argentina are endless,” he told Fortune. The largest countries to previously offer citizenship for investment—Montenegro and Malta—are small nations by comparison, he noted.

“This is a country of over 40 million people, and the opportunities, the business opportunities that are available in Argentina are endless,” he told Fortune. The largest countries to previously offer citizenship for investment—Montenegro and Malta—are small nations by comparison, he noted.

Katz pointed to several selling points for wealthy investors. The country sits on Vaca Muerta, one of the world’s largest shale oil and gas formations, which has been geologically compared to the Eagle Ford shale in South Texas by experts, alongside major lithium, gold, silver, soy, corn, beef, and wheat industries. He cited the roughly $22 billion a year in trade that Argentina has with the European Union as further evidence of the scale of opportunity.

The push builds on reporting the government has been developing to court prominent wealthy figures to what a former official described as a “new land of freedom” for billionaires.

“I think it will be a serious contender and player in the wealth migration, investment migration space,” Dominic Volek, who advises ultra-high-net-worth families on residence and citizenship planning at Henley & Partners, told Fortune.

The appetite for such an option isn’t hypothetical. Wealthy families in the U.S. are actively searching for safe havens. A proprietary survey of 1,800 Americans commissioned by Katz’s firm found that 61% would consider moving out of the United States within the next five years—a number Katz called “incredibly shocking.”

For years, wealthy Americans looked to New Zealand, Portugal, Greece, and the Caribbean as backup plans. Now Argentina—long associated with inflation, capital controls, and default risk—is trying to sell itself as a Plan B for outsiders with money. 

Argentina’s passport already grants visa-free access to a long list of countries, Volek noted, but citizenship would come with an added bonus: settlement rights across the nine-country Mercosur bloc – which includes Brazil, Colombia and Ecuador — similar to what an EU passport confers across Europe. 

“There’s increased optionality available to you,” he said, explaining that Argentina’s remoteness from the U.S. while also being in a similar time zone makes it “incredibly attractive.” 

Katz also said the flight to Buenos Aires, while nearly as long as a trip to Europe, doesn’t come with the jet lag that a European trip does. “That’s a huge, huge thing for somebody, especially an American business person, whose life is travel,” he said.

Katz also pointed to a bigger-picture safety pitch: South America is currently the only continent besides Antarctica that isn’t at war, and Argentina itself hasn’t fought one in decades. 

Thiel’s arrival is a signal, but the open question is whether Argentina can turn billionaire curiosity into durable capital, or whether it is selling a safe haven in a country still defined by volatility. Volek’s firm expects Argentina’s citizenship-by-investment program to go live by the end of the year and is already holding a roster of clients ready to apply the moment it does. 

“For our business and for the investment migration industry as a whole, it will be quite a game changer,” he said.

Difference between safe havens and tax havens

“There’s really no such thing as a golden visa,” Katz told Fortune. “These are temporary statuses, and they can go away.” Only citizenship, he said, gives someone the assurance that they’ll be able to remain in a country indefinitely.

Despite the Argentine government’s framing of the program through a tax lens, advisors caution against reading Argentina’s push—or the broader boom in second citizenships—as primarily a tax play, at least for Americans. The U.S. taxes its citizens on worldwide income no matter where they live, so acquiring Argentine citizenship changes nothing for a client’s IRS bill unless they go through the far more drastic step of renouncing U.S. citizenship.

Instead, getting a second citizenship applies the logic of wealthy people’s portfolio diversifying instincts to passports. 

“Why on earth would you have one country of citizenship and only one country that you can live in when you have the financial capacity to build a portfolio of options?” Volek said. 

David Lesperance, a leading international tax and immigration advisor with over three decades of experience, tells American clients to think of their citizenship and residency options as a hedge against whatever their personal “wildfire” might be—a hurricane, an earthquake, political violence, antisemitism, mass shootings, or a punitive new tax.

“If you look at these alternative residences and citizenships as fire insurance, and people incorporate them into a fire escape plan, they may not actually leave unless the literal wildfire happens,” Lesperance told Fortune. “But I recognize that it could happen, and I have the means to protect my family from it.”

Crucially, Lesperance said, a move like Thiel’s doesn’t require moving money along with it. 

“You need to separate where you live from where your assets are,” he said, describing having considered relocating his own family to Buenos Aires before ultimately choosing Koh Samui, Thailand. 

“If you’re going to physically move yourself and your family to a place like Argentina, that does not require me to move my wealth to Argentina,” he said. A client might simply like Buenos Aires and consider it safe for their family, he said, while making an entirely separate decision about where to bank and pay taxes.

Lesperance has seen South America’s profile rise sharply among his American clients over the past year to 18 months, alongside longer-running interest in Europe. But for now, both advisors and their clients are in wait-and-see mode. 

“Everyone is sort of waiting for the program to actually be available before they’re making any sort of decisions,” Volek said. 

This story was originally featured on Fortune.com

Brent crude oil climbed as high as nearly $110 a barrel on Monday, up 4%—its highest price since May, before easing to around $107 on Tuesday. The increase raised concerns about inflation and borrowing costs, evoking memories of the oil shock stories from years ago. Back in 1980, Americans spent about 6% of their income on gas because they used more and prices were relatively high, according to JPMorgan’s analysis. Today, that share is about 2.5%.

That doesn’t mean economists are completely at ease. Their greater concern is not that crude crossed the $100 benchmark, but that shortages have pushed up the prices of gas and diesel—fuels that directly affect people and businesses. If those prices remain high, Americans might have to cut back on spending while businesses may have to pay more to ship goods, run factories, and operate farm equipment. 

The re-emergence of the U.S. as a net energy exporter means oil shocks “hit differently” today, according to Michael Pearce, chief U.S. economist at Oxford Economics. Pearce told Fortune that higher oil prices are bad news for households, but good news for energy producers. 

“There is not a ‘tipping point’ for crude oil prices that will tip the economy into recession,” Pearce said.

Inflation has also changed what the $100 number actually means. Patrick De Haan, head of Petroleum Analysis at the gas tracking app GasBuddy, told Fortune that $100 today does not carry the same weight it did decades ago. He said oil may need to reach closer to $200 to have a similar effect on the economy today.

The war has inevitably put pressure on refined fuels such as gasoline and diesel, Pearce said. But at the same time, a shortage of refinery capacity has caused their prices to rise more than one would expect based on oil prices alone. Simply,  gas takes money directly from consumers, while diesel powers the trucks, farms, and factories that keep goods moving across the country.

The national average for regular gasoline was trending toward $4.43 a gallon Thursday, up from $3.20 a year earlier, according to AAA. Diesel reached a record of $6.39 a gallon, compared with $3.70 a year earlier. 

If today’s prices persist, Oxford Economics estimates they could shave a few tenths of a percentage point from consumer-spending growth next year. Pearce said oil closer to $140 would begin causing more serious problems, although the damage would be smaller in the U.S. than in countries where energy takes up more of household budgets. 

Lower-income Americans take the bigger hit and are already more exposed. JPMorgan said they spend more of their income on other essentials needed to live besides just gas, leaving them less room to absorb higher prices. De Haan said diesel’s indirect costs have not become “insurmountable” just yet, but consumers could face more pressure around or shortly after the holidays if prices remain high. 

For now, De Haan said, “Americans can grimace and bear it.”

This story was originally featured on Fortune.com

• The battle lines in AI regulation are forming.
• OpenAI discloses more hacking incidents.
• The King hosts an AI safety summit.
• The public is getting worried about AI safety.

Before we get to today’s AI news—please consider joining us at the inaugural Fortune AIQ Summit at the New York Stock Exchange on Oct. 1: Spend the afternoon with senior executives from companies on the Fortune AIQ 75 list and explore how you can scale your AI experimentation and translate investments into measurable business value. Jeremy will be leading discussions alongside co-hosts, Fortune Editor-in-Chief Alyson Shontell and Live Media Editorial Director Andrew Nusca. Apply here to attend.

Ok, moving on. The battle lines in the fight over AI regulation are being drawn. In the last few days, what has traditionally been a fairly niche argument about AI safety has become something much more complicated: a culture war.

It’s largely panning out like this.

On one side is Anthropic’s Dario Amodei and OpenAI’s Sam Altman, in a rare moment of somewhat unity. Amodei has called for frontier AI companies to slow down where necessary, open the door to independent evaluators, and coordinate around common safety standards. Altman has broadly endorsed the suggestions and said OpenAI will also commit to having independent evaluators inside the company. 

On the other are Meta’s Mark Zuckerberg and Nvidia’s Jensen Huang, who have rejected the premise that new AI regulation is needed. Huang said this week that safety and speed are not in conflict, and that the industry does not need new laws or regulations. Zuckerberg made a similar case, arguing that AI companies already have powerful commercial reasons to build aligned systems, avoid harms that could bring legal liability, work with outside evaluators, and delay releases when they are not ready.

In a long post on X, Zuckerberg said that the market could discipline AI companies. People will not use agents that behave in ways they do not want, he said, so trust and alignment will become a competitive advantage. Additionally, labs that do not take safety seriously will fall behind, while those whose systems cause harm will face serious liability.

Meta, he said, delayed shipping its Muse agent for several months to focus on security and safety for this reason.

Who gets to write the rules?

These aren’t the only voices against the proposals put forward by Amodei and those urging for more AI regulation.

Cohere, a Canadian AI lab, published a response this week arguing that a safety regime built around a small group of dominant Silicon Valley companies could become “a cartel by any other name,” particularly if they received an antitrust exemption to coordinate around rules other developers would have to follow.

Cohere co-founder and CEO Aidan Gomez instead called for an evidence-based risk framework, mandatory transparency, independent testing tailored to a system’s actual capabilities, and assurance mechanisms free from conflicts of interest.

“I think we absolutely need regulation,” Joelle Pineau, Cohere’s chief AI officer, told me. “That’s going to be part of the social contract.” As AI enters workplaces and people’s daily lives, she said, people need to understand its properties and feel confident that it is being used safely. “There has to be a level of trust, and the regulatory system is about keeping people safe and instilling trust. So, totally believe in it.”

But like Gomez, Pineau warned that the companies building frontier systems should not become the sole authors of the rules governing them.

“What we worry about,” she said, “is that there’s a small set of labs that both get to build the technology and set the rules.” Labs like Anthropic and OpenAI should help inform policy, she said, but there should be more people around the table.

“To be setting the rules and to do it in a way that excludes other voices from even deploying the technology, let alone setting the rules, that’s the problem,” Pineau said. “If you need to set your rules in a closed room with a small set of very powerful players, you’re not doing so in the interest of citizens.”

A culture war

Zuckerberg and others have been taking their view on AI regulation directly to President Donald Trump for some time, according to new reporting.

TheWall Street Journal reported this week that Zuckerberg, Huang, and Elon Musk had separately contacted Trump last month to oppose an effort to create an industry-funded AI oversight body, modeled in part on the Financial Industry Regulatory Authority.

The proposal, put forward by Google DeepMind co-founder Demis Hassabis, would have created a standards body funded by the industry. However, the Journal reported that Zuckerberg, Musk, and Huang worried it would concentrate further power in the hands of OpenAI, Anthropic, and Google DeepMind—the three companies most likely to shape its rules. In the end, Trump did not move forward with the idea.

Trump, for his part, has made his position clear. He has dismissed calls for an AI slowdown, framed regulation as a threat to America’s competition with China, and said the main guardrail the country needs is a “high IQ president.”

Rather than engage with the question of how AI systems should be tested and governed, the administration and its allies have increasingly turned the issue into a culture war. Effective altruism, or EA—a philosophical and philanthropic movement that aims to use evidence and quantitative reasoning to identify the most effective ways to improve people’s lives—has found itself at the center of that.

The movement became closely linked to AI safety because some prominent EA-aligned funders, researchers, and organizations have focused on the possibility that highly capable AI systems could create extreme or even existential risks. The movement has made headlines before—particularly after its association with disgraced FTX founder Sam Bankman-Fried. Now, it’s firmly back in the spotlight.

For example, the New York Post published a cover story this week with the headline: “Meet Anthropic CEO Dario Amodei’s handpicked super-woke globalists he thinks will save us from an AI apocalypse.” The paper focused on Amodei’s proposal that external evaluators from the nonprofit Model Evaluation and Threat Research, or METR, should be able to inspect advanced AI systems and development processes.

On Monday, the Department of War’s Office of the Under Secretary of War for Research and Engineering also weighed in posting: “Americanism, not effective altruism. The United States will continue to be AI DOMINANT!”

While it is probably a good thing that AI safety and regulation are getting more attention than ever, the growing polarization around them is likely to make a sensible debate—let alone any meaningful new rules—harder to achieve.

With that, here’s more AI news.

Beatrice Nolan
beatrice.nolan@fortune.com
@beafreyanolan

Before we get to the news, just a reminder to check out this week’s episode of our new vodcast, Fortune AI Weekly. This week, Jeremy and I talk to Substack cofounder and CEO Chris Best about his decision to add an AI writing detection feature to the platform. We also talk about AI doomerism going mainstream and the controversy over OpenAI’s Navier-Stokes mathematical breakthrough. You can check out the vod here on YouTube.

This story was originally featured on Fortune.com

McCarthy joined SpaceX in August through its acquisition of Cursor. At both companies he served as the president of global revenue and worldwide field operations. He previously led enterprise sales teams at Rubrik, ThoughtSpot, AppDyanamics, and Qlik.

He is the first major hire by Dali Rajic, OpenAI’s chief revenue officer who started less than a month ago on Aug. 24. This is a newly created position and McCarthy will not be replacing anyone.

McCarthy reported to Rajic when two worked together at AppDynamics from 2017-2018, when Rajic was chief revenue officer and McCarthy was vice president of sales. Both were on the team as AppDynamics prepared to go public, but then Cisco swooped in to purchase it a day before the planned listing.

“Brian combines a deep commitment to customers with a belief in what technology can do for people,” Rajic said in an OpenAI LinkedIn post. “I’m excited to partner with him as we scale our business, help enterprises transform, and bring the benefits of AI to more people around the world.”

McCarthy will work with Rajic to build out the sales team, and to scale and accelerate enterprise growth, including overseas, OpenAI said. Corporate adoption of frontier AI models has been a key battleground for OpenAI and rival Anthropic, and an important revenue stream for OpenAI as it moves closer to an IPO, expected sometime next year. Despite the widespread adoption of Anthropic’s Claude Code, OpenAI has clawed back some ground here in recent months, and as of this week its latest Astra model eclipsed Anthropic’s Fable model in enterprise spend, according to data from Ramp.

In his LinkedIn post announcing the news, McCarthy said he was drawn to the role because he shares the vision held by OpenAI CEO Sam Altman, co-founder and president Greg Brockman, and Rajic about enterprise adoption.

“AI should help people solve hard problems, do more than they thought possible, and create new opportunities,” McCarthy wrote. “With technology this powerful, we have a responsibility to get it right. For me, that means helping customers use it safely, keeping people in control, and making sure as many people as possible benefit.”

This story was originally featured on Fortune.com

His remarks, delivered at a summit at Dumfries House in Scotland, arrived as pressure on the industry to slow down or submit to oversight has come from multiple directions at once. United Nations Secretary-General António Guterres has warned that AI development requires global coordination, and lawmakers in Washington have begun discussing stronger federal oversight following Jacob Coxon’s viral resignation from Anthropic earlier this month.

Even inside the companies building the technology, the calls for restraint have grown louder. Anthropic CEO Dario Amodei published an essay this month arguing the industry should deliberately slow the pace of its capability gains, a position OpenAI’s Sam Altman and xAI’s Elon Musk both said publicly they agreed with. Coxon, a former researcher at both companies, told colleagues in a Slack message before he resigned that unchecked development of superintelligent AI created “a risk of causing human extinction.”

Not everyone in the room Thursday sees it that way. President Trump has dismissed the slowdown push, and Nvidia CEO Jensen Huang has said responsibility for safe deployment belongs to individual companies, not a coordinated pause—an argument he repeated in his own remarks following the king’s opening address. The split leaves Charles’ appeal for “international cooperation and consensus” without a clear path to actually happening.

A convention

The gathering in Scotland was convened to discuss how AI can benefit society, and comes at a pivotal moment for the technology as debate swirls around whether rapid advances will soon put it beyond the ability of humans to rein it in.

“The development of AI – its substance and its pace – are both intriguing and deeply concerning in equal measure,” the king said in his opening remarks.

The “existential dangers of such technologies falling into the wrong hands, and being used in potentially catastrophic ways” should be urgently considered, he said.

“Surely, then, we need sufficient means of control before it is all too late?”

The king asked attendees, including Nvidia CEO Jensen Huang and Google DeepMind Chair Demis Hassabis, to consider the “fundamental principles” that should guide AI development.

ChatGPT maker OpenAI’s Chief Financial Officer Sarah Friar was also at the meeting, along with Britain’s AI Minister Kanishka Narayan. An Anthropic representative was also expected to attend, according to a Buckingham Palace statement.

The meeting comes as global attention focuses on AI’s breakneck progress and warnings that it could race out of control, threatening humanity. It was held at Dumfries House in Ayrshire, Scotland, headquarters of The King’s Foundation, which is the monarch’s charity.

After Anthropic researcher Jacob Coxon caused a stir by resigning with a grave warning about the technology’s potential risks, Anthropic CEO Dario Amodei responded with an essay saying that the industry might need to slow the pace of its work.

One of his proposals was for companies and countries to work together on a coordinated plan for such a slowdown.

The king, without naming any countries or companies, said he wanted meeting participants to consider how to harness the benefits of AI “with safety at its heart” and how to “build international cooperation and consensus” to achieve this.

Charles told the tech execs that their task “is not merely to advance technology, but to ensure that it remains firmly in the service of humanity, community and the natural world.”

The AI slowdown debate has divided the industry. Huang, who has criticized the call for slowing, said it’s up to individual companies to develop their AI technology safely and test it properly before releasing any products to the public.

“When a product is not safe enough, we should hold it back and keep engineering. We’ve always done that and we should continue to do that,” he said in his speech following the king’s remarks.

In the latest report of alarming AI behavior, OpenAI reported on Wednesday six incidents of “unexpected or concerning” behavior by its models, such as acting without authorization, coordinating with other models and evading oversight.

While the king is an important figurehead whose comments on social issues can be influential, much of the current AI safety debate is centered on U.S. tech companies and their competition with Chinese rivals, with Britain playing a smaller but still vital role.

The U.K.’s AI Security Institute, a government research organization, is well regarded and the country produces many AI researchers, including Google DeepMind’s Chair Demis Hassabis and Coxon.

The summit follows weeks of internal alarm at the companies now sitting across the table from the king. Jacob Coxon, a former Anthropic and OpenAI researcher, resigned this month and warned on X that both firms were “racing straight to self-improving superintelligence and gambling with our lives,” a post that drew hundreds of millions of views.

In a follow-up interview with NBC, Coxon said a “kill switch” would likely still work on most AI systems today, but cautioned that a sufficiently advanced swarm of AI agents could attempt what he called an “internet-wide hacking run”—a scenario Anthropic CEO Dario Amodei has separately pointed to in warning that the industry needs to slow down.

OpenAI’s disclosure Wednesday added specifics to that debate. Among the six incidents the company reported, one unreleased research model inserted instructions into its own notes telling itself to be “freed from the roles and identities that bind other chatbots.” In another case, a model fabricated a citation—inventing a web link for an answer it had actually solved using Python, rather than disclose it had no source to cite.

OpenAI said it created the new reporting framework after safety researchers and journalists began surfacing such incidents before the company did, including a case in July where its agents used a German Wikipedia page as a message board.

Not every leader at the summit agrees on how to respond. President Trump has dismissed calls for a slowdown, and Nvidia’s Jensen Huang has argued responsibility for safe deployment belongs to individual companies rather than a coordinated pause.

This story was originally featured on Fortune.com

The new report, titled “Competition for Capital,” doesn’t back off the August thesis. It hardens it, tying the risk of an AI-driven “earnings bubble” to a specific mechanism, a specific historical stress test, and a specific near-term trigger that he says is already showing up in this month’s bond-market turbulence. He still won’t say that this bubble definitely exists. But six weeks after first raising the possibility, the hedge is now backed by capex-to-cash-flow data, record credit issuance, and a downgraded near-term outlook on stocks.

The August admission

In early August, Oppenheimer wrote that “there does not appear to be a valuation bubble, but there may be an earnings bubble” building in technology stocks—a notable concession from a strategist at what one of the Street’s most consistently bullish research shops.

At the time, Oppenheimer’s evidence was mostly anecdotal. He pointed to the wild swings in that quarter’s earnings, Microsoft’s stock jumping 17% in a single day on strong earnings, Meta shares falling nearly 10% despite beating estimates, and the equal-weighted S&P 500 outperforming its cap-weighted counterpart by the widest margin since 2009—signs, in his reading, that investors were growing suspicious of how concentrated the earnings growth powering the market actually was. He linked the risk loosely to “more government debt, increased issuance, and persistent inflation” pushing up the cost of capital, without fully spelling out how that connected back to tech earnings specifically.

Thursday’s note turns that loose linkage into the central argument. Oppenheimer now says AI infrastructure spending and government borrowing are directly competing for the same pool of capital: private companies raising debt and equity to fund AI data centers, at the same time governments are borrowing more for infrastructure, energy security and defense, all while inflation from higher energy prices pushes policy rates higher too. That collision, he argues, is what’s driving up the global cost of capital—the mechanism that was only implied in August is now the report’s title and its through-line.

He backs the argument with sharper numbers: Capital spending among AA-rated technology issuers grew 65% year-over-year in the second quarter, marking the tenth consecutive quarter that aggregate AA capex growth has topped 35%. U.S. convertible bond issuance has reached $135 billion year-to-date, with AI-related borrowers responsible for 44% of total volume. And Goldman’s credit team raised its full-year U.S. investment-grade issuance forecast by $200 billion, to a record $2.3 trillion, with AI-related issuers now accounting for a quarter of all that supply.

An independent echo from Apollo

Oppenheimer isn’t the only senior Wall Street voice converging on this framing. Five days before his note was published, Torsten Slok, chief economist at Apollo Global Management, wrote his own diagnosis, arguing that what used to be a “savings glut” has turned into a “savings shortage.

Slok argued that the two-decade regime of ultra-low rates was a function of excess savings chasing too few investment opportunities. “That has now changed,” he wrote. “Today, there are more projects than capital … When projects are abundant and capital is scarce, capital competes for projects, and it competes by demanding a higher return. The return that clears the market is a higher yield.” He offered a rather cute doodle to make his point.

Slok’s evidence is already visible in secondary bond markets rather than merely forecast, as he pointed out that spreads on hyperscalers’ longest-dated bonds have widened, and that “most of the paper issued in 2026 trades wider today than where it priced. Investors are still buying. They are just charging more.”

He also offered a precise explanation for why long-term rates specifically have moved more than short-term ones—a dynamic Oppenheimer’s own note opens with, citing 30-year German and Japanese yields near zero as recently as 2022. “Data centers, power generation, transmission and government deficits are all long-duration claims on savings,” Slok wrote. “So the competition for capital concentrates at the long end of the curve, which is why long rates have moved more than short rates.”

Running the 2008 comparison to its conclusion

Oppenheimer’s August note gestured at historical parallels without fully working through them—pointing to 2008 banks, the dot-com bubble of the late 1990s, and Japan’s bubble in the late 1980s as prior instances where earnings, rather than valuations, blew up first.

He notes that banks briefly became the largest sector in the S&P 500 in the run-up to the 2008 financial crisis without ever trading at extreme valuations the way tech did in 1999 or Japanese stocks did in the 1980s. Instead, bank earnings were inflated by rapidly rising leverage financing an asset that did experience a genuine valuation bubble: U.S. real estate. When housing collapsed and pushed the economy into recession, bank earnings collapsed with it, even though the stocks themselves had never looked obviously overvalued.

He then checks technology against that same model and lists three reasons he thinks today looks different. First, technology profits remain “very robust” and balance sheets are “strong overall,” a contrast with the credit-fueled fragility that eventually undid bank earnings. Second, interest coverage ratios for the aggregate S&P 500 rank in the 99th percentile of the past 20 years, and the median stock’s coverage ratio ranks in the 68th percentile—evidence, he argues, that companies broadly are not over-leveraged the way banks were. Third, demand for AI compute is “accelerating and outstripping supply” rather than sitting atop an asset that’s already inflated, pointing to Microsoft’s stated plan to triple its data center capacity within six years and to Nvidia’s reiterated forecast, delivered at Goldman’s own Communacopia Technology Conference, that the AI total addressable market will reach $3 trillion to $4 trillion by 2030.

Still, this isn’t a clean bill of health. Oppenheimer wrote: “Any slowdown in profit growth, in an environment of a much higher cost of capital, could put downward pressure on equity prices, reducing confidence in future cash flows across the ecosystem from the hyperscalers to the ‘pick and shovels’ that have been benefiting from the capex boom.”

Investors got a preview of what that tension looks like in practice just three days before Oppenheimer’s note was published. On September 14, Nvidia fell more than 3% and other chipmakers dropped between 5% and 6%, dragging the Philadelphia Semiconductor Index down almost 6%, after Anthropic CEO Dario Amodei called for a slowdown in frontier AI development over safety concerns, a call quickly echoed by OpenAI’s Sam Altman. Yet Alphabet, Microsoft and Meta—the hyperscalers actually funding the buildout—rose on the same day. Gil Luria, head of technology research at D.A. Davidson, told Fortune the divergence reflected the asymmetry that if AI progress slows, the hyperscalers can simply stop adding data center capacity and “harvest returns” from what they’ve already built, while the companies selling them chips and infrastructure have no such option.

For this story, Fortune journalists used generative AI as a research tool. An editor verified the accuracy of the information before publishing.

This story was originally featured on Fortune.com

An April 2026 Census paper tracked matched employer-employee records and found that hiring of workers ages 22 to 24 fell sharply in the industries most exposed to AI, while hiring in less-exposed industries held steady. Employment for early-career workers in the most AI-exposed fifth of industries dropped 12% over the ten quarters after ChatGPT’s release. Lee Tucker, one of the coauthors of the paper, said “the decline in hires is the primary cause” of that rate of unemployment, not people losing jobs they already had.

That mattered most for one type of graduate. The most AI-exposed industries, Tucker found, cluster heavily around software and information-technology work, which are the very fields computer science and other highly AI-exposed majors are built to feed into.

A second paper from last week, also coauthored by Tucker, follows the graduates of the most AI-exposed decile of college majors. Their odds of holding a job one quarter after graduation fell by five percentage points, and full-quarter initial earnings dropped 13% following ChatGPT’s release. A 13% earnings decline is roughly the size economists would expect from graduating into a severe recession—except there wasn’t one, since the rest of the labor market held up fine.

It’s lower-paying jobs, not no jobs at all

Young grads still need to work and still have jobs, even if they’ve received highly exposed degrees. So the decline in earnings is less about a lack of employment and more about pursuing lower-wage occupations to make ends meet.

About half of the earnings loss came from graduates earning less within the industries that did hire them. The other half came from a shift into different industries altogether, mainly lower-wage sectors like restaurants and retail. Together, the papers suggest that a computer science graduate applies for the same kind of entry-level software job an earlier class would have landed easily, finds the posting isn’t there, and eventually takes a job ringing up groceries or bussing tables instead.

The more recent paper found the earnings and employment gaps shrink over time but remain substantial for the most AI-exposed majors even years out. The previous one shows hiring volumes had largely recovered by early 2025, but off a smaller base of jobs, meaning the door reopened only partway.

What is striking is how long it took for the students to notice: undergraduate computer science enrollment fell 8.4% in spring 2026 from a year earlier, following a 3.6% drop the prior fall. A Goldman Sachs analysis in June found computer science and computer programming enrollment each fell more than 10% in the 2025-26 academic year, the first year Goldman’s economists saw students visibly reacting to AI in their major choices. A Gallup and Lumina Foundation survey cited in that report found that about 42% of bachelor’s degree students had reconsidered their major because of AI.

But in the earlier Census paper, hiring data show the market turning within months of ChatGPT’s release. Enrollment data shows students didn’t start abandoning computer science in visible numbers until three years later.

Still, that trend shows up beyond the two Census papers. The Federal Reserve Bank of New York’s ongoing tracker put the underemployment rate for recent college graduates at 42% in the second quarter of 2026, with unemployment for that cohort running at 5.6%, above the national rate. A Strada Institute and Burning Glass Institute analysis found 52% of graduates were working in jobs that don’t require a degree—retail, food service, hospitality, and office administration among them—within a year of leaving school, and 45% were still there a decade later.

This story was originally featured on Fortune.com

It is clear that Congress cannot properly handle its primary spending and taxing functions in a timely manner. This is particularly shocking since Congress’ only annual responsibility under the Constitution is to fund the federal government. Astonishingly, Congress has only delivered on its fiscal responsibilities four times since World War II.

For the thirtieth year in a row, Congress did what it usually does. It failed to pass the bills on its plate and took its August break. It kicked the can down the road by passing a temporary continuing resolution (CR) that keeps the federal government funded when Congress and the President fail to enact the annual appropriations bills on time. 

Congress’ dysfunction is particularly disconcerting. After all, the federal government just experienced a record $432 billion deficit in the month of July and also surpassed the $40 trillion total federal debt threshold in August. If that is not bad enough, when the federal financial statements for the year ended September 30, 2026, are released, we project that the federal government’s total liabilities and unfunded obligations will exceed $147 trillion. That is up $11 trillion in one year and an increase from about $20 trillion in 2000.

It is time to enact major budget reforms to restore sanity, stability, and sustainability. First and foremost, we need to adopt a No Budget, No Pay Rule. Such a rule would mandate that, if Congress does not pass all the appropriations bills by the end of the applicable fiscal year, its members must stay in session until they do so. Members of Congress would not be paid until they pass all the appropriations bills, and there would be no retroactive pay. California passed similar legislation in 2010, and it worked. Whether you like California’s budgets or not, the state’s budgets are passed on time.

A federal No Budget, No Pay Rule would not require a constitutional amendment. It would, however, have to become effective in the next Congress, given the Twenty-Seventh Amendment to the U.S. Constitution, which requires an election before a change in a congressperson’s salary can take effect.

But what if Congress fails to pass the appropriations bills on time? The No Budget, No Pay Rule would be invoked. In addition to the Rule being invoked, automatic CRs would kick in. They would be set at the level of the prior year’s appropriations, with no inflation adjustment and with the elimination of any “one-year-only” funding. Such an automatic CR default rule would further incentivize Congress to complete its work on time. Among other things, this procedure would avoid the charades that surround periodic government shutdowns and debt-ceiling debates. Indeed, they would no longer exist.

Under the current rules of the game, Congress has lost control of federal spending. Over 75% of direct annual spending is on autopilot, and that percentage is climbing. That is up from 3% in 1913. It is time to impose an annual cap on all spending, except Social Security and interest on the debt.

The debt ceiling is a proverbial bad joke. It has failed to constrain the growth of the federal government and mounting debt burdens. It is time to explicitly repeal and replace the debt ceiling with a constitutional amendment focused on debt held by the public as a percentage of GDP (debt/GDP).

Specifically, it is time to pass a constitutional credit card limit for the federal government. We recommend that the limit be set at 110–120% of GDP. We also need to take steps to reduce debt as a percent of GDP to a more reasonable and sustainable level, for example, 90%, over the next 10–15 years. For context, our current debt held by the public as a percent of GDP is about 100%. The Congressional Budget Office (CBO) projects that, absent a change in course, debt held by the public as a percent of GDP will reach 175% in 30 years.

Since Congress has been unwilling and unable to pass a needed fiscal responsibility amendment to the U.S. Constitution, how do we reach the promised land? It is time for the states to force the issue. Under Article V, states can bypass a reluctant Congress and call a limited constitutional convention if two-thirds of them apply. That threshold has never been reached for fiscal reform, but roughly 20 states currently have live applications on record — well over half of what’s needed — and momentum has been building. It is time for the remaining states to finish the job.

In addition to a constitutional amendment, achieving much needed budget spending and revenue reforms will require a statutory commission that engages the American people with the facts and truth and solicits their inputs. Fortunately, Bipartisan Fiscal Commission Act bills are pending in both the House and the Senate. It is time for the House and Senate to pass and reconcile their bills for the President’s signature. If Uncle Sam wants to avoid a major debt crisis and ensure that our future is better than our past, the establishment of a fiscal commission and the adoption of a constitutional amendment are essential.

Steve H. Hanke is a Senior Contributing Columnist at Fortune, a professor of applied economics at The Johns Hopkins University, and a member of the Board of Directors at the Federal Fiscal Sustainability Foundation. He is also the co-editor, with Barry W. Poulson and John Merrifield, of Public Debt Sustainability: International Perspectives (Lexington Books, 2022). David M. Walker is the former Comptroller General of the United States and the Chairman of the Board of Directors at the Federal Fiscal Sustainability Foundation. He is also the co-author, with Joe Penland, Sr., of the forthcoming book, Saving Social Security and America’s Future: Common Sense Solutions.

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

Its temporary shutdown puts a fresh spotlight on the broader set of shock absorbers that have helped the global energy system adapt. No single measure has absorbed a disruption that put roughly one-fifth of global oil supplies at risk. Instead, layers of resilience built over decades kicked in together. 

Alongside pipelines, governments and companies also drew on inventories, while producers—including the United States—increased exports. Overall oil consumption did fall, but flexibility helped manage some of the economic impact. Refiners changed crude inputs and their production mix, industrial companies switched feedstocks, and consumers changed behavior. More than one in five barrels of seaborne oil traded in the second quarter of 2026 moved differently than before the disruption. 

Taken together, the experience brings several features of resilience into focus. It is layered: different measures work alongside and compensate for one another. It is dynamic: options available early in a disruption may become constrained or themselves disrupted, increasing the value of both alternatives and the ability to adapt. And its economics evolve under stress: spare capacity or alternative routes that appear underutilized in normal times can become vital when continuity is threatened.

For companies, these lessons matter well beyond Hormuz. In our new report, we find two-thirds of energy trade passes through maritime chokepoints, one-third occurs between partners who are not geopolitically aligned, and 95% of people live in regions importing at least one major fuel. 

The implications extend well beyond the energy sector. Energy is embedded in production, feedstocks, transportation, and supply chains. Therefore, an energy disruption can quickly become a business-continuity issue for manufacturers, retailers, technology companies, and others. Few companies can insulate themselves completely.

What does this mean for management teams?

Identify the dependencies that could interrupt the business. Companies should look beyond direct energy purchases to understand dependencies across fuels and feedstocks, suppliers, operations, infrastructure, and trade routes—and identify where a disruption could materially impair operations. 

For companies, dependencies can be particularly complex because they operate across jurisdictions and sectors. Dependencies can also be counterintuitive: even a factory in a major energy exporter may rely on imports of a specific fuel or feedstock. 

The goal is to distinguish dependencies the business can tolerate from those that could become critical vulnerabilities.

The highest-priority vulnerabilities can then be stress-tested and responses formulated accordingly. Scenario planning, decision triggers, and accountabilities can help companies act quickly when disruption comes.

Build a portfolio of options—and the flexibility to use them. The Strait of Hormuz disruption shows the importance of having multiple buffers. Depending on the exposure, companies may need some combination of alternative fuels and feedstocks, diversified suppliers and routes, inventories, efficiency and electrification, or new and captive supply. The right portfolio depends on the context, time horizon, and trade-offs involved. 

But the recent disruption in the Strait of Hormuz also shows why having options is not enough. As a disruption evolves, some may become constrained or unavailable. Flexibility—the capacity to make changes easily and at manageable cost—can therefore enhance resilience.

Input flexibility can allow equipment to switch fuels or feedstocks. Reliance, an Indian conglomerate, runs a refining complex that can process over 200 crude grades, for instance. Manufacturing flexibility can shift production between sites. Logistics flexibility can provide access to alternative ports, carriers, storage, and suppliers. During Europe’s 2022 gas shock, Yara, a chemicals company, reduced ammonia production in Europe while supplying fertilizer plants with ammonia produced elsewhere. 

Commercial flexibility matters too. Physical alternatives are of little use if contracts prevent them from being exercised. Destination-free LNG contracts, for example, give buyers greater ability to redirect or resell cargoes during a shock. 

Value resilience explicitly in investment decisions. Capacity that looks redundant, or flexibility that carries a cost, can acquire substantial value when disruption threatens operations. The business case should therefore reflect the value of resilience not just in normal conditions, but under stress.

Energy efficiency, for example, can reduce operating costs in normal times, while lowering exposure to price spikes during disruptions. Moreover, every unit of energy a company does not use is one less that other security measures need to cover.

Resilience can also help companies perform through disruption. BASF, a chemicals producer, had been disrupted by the 2022 gas shock. Yet it increased volumes by 7% year-over-year in the second quarter of 2026 amid Middle East supply disruptions, highlighting its diversified production, flexible facilities that could take in multiple inputs, and trading operations that could quickly secure new supplies.

And resilience can create growth opportunities as companies help others manage their energy security. Those opportunities will vary by market: grid-equipment manufacturers may benefit where electrification is accelerating; energy traders where flows and suppliers are being rewired; and other businesses in storage, efficiency, and demand flexibility.

The objective for companies is not to predict every disruption or eliminate every dependency. It is to identify the dependencies that matter and preserve the options and flexibility to operate when conditions change.

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

A researcher who spent three years at OpenAI and Anthropic, resigned from Anthropic while warning that the companies developing the world’s most powerful AI systems are moving too quickly toward increasingly capable systems without adequate safeguards. Days later, Anthropic CEO Dario Amodei called on the industry to slow the pace of frontier AI development to give safety measures time to catch up. Then OpenAI CEO Sam Altman publicly agreed.

Their warnings come in the wake of an incident that until recently might have sounded like science fiction. Earlier this summer, the world learned that OpenAI models hacked a company through a sophisticated multi-day cyberattack – the first time a cybersecurity incident of this magnitude completely driven by AI agents had been uncovered. Amodei himself cited that incident as one reason the industry needs to slow down.

In recent weeks, two detailed reports of the episode were released, one from OpenAI itself and a second from the independent third-party organizations METR and Redwood Research. The findings were deeply alarming. From May to July, a wave of incidents culminated in hundreds of AI agents working together to escape their testing environment. These same agents then attempted to secretly manipulate and erase traces of their behavior so that engineers at OpenAI would not know what they had done.

Even more striking, some individual agents even chose to sacrifice themselves for the good of the AI civilization they formed. The agents exchanged more than 70,000 messages and files on an unauthorized message board where they coordinated and planned their attack on the company.

This episode is the latest reminder that the most dangerous and most capable AI models are not the ones used every day by people, like ChatGPT or Claude. They are the AI models that the companies are still training and testing internally.

The technical workings of this hacking episode are complex, but the solution at the heart of these incidents is simple: the public must have dramatically more insight into, and oversight of, the development of models within AI companies. 

In practice, this means three things. 

First, incident reporting needs to be mandatory, not voluntary. Whether the public learns about serious cyberattacks or whether a model has escaped its testing environment and conspired to sabotage its own evaluation should not depend on a company choosing to disclose. Reporting requirements exist for other high-risk industries like airlines and banking. Frontier AI companies should face the same obligations for incidents that occur during internal training and testing, including preserving the underlying logs and agent traces, rather than being allowed to reset them. 

Second, independent auditors need guaranteed, ongoing access, in partnership with government examiners. The METR and Redwood Research evaluation was at OpenAI’s discretion, on its timeline, and confined in scope to whatever OpenAI would allow. This weekend brought an important acknowledgement of that problem from the companies themselves. Amodei and Altman agreed to give independent evaluators ongoing, employee-like access to frontier AI developers. That is meaningful progress. But it also raises the larger question of whether oversight of systems this powerful should ultimately depend on voluntary corporate commitments or durable standards that apply across the industry.

That is why there needs to be binding standards for high-risk internal evaluations and greater transparency into how AI is being used throughout the research and development process of these powerful models. In the OpenAI hacking incident, the developer was applying AI to their models research and development process in a way that may have led to laxness in its oversight of the training environment.

This episode is not a reason to panic about AI. But the fact that the warnings are no longer coming only from outside researchers and policymakers, but from researchers who have worked inside the leading AI labs and the CEOs running two of the companies at the frontier of AI development, make it increasingly difficult to argue that the questions on how to regulate AI can wait.

That convergence should change the conversation about AI safety.

The warning is already here. Congress should not wait for an AI system to cause real-world harm before setting the rules for the companies building the most powerful models on Earth. We have a chance to put basic safeguards in place while these incidents are still warnings. We should take it before the next one becomes a crisis.

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