All News (Page 434)

Ireland’s Prime Minister, or Taoiseach, Micheál Martin is expected to unveil more than $6 billion in deals Wednesday when he meets with U.S. President Donald Trump in Washington. But the usual St. Patrick’s Day festivities will no doubt be dampened by the Iran war and lingering tensions over trade, tariffs, and Irish policies that the White House has called a “tax scam” for U.S. companies.

Few understand the challenges better than Michael Lohan. As CEO of IDA Ireland, the nation’s foreign direct investment agency, Lohan is charged with attracting companies to a country that has long relied heavily on U.S. capital and companies. “You know, Ireland is closer to Boston today than Berlin,” says Lohan, repeating a long-held trope about its economic similarities to the U.S. in terms of taxes, talent, and ease of doing business. (Technically, Boston is almost three times as far as Berlin.) “Last year was a record year for FDI investment in Ireland—against the backdrop of everything that was happening—and 65% of that investment came from U.S. multinationals.”

That flow of capital is not viewed as a good thing by Trump, who accused Ireland of “taking our pharmaceutical companies” during last year’s St. Patrick’s Day meeting. While his pressure on U.S. companies to double down at home is being heard–U.S. FDI to Ireland dropped 20% in 2024 to $467 billion–America remains the country’s largest investor. Lohan’s job is to attract more of that capital by “telling the story of Ireland” as a place to get talent, agility and easy access to the 27 member states of the European Union.

Pharma-fueled trade surplus

Ireland has proven to be a desirable place to book profits and pay taxes. The country’s 12.5% corporate income tax rate, and prior tax benefits for companies like Apple, have generated both investment and unwanted attention. Among other things, Apple and Microsoft’s intellectual-property rights are held in Irish subsidiaries that collect royalties from elsewhere. Pharmaceutical companies like Eli Lilly manufacture key ingredients of blockbuster drugs like Mounjaro and Zepbound on the emerald isle for the same reason, shipping those drugs to U.S. consumers and booking the revenue in Ireland.

Ireland’s budget watchdog says three U.S. companies accounted for almost half of the country’s corporate tax revenues last year. Although unnamed in the report, they’re known to be Apple, Lilly, and Microsoft. Lilly, for one, paid $6.6 billion in tax to Ireland in 2025, about double what it paid in the U.S.—a country with 65 times the population and the bulk of its customers. With 4,000 employees in Ireland, Lilly’s workforce is also less than a fifth the size of its U.S. operations. Pharmaceutical sales helped Ireland’s exports of goods to the U.S. grow 52% last year to about $132 billion, more than doubling the goods trade surplus to $114.2 billion. (Trade in services between the two countries is essentially the opposite, with Ireland buying more than it sells.)

One man’s trade surplus is another man’s trade deficit, especially if that man is Donald Trump. The U.S. President has paid particular attention to physical goods when it comes to trade flows, and has called out Big Pharma for rising drug costs. Even with the U.S. now at war, Ireland’s reputation as a corporate tax haven is unlikely to escape attention during the White House visit.

That may be why the prime minister, much like his IDA emissary Lohan, has shifted the emphasis from inbound investment to money flowing the other way. Lohan talks about how Ireland invested a historic $389 billion in the U.S. in 2024, making it America’s fifth largest source of FDI. On a per-capita basis, the nation of 5.4 million claims to be number one. “The U.S. continues to be the most innovative economy in the world. It continues to be where capital is readily available and supportive,” says Lohan. “None of those things have really changed.”

What has changed, of course, is Trump’s focus on “America First,” which is why Martin is expected to present $6.1 billion in new Irish investments to the U.S. alongside the traditional bowl of shamrocks. While nurturing European alliances is also not a priority for the White House, Trump’s calls for NATO and Europe to step up in protecting the Iran-controlled Strait of Hormuz could occupy much of Martin’s discussions with Trump. Ireland will begin a six-month stint in holding the presidency of the Council of the European Union in July, which will give it a central role in E.U. decision making.

Martin may want to talk trade this time as the war with Iran is an issue that few leaders want to tackle in public, especially during a White House press conference. IDA’s Lohan is also not oblivious to the fact that consumer sentiment in Ireland is decidedly mixed when it comes to Trump, tariffs, and the tech giants that have raised the cost of housing while putting pressure on the energy grid at home. And Europe’s approach to tech innovation is decidedly different than what’s coming out of D.C.

“We want to push the innovation and technology agenda, but we have to do it safely and ethically,” said Lohan. And that applies to all potential investors, including China.

“We want to see a fair, level playing pitch between China and its counterparts, with Ireland being part of that,” he said. “But I do think we can’t turn our back on what is a very significant economy where there is a significant amount of innovation.”

This story was originally featured on Fortune.com

Businesses are pouring billions into AI to boost productivity and cut costs—and to fund it, they’re slashing hiring, training, and employee support. Headcount isn’t safe either: as Jack Dorsey’s Block recently demonstrated, a growing number of executives are citing AI as justification for substantial layoffs.

That approach may lift short-term margins. But it is a dangerous long-term strategy—and the data makes clear why.

As president of SHRM Foundation—the philanthropic arm of the largest HR association in the world— I’ve seen firsthand how organizations thrive when they invest in human potential alongside AI. Technology can accelerate work. Competitive advantage comes from the judgment, adaptability, and trust that only people provide.

The gap is glaring. Nearly three-quarters of knowledge workers globally now use AI at work, yet 60% say they have not received formal training to use it effectively. AI spending is projected to rise 44% in 2026—while training budgets are expected to grow just 5%, and average learning time is actually falling—from 47 to 40 hours per employee. Companies are deploying powerful tools while quietly disinvesting in the humans required to use them.

At the same time, employees are navigating rising pressures inside and outside the workplace. Burnout and stress remain widespread, the specter of AI-driven layoffs is increasing workers’ sense of precariousness, and millions of Americans balance their jobs with responsibilities like caregiving outside of the workplace. Without support, these pressures carry real costs: Gallup estimates disengaged and stressed workers cost the global economy nearly $9 trillion annually. Unaddressed stress drives absenteeism, presenteeism, and turnover — hidden costs that can exceed an employee’s annual salary. Against this backdrop, it’s no surprise that ADP’s employee motivation index just reported its sixth straight month of decline.

In an AI-driven economy, unlocking business potential and long-term growth requires investing in human potential. That means not just keeping workers around, but equipping workers with the skills to use emerging technologies effectively. It also means addressing the conditions that determine whether people can bring their full capacity to work: ongoing skills development, mental health support, caregiving flexibility, financial stability, and workplace cultures that foster trust and psychological safety.

The Business Case Is No Longer Optional

Employers are uniquely positioned to provide this support — and the business case is increasingly clear. Johnson & Johnson’s long-running employee wellness initiatives generate an estimated $250 million in healthcare savings and returns nearly $3 for every $1 invested. Companies providing childcare support have reported returns exceeding 400% through improved retention and productivity. Organizations that support employee well-being and life responsibilities see stronger retention, higher productivity, and better long-term performance.

By contrast, companies that reshape their workforce around AI and treat workforce investment as discretionary spending often face higher turnover, lost productivity and prolonged vacancies, more safety incidents, and weakened customer experience — costs that compound quietly but relentlessly.

The question for business leaders is no longer “How can AI help us automate more tasks and reduce headcount?” The smarter question is: “Which human capabilities become more valuable as AI absorbs routine work—and how do we redesign roles to strengthen them?”

What Happens When AI Maximizes People Instead of Replacing Them

The productivity upside is real—but only under the right conditions. Research shows tasks completed 25% faster and with 40% better quality, 60% greater productivity, up to 36% more time for higher-order work, and more effective—and cheaper—learning and development programs. But these gains only materialize when workers are trained, supported, and trusted to apply judgment.

Some companies are already showing what this looks like in practice. IBM CHRO Nickle LaMoreaux, bucking the layoff trend, announced plans to expand entry-level hiring and redesign roles around durable skills.

“The companies three to five years from now that are going to be the most successful,” LaMoreaux said, “are those that doubled down on entry-level hiring in this environment.”

IBM isn’t alone. Amazon has committed more than $1.2 billion to upskill hundreds of thousands of workers for technology-enabled roles. Mastercard has deployed an AI-driven internal talent marketplace to match employees to growth opportunities, reducing external hiring costs while increasing retention. SAP has embedded continuous learning time and wellbeing supports into the workweek to sustain productivity and attract scarce talent.

Short-term gains may come from cutting labor costs and accelerating automation. But long-term performance depends on resilience, trust, institutional knowledge, and the capacity to adapt—all of which are built through sustained investment in people.

AI will shape the future of work. Humans will drive it. The organizations that come out ahead will be the ones that treat workforce investment not as a line item to cut, but as the strategy itself. The $500 billion bet on AI only pays off if the people running it are trained, supported, and set up to succeed.

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

Earlier this month, Andrej Karpathy, a well-known AI researcher who was one of the founding employees of OpenAI and later headed up AI for Tesla, went viral on X. This alone isn’t so unusual. Karpathy—who now works as an independent AI researcher and is also the founder of Eureka Labs, which says it is creating a new kind of school for the AI era—has 1.9 million followers on X and his reputation is such that almost anything he says about AI is treated as either gospel or prophecy.

But this post was about an experiment he’d run where put an AI coding agent to work running a series of experiments to figure out how to improve the training of a small language model. He let the AI agent run continuously for two days, during which time it conducted 700 different experiments. Over the course of those experiments, it discovered 20 optimizations that improved the training time.

Karpathy found that applying the same 20 tweaks to a larger, but still fairly small, language model resulted in an 11% speed up in the time it took to train the model. Karpathy called the system he built for conducting this experiment “autoresearch.”

Tobias Lütke, the cofounder and CEO of Shopify, posted on X that he tried autoresearch to optimize an AI model on internal company data, giving the agent instructions to improve the model’s quality and speed. Lütke reported that after letting autoresearch run overnight, it ran 37 experiments and delivered a 19% performance gain.

What caught many people’s attention was that the autoresearch is close to the idea of self-improving AI systems that were originally broached in science fiction and that some AI researchers fervently desire and others deeply fear. The concern is that “recursive self-improvement,” where an AI continually optimizes its own code and training in a kind of loop, could lead to what AI safety researchers sometimes call a “hard takeoff” or an “intelligence explosion.” In these scenarios, an AI system rapidly improves its own performance, leading it to surpass human cognitive abilities and escape human control.

Karpathy’s experiment wasn’t quite this. The AI agent at the heart of autoresearch set up isn’t refining its own training set up, it’s adjusting the training code and initial neural network settings for a different, much smaller and less sophisticated, AI model. But Karpathy rightly noted that his experiment had big implications for how AI labs will do research going forward, and this might accelerate their progress.

“All LLM frontier labs will do this. It’s the final boss battle,” Karpathy wrote on X. He acknowledged that “it’s a lot more complex at scale of course,” since his autoresearcher only had to worry about adjusting a model and training process that was contained in just 630 lines of Python code, whereas the training codebase of frontier AI models is orders of magnitude bigger. “But doing it is ‘just engineering’ and it’s going to work,” he continued. “You spin up a swarm of agents, you have them collaborate to tune smaller models, you promote the most promising ideas to increasingly larger scales, and humans (optionally) contribute on the edges.”

He said that while the current autoresearch system he built was designed for a single agent to continually improve a piece of code along a single path, in the future he imagines multiple AI agents will be able to explore different optimizations and different experiments in parallel. “The next step for autoresearch is that it has to be asynchronously massively collaborative for agents,” he wrote. “The goal is not to emulate a single PhD student, it’s to emulate a research community of them.”

Karpathy also said something else about autoresearch which got many people excited. “*any* metric you care about that is reasonably efficient to evaluate (or that has more efficient proxy metrics such as training a smaller network) can be autoresearched by an agent swarm,” he wrote. “It’s worth thinking about whether your problem falls into this bucket too.”

Some commentators pointed out that the basic components of autoresearch could be used for many other agentic systems to optimize a process. Janakiram MSV, principal analyst at Janakiram & Associates, writing in tech publication The New Stack called this “the Karpathy Loop.” It has three components: an agent with access to a single file that it can modify; a single metric, objectively testable metric, that the agent can optimize for; and a fixed time limit for how long each experiment can run. He also highlighted that the instructions Karpathy gave the AI agent in autoresearch were also good models for anyone interacting with any AI agent. The plain text file Karpathy used included clear instructions for what the agent should do, constraints, telling the agent what it should not do or change, and a stopping criteria, indicating how long each loop should run and when the agent should stop looping and report its results.

But some critics said that Karpathy had done little more than rediscover part of a process known as AutoML that researchers at Google, Microsoft, and other AI labs have already been using for years. AutoML also uses an optimization loop and series of experiments to find the best data to use for AI, the best model architecture to use, and to tune that model architecture. But it doesn’t use an AI agent that can read AI research papers and develop hypotheses for which improvement to make. AutoML systems tend to depend on random variations or various evolutionary algorithms to decide which changes to try. 

Karpathy replied to some of these comments, saying that some AutoML methods, such as neural architecture search, which is an automated way to optimize the design of an AI model, were not nearly as powerful as his autoresearch. “Neural architecture search as it existed then is such a weak version of this that it’s in its own category of totally useless by comparison,” he wrote. “This is an *actual* LLM writing arbitrary code, learning from previous experiments, with access to the internet. It’s not even close.”

This story was originally featured on Fortune.com

Six of California’s 214 billionaires have been widely reported to have left the state in time to avoid a proposed 5% wealth tax—but that small cohort would have collectively generated $27 billion in tax revenue, roughly a fourth of the initiative’s projected $100 billion haul.

Last November, panic erupted over the announcement of a proposed billionaire’s tax in California. The tax would levy a one-time 5% tax on the net worth of California residents with assets worth at least $1 billion. California’s progressive governor, Gavin Newsom, emerged as the measure’s biggest opponent and vowed to stop the tax to “protect” the state’s tech industry and overall economy. 

Before the Jan. 1, 2026, cutoff proposed in the initiative, Google cofounders Larry Page and Sergey Brin, and venture capitalist Peter Thiel, left California for Miami. Car loan magnate and L.A. native Don Hankey left the state for Las Vegas. Former Uber CEO Travis Kalanick recently announced he had left California for Texas in December. Director Steven Spielberg became a New York City resident on New Year’s Day, according to the Los Angeles Times, although his representative said the Jaws director had long planned to move to be closer to family.  

The tally of departed billionaires likely understates the extent of the flight. Meta CEO Mark Zuckerberg has also reportedly left the state, but not before the Jan. 1 deadline. Venture capitalist David Sacks, whose net worth has been reported to range from $250 million to $2 billion, also left the state as his company Craft Ventures moved to Austin. Zuckerberg would take another roughly $10 billion of tax revenue with him. 

If the state were to tax Page’s $260 billion net worth at 5%, it would rake in $13 billion in tax revenue. Brin’s taxes would bring in about $12 billion. While Thiel, Kalanick, and Hankey may not rank among the top five richest men in the world, together they would have generated $1.775 billion.  

The loss of a fourth of the proposed tax revenue is a major hit to the initiative, which intends to use the funds toward health care, education, and food assistance.   

Billionaires are backing a fight 

Billionaires in and outside of California are working to fight the tax, which has been a harbinger for more wealth taxes across the country. 

Brin donated $20 million to a group called Building a Better California that is giving out $15 to people who sign their three countermeasures. The group’s proposals would prevent retroactive taxes and narrow the definition of California residency to fight against the 2026 Billionaire Tax Act’s application to anyone who lived in the state as of Jan. 1, 2026. 

Two billionaire-backed political action committees, Stop the Squeeze and Golden State Promise, have launched to stop the proposal. 

Chicago-based venture capitalist Daniel Tierney donated $200,000 to Stop the Squeeze, and crypto billionaire Chris Larsen is backing Golden State Promise, the New York Times reported. 

Since the California initiative was announced, other states have proposed higher taxes on their high-earning residents. In January, Rhode Island Gov. Dan McKee backed a 3% tax increase on millionaires. Last week, Washington, which is one of nine states without an income tax, passed a 9.9% tax on personal income above $1 million per year. 

“We’ve got more millionaires and billionaires than we’ve ever had, and they’re paying, effectively, a 4% tax rate,”  Rep. Brianna Thomas, of Seattle, a Democrat who supported the measure, previously told Fortune. “Meanwhile, you got working folks paying 11% of their income, and the lowest-income people paying 14%. Isn’t it unfair for those who have the most, to pay the least, and those who have the least to pay, the most, proportionally?”

This story was originally featured on Fortune.com

It’s a scene straight out of a science fiction show: robot dogs. Think K9 from the sci-fi series Doctor Who, or Goddard from the cartoon Jimmy Neutron.

Now, robot dogs are standing guard for tech companies, patrolling the massive data centers across the country that power AI operations, according to Business Insider. These four-legged robots, known as quadrupeds, are in high demand from AI firms, according to robotics company Boston Dynamics, which manufactures a quadruped called Spot. These systems are able to navigate complex landscapes on their own, alert authorities about security threats, and can provide around-the-clock video surveillance.

“We’ve seen a huge, huge uptick in interest from data centers in the last year,” Merry Frayne, senior director of product management at Boston Dynamics, told Business Insider, “which is probably not surprising given the investment in that space.”

Companies are pouring nearly $700 billion into the AI infrastructure buildout, a sum that rivals the GDP of developed countries like Sweden. And some data centers are the size of multiple football fields. One data center—Meta’s Hyperion—will sprawl out to about four times the size of Manhattan’s Central Park. Aside from requiring loads of energy and millions of gallons of water, the vast size of the data centers means the cost of security to protect their around-the-clock operations is inspiring some firms to look to alternative security resources.

According to Frayne, Spot’s pricing ranges from $175,000 to $300,000, depending on their client’s needs. But despite that high price, the company estimates that the quadrupeds would compensate for their cost within two years.

The robot dogs are actually capable of doing more than just perimeter patrol. Frayne told Business Insider data center customers are looking for the quadrupeds to conduct industrial inspection, site mapping, and construction monitoring. These tasks could help facility managers to more easily detect hazards, such as puddles or leaks. Boston Dynamics says Spot has “360° perception and athletic intelligence.”

Quadrupeds like Spot have actually existed for some time now, assuming roles in public safety and law enforcement. Another robotics company—Ghost Robotics—advertises quadrupeds as a business solution for construction sites to streamline inspections and enhance safety monitoring. The company also advertises the robots for reconnaissance, intelligence, and surveillance use by the military. 

Forget the age of AI—the dawn of the robotics era

Some tech leaders predict the AI revolution could usher in a new era of robotics, with some predicting they’ll soon outnumber humans. The current state of robotics is a bit far off from that reality. A Deloitte research report titled “AI for industrial robotics, humanoid robots, and drones” found that annual sales of new industrial robots have remained flat since 2021, at roughly 500,000 units. 

But their longer-term projection suggests massive growth in the future, with robot shipments doubling to 1 million by 2030 and revenues of $21 billion. That prediction jumps to $5 trillion by 2050.

In a recent interview with Fortune, Zak Kidd, founder of AI company AskHumans—which has been used by organizations like the World Bank and Fidelity—said that while AI threatens white-collar work, robots could one day poach jobs that require physical labor.

“I see AI as an augmentation of knowledge work,” he said.  “But I see robotics, humanoid robotics, as a replacement for manual work.”

This story was originally featured on Fortune.com

U.S. President Donald Trump is considering delaying a key meeting with Chinese President Xi Jinping “by a month or so” as he struggles to manage the surging conflict with Iran.

The meeting was set to take place between March 31 and April 2, building on the two leaders’ previous face-to-face dialogue in South Korea last October. 

On Monday, Trump pushed back against claims that he was considering postponing his visit to pressure China to intervene in the Strait of Hormuz, a key strategic waterway currently closed by Iran. “I’m looking forward to being with [Xi],” Trump told reporters at the White House on March 16. “[But] it’s very simple, we’ve got a war going on, and I think it’s important that I be here.”

Still, a delay to the meeting will mean that Trump and Xi will have to wait to discuss a number of factors dragging down the U.S.-China relationship, such as China’s continued export controls on critical minerals, the U.S.’s export controls on semiconductors, and U.S. demands that China buy more agricultural products.

Analysts say the U.S. President’s decision appears driven by the Iran conflict—and a need to manage a fast-escalating conflict and the fallout in energy markets—instead of an attempt to pressure China. 

“Trump’s delay seems to be genuinely about managing the Iran war,” argues Kyle Chan, a fellow at the Brookings Institution, an American think tank. “The Iran war has escalated dramatically, and it would make sense for the U.S. commander-in-chief to give this his full attention.”

Kevin Chen, an associate research fellow at Singapore’s Nanyang Technological University (NTU), also thinks the mooted delay is not due to China’s unwillingness to help the U.S. unblock the Strait of Hormuz, especially given that the U.S. Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng “just had a very productive meeting in Paris”. 

During the meeting, China expressed willingness to buy more agricultural produce from the U.S., Reuters reported, citing unnamed sources. U.S. officials also said there were discussions about setting up new mechanisms, like a U.S.-China “Board of Trade” to manage the economic relationship between the two countries. 

A shift in priorities

Some Trump officials had suggested withdrawing from some U.S. diplomatic endeavors, like its support for Ukraine, in order to devote more resources and attention to countering China. 

But now analysts say Trump’s move indicates that the Iran war—and the resulting blockage of oil and LNG exports from the Gulf—has eclipsed the U.S.’ other geopolitical priorities. 

“President Trump’s move to delay the late-March summit with Xi reflects a shift in priority towards the ongoing military campaign in Iran,” says Dylan Loh, an international relations expert from NTU. 

The U.S. likely has a narrow window to shape events in Iran. “The next two to three weeks are likely to be the most critical period before the Middle East situation stabilizes,” explains Khuong Minh Vu, a public policy professor from the National University of Singapore, adding that the United States and Israel will try to weaken Iran’s strategic capabilities related to nuclear weapons and missile systems.

Vu adds that Trump may also want the Iran conflict to be resolved prior to his negotiations with China to strengthen his bargaining position, particularly if Trump successfully pressures Iran to accept a negotiated agreement.

China has yet to comment on Trump’s proposal. But Beijing might be pleased with a delay; originally, it had requested a later meeting date to give officials more time to prepare. 

“I don’t think the fallout will be large,” says Loh, from NTU. “China will be patient.”

This story was originally featured on Fortune.com

As the Iran war drags deeper into its third week, one seemingly obvious solution for more energy is crude oil from Venezuela after the Trump administration seized former leader Nicolás Maduro and pressed for the reopening of the nation’s oil sector.

The glaring problem is more oil from Venezuela—or any other source around the world—represents only metaphorical drops in the global supply bucket compared to the massive losses each day from the Persian Gulf and the effective closure of the Strait of Hormuz by Iran.

“It’s a math problem,” said Fernando Ferreira, director of the geopolitical risk service at Rapidan Energy Group. “Hormuz flows about 20 million barrels [of oil] a day. Venezuela is currently producing about 1 million [barrels daily].”

The issue is there simply are no alternatives to the de facto closure of the passageway that sees about 20% of the world’s oil and liquefied natural gas trek through it each day.

“Venezuela helps; every little bit helps. But, in the grand scheme of things, it doesn’t change the equation,” Ferreira told Fortune. “There is no medium-term solution other than reopening the straits. Nothing else is going to solve the crisis.”

Arguably the best-case scenario for Venezuelan oil production is it grows from producing nearly 1 million barrels of oil a day late last year to churning out about 1.2 million barrels daily by the end of 2026, said Francisco Monaldi, director of the Latin America Energy Program at Rice University’s Baker Institute for Public Policy.

“I’m expecting less than 250,000 barrels added over the whole year, if at all. That is of course significant for a country that produces just 1 million, but it’s nothing for the world market. It’s less than 0.3%,” Monaldi said, considering the world consumes about 103 million barrels a day. “In particular, it’s very insignificant compared to the disrupted market.”

In the meantime, the White House is aiming to build a coalition of allies to control the strait and escort tankers. The U.S. is also temporarily lifting sanctions on some Russian oil—but that only impacts the destination and prices, not the volumes of oil. And member countries of the International Energy Agency agreed to release a record-high, 400 million barrels of oil from strategic reserves, including 172 million barrels from the U.S.

Pulling that oil from storage will take at least four months however. And while the planned emergency releases are helping keep oil prices from hitting all-time highs, crude oil benchmarks are still hovering near $100 a barrel—up almost 70% from the beginning of the year.

The average price of a gallon of regular unleaded gasoline is $3.80 and rising in the U.S.—up nearly 40% since its January low—but that’s nothing compared to the Asian nations suffering from much higher prices and long lines for fuel, closed schools, and shortened work weeks because of their greater reliance on Middle Eastern oil and Qatari natural gas.

The most successful approach thus far is Saudi Arabia and the United Arab Emirates redirecting as much of their oil flows as they can away from the Strait of Hormuz via the Saudi Arabia East-West Pipeline and the UAE’s Habshan–Fujairah pipeline.

Still, close to 14 million barrels of oil per day remain blocked, according to energy analysts.

“If those pipelines are attacked, then it could be even worse,” Monaldi said.

An Iranian drone attack hit Fujairah on March 16—though not the pipeline itself—triggering the temporary suspension of oil-loading operations.

Positive momentum in Venezuela

Even if Venezuelan supplies won’t help solve the global energy crisis, the country’s oil and gas industry is making notable gains quite quickly, analysts said.

And the growth of oil and gas in South America overall eventually can help the world reduce its reliance on Middle Eastern supplies, Monaldi said.

“In the very long term, it does derisk the oil markets if Venezuela produces much more,” he said, citing other key oil-producing countries. “Venezuela and Brazil and Guyana and Argentina are far away from these geopolitical conflicts.”

Venezuela is still home to the world’s largest proven oil reserves on paper. But the dilapidated industry peaked decades ago with an output of nearly 4 million barrels and needs well more than $100 billion in investments to even approach its past glory. Doing so would take several years to bring to fruition.

“Production is moving up, but it’s moving up gradually. There isn’t a secret pool of oil that Venezuela can tap into and immediately unlock hundreds of thousands of barrels a day,” Ferreira said. “The potential is there, but this is years’ worth of work.”

Momentum is building with Venezuela passing new laws to open the industry to outside investment. Chevron, which was the only U.S. producer that didn’t abandon the country during periods of asset expropriation, has agreed to expand its largest project in Venezuela’s oil-rich Orinoco Belt.

Also, Shell plans to develop gassier regions of Venezuela—both onshore and offshore, which would be closer to Trinidad.

Exxon Mobil plans to send a small team to Venezuela to assess the situation, although CEO Darren Woods drew President Donald Trump’s ire in January when was said Venezuela was currently “uninvestable” until major reforms were enacted.

The ongoing political transition with acting Venezuelan president Delcy Rodriguez is going about as well as it possible could thus far, Ferreira said. Changes should continue and eventually usher in elections.

“Folks that have been to Caracas say it’s open for business,” he said.

This story was originally featured on Fortune.com

When Apple CEO Tim Cook and his executive team received their performance targets for fiscal 2025, the board set a modest bar for bonus payouts. The new targets, including sales and operating profit, did not require Apple’s leadership to expand the business—the board set goals at the same level or below the prior year’s results, citing “trade policy” and an “uncertain macroeconomic outlook.”

At the end of the fiscal year, Cook and his team delivered lights-out, extraordinary results anyway, not only blowing past the lackluster bar set by the board, but handily surpassing the prior year’s results, with net sales increasing 6% and operating income increasing 8%.

Cook collected the maximum bonus payout of $12 million—just as he would have, had the company not performed as well, thanks to the safety net offered by Apple’s board.

Apple’s board is hardly unique. An exclusive analysis of pay data from 50 public companies by Compensation Advisory Partners (CAP), published Friday, reveals how corporate boards across America use a range of techniques—more-conservative targets, widened performance curves, and flattened payout ranges—to protect CEO compensation from uncertainties like the chaos of President Trump’s Liberation Day tariffs in 2025. According to CAP’s findings, total pay for CEOs in 2025 rose 8% year-over-year, with annual bonus payouts up 4%. Meanwhile, median financial performance was generally flat to up, with median revenue growing 2.9% and earnings per share down slightly at negative 1.6%, the analysis found. Even among companies with the weakest payouts due to underperformance, CEOs still collected 87% of their target bonuses, up from 77% the year before. The share of companies that landed in the lowest bonus payout tier was down, from 15% in 2024 to 9% in 2025.

Now, with the Iran conflict erupting weeks after most companies finalized their 2026 incentive goals—and global stock markets down roughly $3.5 trillion—some market observers expect that boards will soon be holding the same conversations again.

“They’re not necessarily making decisions today, but they’re just having the conversation about the approaches they might consider at year end, and let’s see how the year plays out,” said Joanna Czyzewski, a co-author of the study and principal at CAP.

To be sure, some of the change among the weakest-performing companies in the CAP report is because of improving results. “Some of it is definitely business improvement,” noted Lauren Peek, a partner at CAP and co-author of the study. But, she said, there are a lot of ways companies can soften the blow of uncertainty and curveballs like tariffs. 

“You might have growth in your targets, you might have widened the curve and the wings,” Peek said. “It’s—for lack of better words—easier to get into the money, because at the end of the day, these executives are trying to do the right thing.”

The Escape Hatch

Among the early-filer companies in the CAP study, their fiscal years end between August and October 2025. That means Trump hadn’t even won the election when they were budgeting and planning for the 2025 fiscal year. Company proxy statements, which include compensation details, provide examples as to how some companies dealt with impending tariffs which later came to fruition on April 2, 2025, before the Supreme Court struck them down last month. (Trump has since imposed a global 15% tariff.)

At personal computer and printing giant HP, the company didn’t wait for tariffs to hit before crafting a plan. In January 2025, at the same time the board locked in HP’s fiscal 2025 performance goals, the HR and compensation committee approved an explicit tariff carveout. Then, when it came time to calculate bonuses for CEO Enrique Lores and his executive team at year-end, the committee stripped out the “net impact of tariff-related costs” from both annual and long-term incentive calculations, the company disclosed in its proxy statement. HP said the adjustments “reflect the net impact of the tariffs after management’s actions, including significant and swift movement of the company’s manufacturing and supply chain along with additional cost reductions and price increases.” That included shifting more than 30% of HP’s manufacturing from China to Southeast Asia and Mexico. 

Ultimately Lores and the executive team earned 67.3% of their target bonus on average. HP described the tariff hit as having “unexpected magnitude” on its financial results and annual and long-term incentive plan calculations after the relevant goals had been set, suggesting that the bonus payouts would have been lower had the tariff impact not been excluded. The compensation committee also used discretion to bring down the executive payouts in line with the broader employee pool, a small acknowledgment of the optics of shielding C-suite executives while the broader workforce takes it on the chin. Lores collected $1.9 million and then stepped down from HP in February and joined PayPal as its new CEO this month. 

HP did not respond to a request for comment. 

Peek, who spoke generally and not about any specific companies, acknowledged the reputational pressure some of these decisions can carry.

“If the company is making these adjustments and giving executives a big payout at the same time as significant layoffs, I’m not sure shareholders would formally comment on that, but the overall optics would be seen in the press,” said Peek. 

Hedging the Goals

Other companies handled uncertainty earlier, back at the goal-setting stage before a carveout would be on the table. Unlike HP, Apple didn’t strip out any costs after the fact, the board made goals conservative at the start.

For the past three fiscal years, Apple’s compensation committee has set at least one bonus target at or below the prior year’s actual results. Bonuses pay out at Apple based on hitting threshold, target, and maximum performance. Hitting threshold earns 50% of the target payout for that measure, hitting the target gets them 100%, and clearing the top rung, the maximum, means executives can double their bonus opportunity. 

For fiscal 2025, Apple’s people and compensation committee appear to have faced a conundrum. Apple’s fiscal year begins in late September and the committee sets goals prior to the start of the fiscal year. President Trump was campaigning against Vice President Kamala Harris and his agenda included massive new tariffs on imports that would hit Apple’s China-based manufacturing significantly. Trump’s tariff rollout wouldn’t happen for roughly another six or seven months when the committee was setting goals. In hashing them out, the committee considered financial results from prior years “as a reference point” but chose goals for 2025 that would “reflect strong financial results commensurate with the projected business and economic conditions for the current fiscal year,” the report in Apple’s 2026 proxy statement reads. 

The fiscal 2024 results were key for determining the goals for fiscal 2025. 

In fiscal 2024—which saw Apple launch its iPhone 16 and $3,499 Vision Pro virtual reality headset—Apple delivered record net sales of $391 billion and operating income results of $123.2 billion. The results were 2% and 8% year-over-year increases, respectively. 

For fiscal 2025, the board set the net sales target at $391 billion—the exact same as the prior year’s actual result. The operating income target was set at $118.5 billion, some $4.7 billion lower than the prior year’s actual result. In doing so, the proxy points to “trade policies and impacts and foreign currency” fluctuations as the rationale for the goal-setting but the committee doesn’t specify a certain anticipated hit to profit margins.

Ultimately, Cook and Apple delivered extraordinary results so strong the structural safety net constructed for the year may have been moot. Net sales in fiscal 2025 swelled to $416.2 billion and operating income was $133.1 billion—bashing past the maximum performance thresholds by more than $14 billion in net sales and $9 billion in operating income.  Cook and the other named executive officers took home maximum payouts on their annual bonuses, which for Cook equated to $12 million, according to Apple’s disclosures. 

The same broad pattern also appeared in fiscal years 2023 and 2024 when Apple set at least one target at or below the prior year’s actual results. During fiscal years 2021 and 2022, the targets were both set above the prior year’s actual results. 

Apple did not respond to requests for comment. 

In the proxy, the compensation committee explained the reasons behind the approach, and wrote that it made the decisions after considering business scenarios “and once again focusing on the underlying business performance rather than the absolute growth rates.”

Generally, companies don’t lower the bar in goal setting arbitrarily, they align targets with their financial budgets to measure performance against what they reasonably anticipate, Czyzewski noted. In setting comp-related goals, companies take into consideration their strategic growth for the year, including growth expectations, and headwinds and tailwinds in the industry and broader economy when they set budgets, added Peek. Target incentive goals are usually set at budget and should “be achievable but stretch,” she said.

“If the goals are too easy, then executive pay may not align with the shareholder experience,” Peek said. “If the goals are too difficult or aspirational, then the award may be demotivating if it is believed that all or most of the award cannot be achieved.”

That’s why board committees have conversations at the beginning of the year about financial metric definitions and possible adjustments, Czyzewski said.

“The intent when goals are set is to ‘get it right’ and account for what is both in and out of the executive team’s control—and set realistic goals,” she said. “But when the unexpected happens, it is good to have a plan and parameters for evaluating results and ensuring pay aligns with performance.”

That way, the conversations at the end of the year will be less about pure board discretion and more about evaluating outcomes within the performance framework.

“If companies are setting targets with their budget,” Czyzewski said, “you’re still aligning them to what finance actually thinks is achievable.”

Taking the Hit

In contrast, not every company reached for tools in the kit, although few businesses have the size and risk profile of Apple.  

TransDigm, which produces pumps, valves, and other parts for aircraft, explicitly told investors that it had the authority to ratchet up payouts by 20%, but didn’t do it in fiscal 2025. The company beat its target goals but did not clear the maximum. CEO Kevin Stein collected $2.6 million. 

Similarly, carbon black manufacturer Cabot told investors it “retains the discretion” to adjust payments, but declined to do it in fiscal 2025. CEO Sean Keohane collected $1.4 million after hitting 90% of target against performance and 130% of target for his individual performance, resulting in a 102% payout. 

Executives at lawn and grounds equipment manufacturer Toro landed between threshold and target for its goals, resulting in a payout of 81.6% for CEO, which translated to $1.3 million for chairman and CEO Richard Olson. 

The Iran Choice

Most companies with a calendar fiscal year approved their 2026 incentive goals in February or the first week of March. Iran erupted days later.

“So even the latest of those probably had those goals approved about two or three days before the news broke about the Iran situation,” said Czyzewski, referring to the fact that calendar year companies did not have an opportunity incorporate the conflict into their goal setting process. “It’s impacting everyone this year, but no one knows how much.”

Whether boards respond the same way they did to tariffs depends heavily on how the conflict unfolds, said Czyzewski. Companies are likely to look for precedent, added Peek— including how boards responded to the Iraq invasion in 2003.

“It would be looking into a crystal ball that we just do not know, because we don’t know how long this conflict is going to last,” she said. 

And, of course, tariffs are still on the table too. 

“If you did not make a carve-out last year, you’re probably not going to make one this year,” said Peek. “But if you did—and you still feel like tariffs are going to significantly impact you—you might still consider using that lever.”

The Compensation Advisory Partners analysis published on Friday covered 50 public companies with revenues ranging from $1.1 billion to $416 billion.

This story was originally featured on Fortune.com

Rob Arnott warns that shareholders in U.S. big-caps will make one-fifth the returns over the next 10 years they pocketed since 2016, and those meager gains will barely edge the consumer price index. You may want to take a cold shower, or a shot of tequila, before you hear the convincing logic behind his dour prediction.

Arnott is the founder and chairman of Research Affiliates, a firm that oversees strategies for nearly $200 billion index funds and ETFs for the likes of Charles Schwab and Invesco. He served as editor-in-chief of the Financial Analysts Journal in the early 2000s, and today comanages the Pimco All Asset and All Asset All Authority funds. He’s also the father of “fundamental indexing,” the practice of weighting stocks by their size in the economy rather than chasing expensive “winners” by ranking according to market cap. At RA, Arnott has bred a think tank in its own right featuring sundry PhDs who apply advanced statistical research to forging benchmark-beating vehicles.

So I check frequently with Arnott to get his take on what those buying into the S&P 500, or baskets of big-cap U.S. stocks, are likely to reap in the years ahead. It’s an especially good time to get a sober reading. The S&P has dropped 4.4% from its record close in January, and the Iran war and jump in oil prices and Treasury yields following the attack are raising a new cloud of pessimism.

An advantage to consulting the sage: Though his predictions are based on a sophisticated analysis of past trends, the future math is basic. In our conversation over Zoom, Arnott stressed that returns have three sources: dividends, growth in earnings (that lift payouts in tandem), and expansion in valuations or P/Es. The last 10 years, he avows, were something of a seldom seen golden age for this trio, but especially profits and multiples. “Overall, U.S. large-caps [as reflected in the S&P 500] produced overall gains of 15.5% a year, an extraordinary number,” says Arnott.

The rub: The fantastic profit and P/E performance over the past 10 years virtually guarantees a rough road ahead

Arnott emphasizes the gap between the historic trends in both profits and valuations, and the S&P’s extraordinary outperformance from mid-March of 2016 through today. Earnings per share waxed at over 11% annually, almost twice their long-term average. The S&P multiple ramped by around one-fifth from the low-20s to roughly 27.5, the current number according to FactSet. “In effect, the big returns were front-loaded by that highly unusual scenario,” says Arnott.

But the high times also foreshadowed today’s downside. Starting at these heights in both metrics, he adds, “has the effect of reducing future returns.” The Wall Street market strategists’ view that anything resembling the last decade’s results are repeatable amounts to a fantasy, declares Arnott. “P/Es don’t always go up without limit,” he says. “In no sensible world is that plausible.” Arnott contends that it’s equally illogical to argue that EPS can keep advancing five points or so faster than their long-term average. As everyone from Warren Buffett to Milton Friedman has pointed out, profits can’t outgrow the economy forever, and after they absorb an unusually large portion of national income, shrink back toward the norm going forward.

Here’s the picture Arnott foresees over the next 10 years. Because stocks are so pricey, the dividend yield now sits at a mere 1.2%, way below its contribution in most periods. (The stats are available on RA’s website under “Asset Allocation Interactive.”) As for profits and P/Es, he cites one of the laws governing markets: reversion to the mean. In the RA scenario, earnings will wax at 5.3%, more or less matching their traditional trajectory, less than half the 2016 to 2026 pace. Add those two components, and you get a “plus” of 6.4% a year. That already sounds mediocre. But the big hit’s a shrinkage in multiples that severely reverses the potent upward push that helped generate those 15.5% returns since 2016. Arnott predicts that valuations will shrink by 3.4 points a year, or 40% by 2036. That pressure would reduce today’s P/E of 27.5 to around 17. Although that sounds extremely slender versus what we’ve seen in recent years, it’s more or less the multiple in the boom years preceding the Global Financial Crisis, and close to the 120-year mean.

All told, the overall S&P 500 should then deliver total annual returns of 3.1% (6.5% from dividends and growth, minus 3.4% from a decline in the P/E). That’s one-fifth the mark for the past decade, and exactly one point better than projected inflation of 2.4%. By 2036, the S&P would stand at 8073, just 21% above its reading of 6672 at the close on March 12.

To gauge just how hugely this outlook diverges from the conventional wisdom, consider that the Wall Street consensus calls for the S&P to end this year at between 7600 and 7650, or less than 6% short of where RA expects the index to finish 10 years hence.

Arnott tags the Magnificent Seven and other high-fliers for pulling the big returns forward, and advises to shun them

Arnott also highlights a significant difference in prospects between the S&P value and growth contingents. The RA model predicts 4% annual gains in the former and a shockingly puny 1.4% in the latter, meaning the recent champs’ returns will lag inflation by one percentage point. Much of the drag, he says, arises from the big valuations, on top of earnings so gigantic they’ll be hard to grow big from here. A major reason we saw that double-digit EPS boom rampage, he avows, “is the stupendous growth in the Mag Seven.” Now, he adds, “Valuations for growth stocks are very stretched, driven by the Mag Seven. The market’s saying it’s a foregone conclusion they’ll grow earnings like crazy. But to beat the market, they’d need to grow earnings even faster than those lofty expectations.”

Arnott’s especially skeptical of the premium prices awarded by investors expecting fantastic profits from AI. “The companies making money from AI are the ones selling the tools,” he says. “They’re now lending to their own customers so that those customers can keep buying their stuff. And their customers are having a hard time monetizing that equipment.” Arnott related that he’d just used Perplexity to perform an in-depth study of how various tax increases being proposed would affect marginal rates at different income levels, and paid nothing for the service. “These AI providers will figure out how to make money,” he says. “But not as fast as the expectations that are built into their stock prices. It will be a slow build over a long period, meaning returns on these stocks will be much lower than the market’s baked in.”

Here’s his advice: “If you’ve owned the Mag Seven, say ‘Thank you very much, Mag Seven,’ and get out and don’t ride them back down.” Arnott believes that returns will be much bigger outside the U.S. than stateside. For example, RA posits that developed nation, non-U.S. value stocks will provide 7.4% returns going forward, more than twice the expectation from the S&P 500, and that emerging-markets value shares will do even better at 7.6%. Arnott concludes that the best strategy is to “first, own no U.S. shares or at least lighten up, and second, own no growth stocks anywhere.”

Versus what we’re hearing from Wall Street, and the S&P’s spectacular showing over the past decade, Arnott’s perception is highly contrarian. But the math’s on his side. And when the math contradicts belief and momentum, go with the math.

This story was originally featured on Fortune.com

The richest person in the world and the most well-known person leading the cause against creating more people like him have many differing views on taxing the ultrawealthy.

But now, Elon Musk and Sen. Bernie Sanders, two men on opposite ends of the ideological spectrum, are using the same math to make opposite arguments for how much billionaires should be taxed, and how that money should be allocated.

In Musk’s view, collecting every cent billionaires rake in pales in comparison to federal debt, which is now hurling towards $39 trillion and counting.

“Even if you tax every billionaire in America at 100%, it barely makes a dent in the national debt,” Musk wrote on X in 2023. “In the end, the government will be forced to tax everyone to pay the debt.”

Sanders agrees—but he’s not looking to tax billionaires for all their worth and he’s not trying to eliminate the debt. Instead, he wants enough to give nearly three-quarters of the nation a nice check, offset cuts to federal health programs, and fund social services. 

938 people stand in the way of you receiving $3,000 checks

Sanders, along with Rep. Ro Khanna, introduced a billionaire tax earlier this month and suggested there are only 938 billionaires in the country, who, combined, hold a net worth of $8.2 trillion. 

Simple math would prove Musk’s logic correct: $8.2 trillion will barely plug a fifth of the national debt. 

But that’s not what Sanders and Khanna are suggesting: the two put forward the “Make Billionaires Pay Their Fair Share Act,” which proposed an annual 5% wealth tax on individuals with a net worth of $1 billion or more.

Sanders estimates the bill would generate $4.4 trillion over its first decade. And in the first year, that revenue would fund a one-time $3,000 check for every American in a lower- or middle-income household, defined as those earning $150,000 or less annually, or roughly 74% of the nation.

In the years that follow, Sanders believes the revenue from the tax would reverse the $1.1 trillion in Medicaid and Affordable Care Act cuts, establish a $60,000 minimum salary for public school teachers, and cap childcare payments at 7% of household income for working parents.

“At a time of unprecedented income and wealth inequality,” Sanders said in the press release, “this legislation demands that the billionaire class in America finally pay their fair share of taxes so that we can create an economy that works for all of us, not just the 1%.”

The debt as it stands

The U.S. is paying nearly $1 trillion per year just to service the debt—a figure that has nearly tripled over five years and has surpassed what the government spends on Medicare. The Committee for a Responsible Federal Budget projects interest payments will exceed $1.5 trillion by 2032. America is, at an accelerating pace, borrowing money to pay interest on money it already borrowed.

Musk and Sanders are making two different arguments. Musk’s framing casts a billionaire tax as a debt solution, and by that measure, it fails. Sanders’ framing casts taxing billionaires as a redistribution mechanism, a way to put money back in the pockets of working Americans and fund social services. By that measure, a 5% annual wealth tax generating $4.4 trillion over a decade is significant. 

Musk has warned more broadly that America is on a path to going bankrupt “1000%” if spending isn’t curtailed. The debt crisis is structural, rooted in decades of spending that outpaces revenue, and no single tax can undo that. The national debt has grown by more than $11 trillion over the last five years alone.

But Sanders’ counter is equally pointed: the debt crisis and the affordability crisis are not the same problem, and solving one doesn’t require ignoring the other. A $3,000 check won’t fix the national debt. But for a middle-class family barely keeping up with inflation, it may fix something more immediate.

This story was originally featured on Fortune.com

The AI capital expenditure boom has created a gusher of corporate debt, forcing the Treasury Department to make its bonds more attractive to investors as the U.S. war on Iran adds to the deficit.

Last Tuesday saw the single busiest day on record for U.S. corporate bond sales as President Donald Trump’s hint that the war may end soon briefly calmed markets and sparked a mad dash for companies to issue fresh debt.

By the end of the day, total investment-grade issuance topped $65 billion, exceeding the prior one-day record of $52 billion in 2013. The flood of debt was led by e-commerce giant and AI hyperscaler Amazon, which raised $37 billion, sources told the Financial Times.

That beat the company’s guidance for $25 billion-$30 billion as investor demand far outpaced the available supply, attracting about $123 billion in orders.

The corporate debt surge was enough to move the needle in the Treasury market, where daily trading volume exceeds $1 trillion. Analysts at Deutsche Bank said in a note last week that the bond sales added some upward pressure on the 10-year yield, which climbed 6 basis points to 4.16% at session highs.

Apollo Chief Economist Torsten Slok previously warned the flood of corporate debt could make borrowing more expensive for the federal government.

In a note from January, he pointed out that Wall Street estimates for the volume of investment grade debt that’s on the way in 2026 reach as high as $2.25 trillion.

That’s as the AI boom increasingly sends companies, including hyperscalers and adjacent firms, to the bond market to fund massive investments in data centers and other infrastructure.

“The significant increase in hyperscaler issuance raises questions about who will be the marginal buyer of IG paper,” Slok said. “Will it come from Treasury purchases and hence put upward pressure on the level of rates? Or might it come from mortgage purchases, putting upward pressure on mortgage spreads?”

Much has changed since January. The Iran war is shaping up to be a prolonged conflict that’s sent oil prices spiking. In turn, bond yields are up on exceptions of higher inflation—further adding to borrowing costs.

Bombarding Iran everyday also adds stress to the deficit, which hit $1 trillion in just the first five months of the fiscal year. Pentagon officials told lawmakers last week that the cost for the first six days of the war topped $11.3 billion, according to the New York Times.

Meanwhile, Trump has vowed to boost defense spending to $1.5 trillion a year from $1 trillion, threatening to further blow up the deficit.

The unsustainable trajectory of U.S. debt has raised growing alarms on Wall Street. But for now, investors appear to have a strong appetite for both corporate and government debt.

Days after Amazon’s mega-offering, an auction Thursday for $22 billion in 30-year Treasury bonds drew solid demand, though it was helped by the jump in yields since the war began.

And a Treasury offering last month saw the highest demand ever in the history of 30-year auctions, led by overseas buyers.

“The bottom line is that Treasury auction metrics show that there continues to be very solid demand for the long end in US Treasuries,” Slok said in a note Feb. 20.

This story was originally featured on Fortune.com

A federal judge in Boston has temporarily blocked federal health officials from cutting the number of vaccines recommended for every child, and says U.S. Health Secretary Robert F. Kennedy Jr. likely violated federal procedures in revamping a key vaccine advisory committee.

The decision Monday halts an order by Kennedy — announced in January — to end broad recommendations for all children to be vaccinated against flu, rotavirus, hepatitis A, hepatitis B, some forms of meningitis and RSV.

Leading medical groups voiced alarm at the changes. The American Academy of Pediatrics and some other groups amended a lawsuit filed in July, asking the judge to stop the government from scaling back the nation’s childhood vaccination schedule.

The judge also says Kennedy’s reconstitution of the vaccine advisory panel likely violated federal law. He ordered the appointments — and all decisions made by the reformed committee — put on hold.

___

The Associated Press Health and Science Department receives support from the Howard Hughes Medical Institute’s Department of Science Education and the Robert Wood Johnson Foundation. The AP is solely responsible for all content.

This story was originally featured on Fortune.com

In its early days, the Elon Musk-led Department of Government Efficiency (DOGE) bragged it could cut up to $2 trillion from the U.S. federal budget. In December, Musk conceded the special advisory only saved $200 billion in “zombie payments” for cancelled contracts or fraudulent unemployment claims. 

But a recent estimation of DOGE’s overall impact indicated any savings it found did little, if anything, for the deficit.

In a deposition video from January, which recently went viral, DOGE employee Nate Cavanaugh said cost-cutting efforts fell far short of its original $2 trillion goal. The deposition was part of a larger lawsuit filed by the American Council of Learned Societies, a nonprofit consortium of scholarly institutions, alleging DOGE used OpenAI’s ChatGPT to identify and then cancel more than $100 million in diversity, equity, and inclusion grants.

“You don’t regret that people might have lost important income…to support their lives?” one attorney asked Cavanaugh regarding the grant cancellations.

“No. I think it was more important to reduce the federal deficit from $2 trillion to close to zero,” said Cavanaugh, who is also the founder of AI-powered accounting firm Flow Finance.

“Did you reduce the federal deficit?” the attorney asked.

“No, we didn’t,” Cavanaugh replied.

The White House did not immediately respond to Fortune’s request for comment.

Judge Colleen McMahon of the Federal District Court in Manhattan ordered videos of the deposition, which also included a deposition of DOGE staffer Justin Fox, to be removed online following social media backlash.

DOGE, formed the first day of President Donald Trump’s second term, was part of an effort to root out so-called “waste, fraud, and abuse” from the federal government. Over the course of the 10 months it was operating under centralized leadership, the group eliminated the roles of more than 300,000 federal employees and claimed to have canceled 13,440 contracts. 

The recent lawsuit is the latest instance of DOGE coming under scrutiny. Cybersecurity experts warned the group had access to U.S. payroll systems that presented “unprecedented power and control” over Americans’ information, while the mass layoffs could have created opportunities for countries like China and Russia to recruit informants who had access to classified data. Management experts claimed DOGE’s purported savings were completely overblown.

Signs of DOGE increasing U.S. government spending

From the beginning of Musk’s tenure as head of DOGE, which was not a real department but instead an advisory office, economists were skeptical about its ability to slash the federal deficit as the national debt soared beyond $38 trillion. The Brookings Institution Hamilton Project tool tracking federal spending found that as of Dec. 19, 2025, government spending increased nearly 6% to $7.558 trillion from $7.135 trillion a year earlier. 

A Cato Institute report from December 2025 argued DOGE’s failure to shrink overall spending, despite culling more than 9% of the federal workforce, was in part because most federal spending does not come from salaries. Additionally, the government hired contractors to replace employees. The libertarian think tank calculated that a 10% cut in the workforce would result in a savings of only about $40 billion.

Max Stier, chief executive of government efficiency and workforce nonprofit Partnership for Public Service, told Fortune in April 2025 that DOGE’s cuts could actually mount pressure on America’s coffers, estimating the cost to fire, rehire, and put workers on paid leave cost American taxpayers roughly $135 billion. 

“We do need to have our government work better, but the approaches that have been adopted so far are taking us in the exact wrong direction,” Stier said. “The end result will be that the American public will be holding the bag as Elon Musk goes back to his private enterprises.”

A Yale University Budget Lab report from March 2025 similarly forecasted that if 22,000 Internal Revenue Service employees left their roles, the agency would lose $8.5 billion in revenue in 2026 as a result of fewer personnel available to conduct audits. Over a decade, that loss could snowball to nearly $198 billion in lost revenue. (The U.S. Government Accountability Office reported that more than 17,000 IRS employees left the agency last year.)

An IRS employee previously told Fortune the mass layoffs have decreased the efficiency of employees, increasing call times and slowing down the processing of paperwork.

“When we look back historically, we’re going to see that the gutting of the bureaucracy that keeps the government running, that keeps the country functional, will be the trigger that collapses America,” the employee said.

Scott Kupor, the head of the Office of Personnel Management, indicated the workforce reductions last year were overdone. He told The Washington Post earlier this month the administration is planning to rehire several positions in the 2 million-person federal workforce.

“We probably have some skills that we now need to hire back, quite frankly,” Kupor said. “There’s no question anytime you do restructurings … sometimes you over-restructure, sometimes you under-restructure.”

This story was originally featured on Fortune.com