The White House is now cracking down on these tariff dodgers. In a report on Tuesday, the administration chronicled the magnitude of the problem and outlined how it’s trying to curb it. It claimed the U.S. is losing between $19 billion to $26 billion in tax revenue annually as a result of countries routing exports through other countries in order to evade levies, in a process called transshipment.

But the true extent of the tariff fraud may be even greater than that. Last year, data from China’s General Administration of Customs and U.S. Census Bureau showed a $112 billion gap between what China reported shipping to the States and what the U.S. reported receiving—suggesting efforts to evade these taxes are ballooning even beyond the tens of billions of dollars outlined by the Trump administration.

China appears to be the main culprit behind the tariff dodging, processing exports through more than 40 other countries, according to the report. But it’s not the only one to receive the White House’s attention: The report also pointed to dozens of other nations turning a blind eye to shell importers and foreign importers behind tariff fraud. 

“While the future may be murky, the past is not,” read the report, which came from the White House’s Office of Trade and Manufacturing Policy (OTMP). “The second Trump Administration inherited a Great Transshipment Scam—a witch’s brew of economic incentives, bad actors, and lax enforcement that had been allowed to simmer and grow more toxic over time.”

While previous trade policy has empowered bad actors to find ways to dodge levies, trade experts say there’s still one obvious reason why tariff fraud has increased, and that responsibility rests of the shoulders of the current president: The existence of the import taxes in the first place, exacerbated by Trump’s Liberation Day tariffs last year, have jumpstarted the practice of dodging them.

“The tariffs have created a huge incentive,” Ryan Petersen, CEO of supply chain management platform Flexport, told Fortune. “If your tariff was 0% there’s no need to commit fraud; there’s no tariffs to evade. As those tariffs have gone way up, it’s just created a huge incentive to change your terms of trade, to lie about the valuation or the classification or the country of origin of the goods.”

The rise of tariffs—and tariff dodging

Tariffs have been a cornerstone of Trump’s second administration. The president’s “Liberation Day” tariffs have imposed levies against China of up to 145%. Even after the Supreme Court struck down the lion’s share of tariffs, which were imposed under the International Emergency Economic Powers Act (IEEPA), the White House has tried to replicate high tariff levels through duties imposed under the 1974 Trade Act. 

As of earlier this month, U.S. tariffs on China were around 23%, according to the Penn Wharton Budget Model, more than double the about 11% import tax on the country prior to the start of Trump’s second term.

“Why we’re seeing transshipment as a much bigger issue now is because the tariffs are higher across the board,” Carrie Owens, a partner at law firm Kelley Drye & Warren and former head of the Enforcement Operations Division at U.S. Customs and Border Protection (CBP), told Fortune.

While transhipment has been around for decades, the practice ramped up in 2018, when the president in his first term imposed tariffs on more than $250 billion worth of Chinese goods. The trade war incited a wave of rerouting goods through third-party countries, as well as led to companies under-reporting the value of goods or mislabeling products as alternative goods not subject to as high of import taxes. Goldman Sachs calculated that the U.S. previously lost between $110 billion to $130 billion in revenue from tariff dodgers during Trump’s first term.

Today, Petersen—a vocal critic of the current tariff policy and advocate for interventions against tariff evasion—warned the magnitude of tariff dodging is “massive,” and the Trump administration would agree. The report cited a 2020 study by the Economic Policy Institute, which estimated 3.7 million jobs were displaced between 2001 and 2018 as a result of the U.S.-China trade deficit. Author Robert Scott said the deficit increased by $336.5 billion in that period, and this could eventually cost the U.S. between $60 billion and $606 billion in annual GDP losses, the White House claimed.

More concretely, Owens argued, tariff dodgers are squeezing the companies paying their fair share of tariffs, leaving them to compete with businesses who aren’t burdened by the levies in the same way they are. 

“The good actors that are doing what they’re supposed to and paying their revenue, foreign companies that are importing the United States that are following the rules—they’re being hurt as well,” she said. “So it’s not even just U.S. businesses. Any company that is following the rules is being damaged and harmed by these goods that are coming in.”

Enabling and curbing tariff dodging

If Trump’s raft of tariffs precipitated an onslaught of tariff dodging, preexisting U.S. trade policy laid the groundwork for it. 

The U.S. allows for foreign importers of record, or non-American business entities, to take responsibility for shipment and customs entries. While this allowance was likely made in the name of free trade and limiting regulatory burden, it also gave power to companies looking to avoid tariffs. These entities can effectively act as shell companies to funnel goods between point A and point B, disappearing quickly when regulators grow suspicious of evasion, but also leaving U.S. custom authorities with little to do because they are outside of their regulatory jurisdiction.

“The owners are foreign,” Owens said. “So we don’t have the tools in the United States to get at, to address and penalize those foreign owners when there’s no U.S. assets.” 

The Trump administration has worked to crack down on these foreign importers of record. A June 3 executive order restricted these foreign entities from using continuous customs bonds and required them to use a formal entry procedure requiring more detailed documentation. CBP is also deploying AI to scan shipment data, check routing histories, and flag inconsistencies in documentation.

Owens said results from these crackdowns will be felt swiftly, as early as October. But she warned that until tariff evasion is curbed, there’s the risk of a vicious cycle of the Trump administration hiking import taxes to try to make up for revenue lost from dodgers.

“If everybody paid the tariffs they’re supposed to, I personally don’t think the tariffs will be as high as they currently are,” Owens said. “Part of having those high numbers is because there isn’t the enforcement that there’s the tariff evasion that’s happening.”

This story was originally featured on Fortune.com

And the concern is warranted: Meta’s Ray-Ban Meta Glasses, for example, have been coined “pervert glasses” for recording people undressing, and the company is even being sued after a study revealed Meta’s subcontractors are viewing your most intimate moments. This is all culminating into a real privacy concern as not only is your right to privacy (or lack thereof) in the public realm coming into question, but so is how to stop what you do in public from getting stored on a company’s servers. 

“We are living in weird times,” Jim Waldo, a professor of computer science who teaches several technological privacy courses at Harvard, told Fortune. “The technology is changing. It’s the combination of the Meta Glasses with facial recognition, AI, and a number of other sorts of technologies that are all coming together and putting us in an environment that we just aren’t prepared to deal with yet.”

The privacy concerns are real—from the suit to data collection, and even the social media content made with the tech has left people chalking up Meta Glasses as a form of surveillance. Even Meta’s Instagram has had to act: the platform disabled several accounts thanks to violations of content usage after those accounts amassed millions of followers by streaming live feeds from Meta glasses.

“We don’t want harassing content on our platforms and take it down when we find it,” a spokesperson for Instagram told Fortune.

A new era of wearable technology

Gone are the days of “dumb” smart wear. No longer are wearables contained to just your fitness trackers or your sleep monitors, they now have cameras with AI built in them. Most prominently in this space are Meta’s Ray-Ban Meta Glasses, which have cameras built directly into the frames, allowing users to take photos and videos without pulling out a phone. The glasses also have microphones that capture audio, and Meta has enabled livestreaming directly from the glasses to Facebook and Instagram.

There’s a way to tell if you’re being recorded: the glasses use a white capture LED on the front of the frames that blinks when content is being captured. Meta says the LED cannot be switched off and that the camera is disabled if the LED is covered or blocked.

“We will keep strengthening our protections as our glasses become even more capable,” Meta spokesperson Dina El-Kassaby told Fortune.

But experts are still concerned about the privacy implications of wearable technology. “They’re making it safe for the consumer,” Waldo said. “They’re not making it safe for the people around the consumer.”

There are also legal implications for the use of these smart glasses in public. Gene Kang, partner at law firm Rivkin Radler LLP, told Fortune the technology itself is not necessarily the problem, but that people don’t know they could be filmed.

“If you’re holding up your phone to somebody’s face, they’re going to know,” Kang said. In that situation, he explained, there could potentially be an argument for implied consent if the person knows they’re being recorded and does nothing to object. With the inconspicuous glasses, however, that assumption becomes much harder to make—meaning privacy and consent laws can potentially be invoked.

“If they’re not aware that they’re being recorded, then I think that presents a different issue,” he added. “I think they would potentially have a claim there.”

Discreet recordings, “Pick-up artistry” and data sensitivity

According to a study done by University of Sydney researchers, “pick-up artistry” content has picked up in recent years. This type of content, spread around social media and mainly perpetrated by individuals in the “pick-up artist” community, attracts viewers who wish to watch point-of-view reels of women being approached in public.

The study found 60% of over 350 videos analyzed involved behavior classified as potentially harassing. In 43% of the videos, women were subjected to derogatory commentary, and other subjects were identified or doxxed. The study focuses on what the researchers defined as “ambient capture”—recording people in their everyday surroundings without them realizing that a camera was pointed at them.

The researchers found a relationship between the apparent covertness of the recording devices and the severity of the harassment. “We should all be very concerned,” Dr. Milica Stilinovic, one of the study’s authors, told Fortune.

Fighting back

The harmful content perpetrated online echoed concerns among consumers, leading them to find avenues to protect themselves. People have started to use face markings to confuse the facial recognition system within the glasses, and a theory has circulated online to sing copyrighted songs when under suspicion of being recorded.

Some individuals have even developed software to help notify users for potential smart glass intrusion. Professor. Dr. Yves Jeanrenaud built an open-source, free software app Nearby Glasses, allowing users—as the name suggests—to be notified when Meta Glasses are nearby. According to its open-source repository, Jeanrenaud developed the app in response to “an intolerable intrusion, consent neglecting, horrible piece of tech that is already used for making various and tons of equally truely disgusting ‘content’.”

According to the Google Play store, the app has amassed over 100,000 downloads to date.

And while the app was made to help users stay aware of potential discreet filming, Professor Jeanrenaud included a disclaimer on the use of his technology.

“It’s still an imperfect approach and probably always will be,” he wrote. “It’s not all good only because this app exists now. We need better solutions to curb surveillance tech and privacy intrusion.”

Not all of the methods are feasible, however. A recent social media theory has circulated citing Disney songs can protect you from being filmed. The idea is that Disney’s notorious copyright strikes would be enough to get any unsolicited videos taken down across social media. But according to Kang, hiding behind copyright isn’t an effective way to protect yourself from discreet filming. 

“If you’re the person being recorded, you don’t own any copyright to the composition,” he explained. However, while he did add that copyright may not be an effective claim, he also said individuals who want to protect themselves should look into privacy claims instead.

“It’s really a privacy issue,” he said. “Which still could be applicable here.”

This story was originally featured on Fortune.com

The Boston-based firm just landed an $80 million contract to modernize and AI-enable WarMatrix, the Department of War’s wargaming simulation environment, Fortune learned exclusively. The process only took less than 12 months.

“I have never seen them move this fast, and that’s been really exciting,” CEO Peter Morales told Fortune, adding that after securing $17 million for the initial operating capability—the minimum version of the system that can be usefully deployed—they received the full funding as an Other Transaction Authority (OTA) agreement.

Code Metal—which builds AI-powered software to translate code across programming languages and verify that it works on different types of hardware—was valued at $1.25 billion in February, when it secured its $125 million Series B, led by Salesforce Ventures.

The deal comes as venture-backed companies race to turn investors’ enthusiasm for defense technology into actual government business. It also reflects a broader push to bring the Pentagon’s vast, decades-old software infrastructure into the AI era. 

WarMatrix is designed to help military planners run sophisticated wargaming and analysis, and Code Metal’s work is expected to allow analyses that once took months to be completed in days.

“We’re not replacing decades of validated code,” Morales said. “We’re really making it reachable to the warfighter and to AI.” 

Morales spent roughly a decade working at the intersection of hardware and software, including at British defense contractor BAE Systems, MIT’s Lincoln Laboratory, and Microsoft. At BAE, he worked on software and real-time systems for the controversial but vital F-35 fighter jet, where he said he learned how quickly military technology can fall behind the commercial world.

That has created an opportunity for startups like Code Metal to help the Pentagon preserve decades of validated technology by building a software layer that makes existing infrastructure compatible with newer tools—including AI.

“The reason this industry has been so ripe for disruption is we’re willing to show up with something working rather than a slide deck and gamble that we believe that this is important,” Morales, 38, said.

About 75% of Code Metal’s business is currently defense-related. Morales necessarily doesn’t see the defense primes—the long-established defense contractors—as competition, but as rather partners and even clients. Code Metal has already worked with the U.S. Air Force as well as Raytheon, L3Harris, and Boeing.

“The defense primes do things that only they could do,” said Morales. “They’re not going anywhere. But in some of these forward pushing areas like AI, they understand they need help, and that’s why we’re working with them so closely to help them go faster as well.”

Roughly 10,000 new defense companies have entered the market over the past two years, according to an analysis by the Center for Strategic and International Studies. And in the first quarter of 2026, VCs deployed a record $19.8 billion into defense tech across 262 deals.

Among the 15 highest-valued defense-tech startups—including Anduril and Saronic—Pentagon contract spending tripled last fiscal year from 2022, The Wall Street Journal reported. However, those companies still only account for less than 1% of total dollars for all defense contractors.

For startups like Code Metal—which employs 123 people—the challenge now is turning one-off opportunities, like the new $80 million Pentagon deal, into repeatable, expandable business. 

That’s part of the reason the company brought in Ryan Aytay earlier this year to serve as president and COO. Aytay, the former chief business officer at Salesforce and CEO of Tableau, brings the kind of sales and scaling experience Code Metal will need as it looks to turn its early defense wins into a much larger business. 

Morales described Aytay as a Sheryl Sandberg-type leader who can help turn Code Metal into an “n of 1”—a company that, like Tesla or Palantir, creates its own category.

ICYMI… The deluge of executive departures at OpenAI continues, as the company announced Denise Dresser would be leaving yesterday. Read more here.

See you Monday,

Preston Fore
X: @forepreston
Email: preston.fore@fortune.com

Submit a deal for the Term Sheet newsletter here.

Joey Abrams curated the deals section of today’s newsletter. Subscribe here.

This story was originally featured on Fortune.com

Mangione, 28, entered the plea at a hastily scheduled hearing in Manhattan federal court. “On the morning of Dec. 4, 2024, I shot Mr. Thompson in Manhattan, and he died,” Mangione told the court.

He admitted to emailing UnitedHealthcare ahead of time and posing as a potential investor, using the ruse to confirm the location of the company’s investor conference so he could track Thompson’s whereabouts. Mangione was never a customer of UnitedHealthcare, according to both police and the company.

U.S. Attorney Jamie McDonald told reporters at a post-hearing press conference that it was an open plea and there was no deal with the defense. Now, there’s a question of whether the state can continue its murder trial, scheduled to begin next month. Today’s federal plea does not automatically wipe out the state’s murder trial.

“Mr. Mangione has accepted full responsibility for the death of Brian Thompson,” his lawyer Karen Friedman Agnifilo told reporters outside court, arguing that her client shouldn’t face two prosecutions over a “single tragic event.”

Mangione pleaded guilty to all pending counts in the indictment with no agreement in place, meaning prosecutors face no restrictions on the sentence they can pursue. McDonald said Mangione “pled open,” receiving no concessions, and that the government is free to seek the maximum penalty under the law.

Before accepting the plea, U.S. District Judge Margaret M. Garnett asked Mangione a series of questions and found him to be “fully competent” after he said he felt his mind was clear. Mangione is set to be sentenced Dec. 18, and faces a maximum of life in prison.

The state case now hangs in the balance

Within hours of the plea, Mangione’s defense filed a motion to dismiss the separate New York murder and weapons charges against him. His attorney argued the dual prosecutions arising from the same event violated Mangione’s rights, and the motion alleges the Manhattan District Attorney’s Office and the Justice Department coordinated to schedule the state case ahead of the federal one specifically to sidestep double jeopardy protections. The motion also claims both sets of prosecutors overcharged him.

The Manhattan DA’s office said the office is prepared to litigate the defense motions and has been preparing since December 2024 to pursue a conviction in state supreme court.

Mangione was arrested five days after the Dec. 4, 2024 shooting at a McDonald’s in Altoona, Pennsylvania, about 230 miles west of Manhattan. Prosecutors say ammunition at the scene was inscribed with “delay,” “deny” and “depose,” echoing language used to describe how insurers deny claims. The words helped turn the case into a flashpoint over the health insurance industry, drawing a cult following to his hearings.

A federal judge dismissed additional charges against Mangione earlier this year, including one that could have exposed him to the death penalty. His lawyers have also fought in pretrial hearings to suppress evidence seized from his backpack, including a notebook and to-do list prosecutors say detailed his intent to kill an insurance executive, arguing in an earlier filing that he wasn’t read his rights and that the backpack search was warrantless. In June, his lawyers briefly said they would pursue a psychiatric defense in the state case, centered on claims he was suffering from extreme emotional disturbance at the time of the killing, all before reversing course a day later. That defense isn’t available in federal court.

This story was originally featured on Fortune.com

In the past week, the administration has pushed back against the narrative that Iran has virtually shut down the critical chokepoint with the threat of missiles and drones.

That’s as traffic data since the U.S.-Iran ceasefire collapsed has shown just a trickle of ships are openly transiting the strait, suggesting another supply shock ahead for global energy markets. But more tankers are sailing “dark,” meaning their transponders have been turned off and are no longer broadcasting their location.

Energy Secretary Chris Wright said on Tuesday that the seven-day average for oil leaving the strait was almost 9 million barrels per day and credited the U.S. military as well as Gulf allies.

When combined with another 5 million-7 million barrels per day shipped via newly upgraded pipelines and export facilities, total oil flows average about 15 million barrels per day, he added in a post on X. That compares with 20 million barrels that were exported daily before the war.

Similarly, a U.S. official told Axios on Sunday that about 8 million barrels are quietly exiting the Gulf each night through a southern lane in the Strait of Hormuz with help from the U.S. military.

Before the ceasefire agreement fell apart, the U.S. military guided tankers through an alternate route that hugs the Omani coast and provided some protection. Enough ships made it out of the Gulf, easing pressure on global oil markets. But that prompted Iran to attack vessels trying to bypass its own corridor, reigniting hostilities and leading to the current standoff.

Some experts doubt the recent Gulf shipment numbers are as high as the Trump administration claims—but not by that much. Oil market researcher Rory Johnston estimated that average volumes out of Hormuz peaked at 7 million barrels per day over the past week and acknowledged that could be higher because of the uncertainty around dark transits. Meanwhile, pipelines are exporting about 4 million barrels per day.

In addition to dark transits, another tactic for sneaking oil supplies through the strait is ship-to-ship transfers, a practice Iran and Russia have previously used to get their oil tankers past Western sanctions.

Bessent’s warns of ‘economic isolation’

To slip under Iran’s nose, tankers exit the Gulf, transfer their oil to another ship off the coast of Oman, then shuttle back through the strait to do it all over again. Not all ships go undetected, which explains why Iran is still attacking ships even as it claims the strait is completely shut down.

But the U.S. naval blockade is preventing Iran from exporting its oil via the Strait of Hormuz, depriving the regime of a vital revenue lifeline. At the same time, other Gulf oil producers like Iraq, which depends heavily on the strait, are getting their barrels out by way of dark transits and ship-to-ship transfers

“Hefty chunk of non-Iranian crude still getting out, unlike the Iranian crude that isn’t,” Johnston posted on X.

Global oil markets still face a supply deficit, forcing consuming countries to tap reserves that are already low and heading toward critical levels soon.

But the oil that’s coming out of the Gulf provides additional wiggle room. In fact, crude prices have declined since spiking last month when the ceasefire ended and fighting flared up again.

An oil market reprieve also gives Trump more time to squeeze Iran’s economy with his naval blockade, which some officials in Tehran have admitted is causing an economic collapse. And even more pressure could be on the way.

“It will be a combination of economic isolation like ‌the world has ​never seen before, ​and ​the continued blockade in ‌the Strait of ​Hormuz that will ​keep anything from going in or out of ​the ‌Iranian ports,” Treasury Secretary Scott Bessent told Newsmax without elaborating.

This story was originally featured on Fortune.com

• Juicy details OpenAI doesn’t want you to see in its new report
• Anthropic reportedly plans a $2 trillion IPO in October—the largest ever
• OpenAI replaces its chief revenue officer after less than a year
• Google pronounces Sam Altman dead—for 41 minutes

It really sunk in for me this week just how much money is flowing in the AI industry.

I spoke with two former OpenAI employees who made about $10 million in a day by selling shares in an internal tender offer, which Bloomberg reports totaled $7 billion across the staff. Then, this morning, on Fortune‘s weekly AI podcast, my coworker Beatrice Nolan and I interviewed the CEO of Lovable. This week, the old Stockholm-based firm, which is only three years old, doubled its valuation to $13.3 billion.

A couple million, a hundred billion, a trillion (or two, in the case of Anthropic’s upcoming IPO)—what’s the difference at this point? There’s just one big problem looming in the background: The ROI of AI adoption is still not clear for companies.

OpenAI grapples with this existential question in a 69-page report published on August 11 on the enterprise adoption of ChatGPT. On its face, the report tells the a story of exponential AI usage growth across all seniority levels and job functions, highlighting what it calls a “frontier gap,” in which companies who are using AI are pulling ahead of those who aren’t. In other words, if you’re not using AI—especially agents you can delegate tasks to—you’re losing.

But the fine print tells a different story.

We still don’t know if AI helps you make more money

In one small table on page 35, the researchers report no statistically significant correlation between the revenue per employee, and how much those employees use AI, measured in messages sent and tokens used.

“Revenue per employee is not meaningfully associated with output tokens per employee or messages per active user once other controls are included,” the report explains.

Importantly, the revenue numbers here are from before the employees began using ChatGPT. It’s unclear if OpenAI is tracking revenue and AI usage, why they would not extend the study to include how ChatGPT has started to affect their cash flow since adoption. This massive question for the business community is left hanging.

Overall, the most lucrative companies have been most likely to be AI early adopters. However, the study doesn’t clearly establish that the more AI they use, the more money they make. Perhaps they just have the most to spend on it.

Executives are using AI the least

Executives may not be best equipped to gauge ROI because they are using it the least—another nugget buried in the report. It’s not just that companies have fewer executives than they do general employees. But what’s interesting about the graph on page 29 is that most senior employees are using it less intensely, with the least weekly messages per user.

Early career employees have by far the most usage, a point OpenAI CFO Sarah Friar highlighted in her LinkedIn post about the report: “For leaders, that’s a reminder that competitive advantage comes from the people closest to the work. Listen to them, learn from them, and help the rest of the organization catch up.”

OpenAI’s enterprise sales had a rough Q4’2025

Surprisingly, OpenAI’s overall usage within enterprises completely flatlined from about October 2025 to December 2025. In a graph (page 26) depicting output token growth, the black line representing “total” growth is almost perfectly flat for that time period. During this time, Anthropic’s Claude Code was taking the corporate world by a storm, becoming the go-to platform at many places.

To the company’s credit, in January 2026, the line thrusts upward into an exponential curve. As one VC told me yesterday, “OpenAI’s run rate in 2026 has been pretty incredible.” OpenAI attributes the growth not only to adding new clients, but also also to its current clients deepening their use. We also know CEO Sam Altman has been reorganizing the company around enterprise sales, and killing what the company called “side quests,” such as the video app Sora.

In a sprint to accelerate this line—Or, maybe to get it going again? Who knows, the graph ends at March 2026—OpenAI today announced it hired a new Chief Revenue Officer, Dali Rajic, who will replace Denise Dresser. It’s an aggressive move; Dresser held the role for less than one year. Rajic’s focus will be accelerating customer adoption and helping businesses measure impact as the company sprints towards its IPO.

OpenAI paid the academics who contributed to the report

Lastly, two of the five authors are academics that OpenAI paid to help with the report. The other three are OpenAI employees. Including academics in a paper like this typically implies greater credibility and the impartiality of an outside research institution, but the waters are a little muddier here.

On the first page, David Holtz and Prasanna Tambe are affiliated with Columbia Business School and Wharton at the University of Pennsylvania, respectively. But a footnote specifies that both “contributed to this work in their capacity as paid contractors for OpenAI.”

Did the researchers find more that they didn’t publish, as they typically would for an academic paper? We’ll never know, but it’s another reminder of what has always been true: You’ll have to measure AI’s impact based on your own first-hand experience—not the hype.

With that, here’s more AI news.

Emily Forlini
emily.forlini@fortune.com
@emilyforlini

This newsletter has been updated to clarify that the revenue per employee figure referenced in the study was taken before the employees began using ChatGPT, not after.

This story was originally featured on Fortune.com

AT&T had spent years expanding into entertainment, including its $85 billion acquisition of Time Warner, and owned satellite television provider DirecTV. By the time Kenny arrived, that strategy was giving way to a retrenchment that would return AT&T’s focus to its telecommunications business. The company would separate DirecTV and eventually spin off WarnerMedia, while directing more capital toward 5G and fiber and reducing debt.

Kenny saw an iconic company that, in her words, had “lost its way a bit” and needed to find its “swagger again.” That complexity was part of the attraction. Kenny, whose career had taken her through Microsoft, Capital One, Uber, and Hilton, wanted her next job to be at a company where marketing would play a central role in its transformation.

“I’m only going to go to a company where, for that company to truly reach its full potential, marketing has to be a key ingredient,” Kenny recalls of her thinking at the time.

Nearly six years later, the CMO, who also serves as AT&T’s chief growth officer, says the company can quantify how brand strength translates into customer growth.

One measure AT&T tracks is what Kenny calls “brand love,” measured by asking consumers to rate brands on a seven-point scale running from hate to love. AT&T counts respondents who select either of the two highest ratings, a six or seven, as consumers who love the brand.

AT&T’s brand love score has risen 13 points over the past five years, according to Kenny, who says a one- or two-point increase in a year is typically considered strong performance.

The more consequential finding came when AT&T compared those survey responses with subsequent customer behavior.

Prospects who say they love AT&T are 1.6 times more likely to become customers within the following 12 months, according to the company’s analysis. Existing customers who love AT&T are three times less likely to leave and roughly 50% more likely to buy a second service, such as adding AT&T home internet to their wireless plan. In markets where AT&T records higher levels of brand love, converting prospects into customers costs roughly 50% less, Kenny says.

The analysis gives AT&T a way to tie an often nebulous measure of brand sentiment to hard business outcomes, including customer acquisition, retention and spending.

“We wanted to sharpen our pencil and make sure that we understood exactly what the financial impacts are,” Kenny says, adding that understanding consumer sentiment is  “essential for the core economics of the company.”

Turning customer research into products

Kenny’s remit as chief growth officer also puts marketing closer to AT&T’s product and growth strategy. As head of growth, she oversees customer research, insights, and analytics, as well as digital, and her team’s work informs product roadmaps and broader operational strategy across AT&T.

Case in point: AT&T found that roughly 40% of consumers are extremely price-sensitive, Kenny says, either because of financial constraints or because they prefer to pay only for what they use. Some felt they were paying for bundled extras they did not want, such as entertainment services and international calling, even when those features were advertised as free.

That insight helped shape AT&T’s Build a Plan offering, which starts at $15 a month and lets customers add the services they want. The offering targets consumers who viewed AT&T as too expensive because its plans included services they did not value.

The same customer research has also shaped how AT&T handles service failures. The company found that customers wanted greater confidence in the reliability of its network and customer support, helping lead to the AT&T Guarantee. Under the guarantee, eligible fiber customers who experience an outage of at least 20 minutes receive a full day’s service credit. 

AT&T also commits to resolving certain technical issues within 24 hours and compensates customers when service falls short of its standards. The program followed a roughly $1 billion investment in customer service and technology improvements that allow AT&T to detect outages, notify customers, and issue credits proactively, says Kenny. 

Moreover, she says, her data shows that customers who have experienced the guarantee after a service problem report higher satisfaction with AT&T than customers who have not had an issue. 

The guarantee also gives AT&T another point of differentiation in a telecom market where competitors make similar claims around coverage, speed, reliability and price, says Kenny.

Marketing infrastructure customers rarely see

One of Kenny’s biggest marketing challenges is explaining the value of infrastructure customers rarely notice when it works as intended. For most consumers, the network shows up as reliable connectivity, while the technology that makes it possible remains largely invisible.

AT&T is trying to make those investments more tangible by alerting customers when connectivity in their area has improved, their home internet speeds have increased, or coverage along their commute has expanded.

FirstNet, the nationwide public-safety broadband network AT&T operates for first responders, is one way the company translates its infrastructure into a more concrete benefit. 

AT&T has found that mass-media advertising for FirstNet improves perceptions of the brand even among consumers who are not first responders and cannot use the service. Knowing that firefighters, EMTs, and other emergency personnel have priority connectivity during crises gives those consumers greater confidence in AT&T, Kenny says.

That gets at the broader case Kenny is making for marketing at AT&T. Brand strength has value when it influences business outcomes, whether it’s a new prospect signing up or an existing customer adding another service. Kenny’s effort to quantify those relationships, she says, gives marketing a way to measure brand in the same terms expected of the rest of the C-suite: through its contribution to revenue growth.

This story was originally featured on Fortune.com

“Once you learn that, it’s just so transformative to everything in life because then you pivot from feeling bad for yourself, feeling sorry for yourself, to how do I go overcome it?” he said in an Instagram post after a group of summer United interns which lessons have shaped his career and personal life.

It’s a lesson many Gen Z may have already learned the hard way, having entered the workplace amid shifting workplace norms, waves of layoffs, and a particularly tough job market. But Kirby’s advice is less about avoiding adversity than how to respond when it inevitably arrives.

It’s a philosophy the 58-year-old first learned while training as a pilot in the U.S. Air Force Academy—and one that was put to the test most publicly when he became United’s CEO in May 2020, just as the pandemic was bringing the airline industry to a standstill. With U.S. passenger traffic plunging by 60%, United operating revenue fell by 64.5% in 2020 and the company posted a $7.1 billion net loss. Kirby was forced to slash flights, burn millions each quarter in cash, and even take out a $6.8 billion loan using United’s loyalty program as collateral. He also took a 100% salary cut.

“You’re going to encounter challenges in business and in life,” he added as an Instagram caption. “You can spend your time explaining why it happened, or you can spend your time figuring out how to overcome it. I’ve always believed the second approach is better.”

From mowing lawns and selling fireworks to leading a $60 billion airline giant

Kirby grew up in a middle-class family in a farming community outside of Dallas, Texas, and has said he never had a grand career plan but developed an entrepreneurial streak early.

“I was always trying to earn money as a kid by mowing lawns, delivering newspapers, and, once I tried to start a company with a buddy,” Kirby told the East Valley Tribune in 2007. “We sold firecrackers—and we nearly blew ourselves up!”

After graduating from the U.S. Air Force Academy in 1989 with a bachelor’s degree in computer science and operations research, Kirby worked as a budget analyst at the Pentagon and later moved into the airline industry. He joined American West Airlines in 1995 and rose through the ranks, eventually becoming president of U.S. Airways in 2006. Following U.S. Airways’ merger with American Airlines in 2013, Kirby became president of the combined airline. He joined United as president in 2016 and was named CEO in 2020.

Above all, Kirby is a believer that self-confidence will lead you down a pathway toward success.

“I also believe in self-fulfilling prophecies,” he said. “By saying you’re going to do incredible stuff, you make it a whole lot more likely that it’s going to happen.”

An additional core part of Kirby’s leadership strategy has been surrounding himself with people who share his approach to work—and who genuinely care about one another. When he was looking to strengthen United’s pilot-hiring process, Kirby asked his head of flight operations to select a dozen well-liked pilots to help interview candidates.

“I told this group of pilots, ‘Your job is just to assess: Is this interviewee someone I would like to take a four-day trip with? And if you say no, then they’re out. You get a veto vote,’” Kirby said in a recent interview with McKinsey.

“The idea is to pick people who care about others, who you want to hang out with, who you want to be with.”

United has since emerged from the pandemic in a much stronger position. Its market capitalization is now roughly $41 billion, just behind rival Delta, at $60 billion. Fortune reached out to United Airlines for further comment.

The CEOs of Delta and Nvidia agree: don’t shy away from adversity

Kirby isn’t the only business leader who believes adversity can be a catalyst for growth.

Delta Air Lines CEO Ed Bastian has similarly emphasized the role of humility in navigating crises.

“Our motto is to keep climbing, and to always keep growing and keep learning and keep aspiring, and keep focused on where we’re going,” he told The Wall Street Journal, adding that crisis—whether the pandemic or the rise of jet fuel, “can make you stronger or they can make you fall back to the pack.” 

“We’ve always tried through learning, through humility to try to take from whatever we’ve encountered [and] become more resilient, become more differentiated, become more distinctive in how we deliver our service.”

Nvidia CEO Jensen Huang has taken a similarly counterintuitive view of adversity, arguing that having low expectations can actually make someone more resilient.

“People with very high expectations have very low resilience—and unfortunately, resilience matters in success,” Huang said at Stanford’s Graduate School of Business in 2024. “One of my great advantages is that I have very low expectations.”

Huang added, “I don’t know how to teach it to you except for I hope suffering happens to you.”

This story was originally featured on Fortune.com

On social media, she teaches followers about food storage and homesteading. Her audience can also buy a homesteading guide, take her food-storage course or shop at Arvoti, the skincare company she launched after creating a tallow balm for her father, who had eczema.

“Whether I’m sharing a day on the homestead or I’m talking on my podcast or making skincare, it’s all rooted in those same values of simplicity and quality and just living closer to nature,” the 29-year-old, who declined to provide a name beyond Gubba, told Fortune.

Her operation sits at the center of a growing creator economy around traditional living. As homesteading, cottagecore and tradwife content converge online, the overlap has made the category more marketable. Creators now offer branded lifestyles, selling courses and products to followers looking for different versions of a simpler life. In other words, the market for a less digital, more self-reliant life is increasingly being built through one of the most digital businesses of all: the creator economy.

Gubba estimated that Gubba Homestead and Arvoti now generate seven figures in annual revenue. She considers herself a tradwife but sees what she does as homesteading first. In her view, homesteading is about developing practical skills and becoming more self-reliant. Tradwife, by contrast, is more about relationships and household roles.

But the lines are blurring, as practices that once belonged to distinct online communities are increasingly grouped together, according to Sara Reinis, a doctoral student at the University of Pennsylvania’s Annenberg School for Communication who studies tradwife culture. 

Cottagecore, an online aesthetic that romanticizes an idyllic rural life through pastoral fashion, baking, gardening and domestic crafts, is often folded into tradwife content even though it does not necessarily carry the same expectations around marriage or household roles, Reinis told Fortune. Similarly, homesteading and tradwife culture can overlap, but they are not interchangeable.

“There’s a lot of the homesteading world that is not tradwife,” she said. “There’s a lot of the tradwife world that’s not homesteading.”

Even some of the creators most associated with the tradwife aesthetic resist the label, such as Hannah Neeleman of Ballerina Farm. Nara Smith, whose elaborate from-scratch cooking videos draw similar comparisons, has also rejected it, saying in a Call Her Daddy podcast, “I don’t think a trad wife would be cooking in Schiaparelli.”

For former tradwife Enitza Templeton, the important difference is not whether someone bakes bread or raises chickens. It is the household structure behind the image.

“You can’t be a tradwife on your own, but you can be a homesteader on your own,” she told Fortune.

Templeton grew up expecting marriage and motherhood to anchor her adult life. She said traditional gender roles were reinforced within her Puerto Rican family and the evangelical churches she attended.

After preparing to be someone’s wife her whole life, she intentionally sought a partner who wanted the same arrangement. She married at 26 and became a stay-at-home mother of four before divorcing in 2020.

Watching polished traditional-wife content now, Templeton recognizes pieces of the workload she once carried. What she questions is how much of it viewers are actually seeing.

“I could chop up my day and make it look that beautiful too,” Templeton said. “Anybody could, but that’s not the reality.”

Templeton’s view of tradwife life is considerably more critical than Gubba’s. But their interviews and Reinis’s research point to the same distinction: Learning to bake bread, preserve food or raise chickens is not necessarily the same thing as adopting a particular household structure.

That distinction matters commercially. The same visual shorthand—a sourdough loaf, an apron or a backyard flock—can attract viewers looking for very different things. Some may want a canning technique or gardening hobby. Others may be drawn to the promise of a more traditional household life. In either case, a creator has an audience for products, courses and content.

“The tradwife fantasy is also about the fantasy of being an influencer who can make money at home,” Reinis said to Fortune. 

A younger market for an older way of life

The market is not confined to social media feeds. Tractor Supply, the largest rural-lifestyle retailer in the United States, told Fortune that interest in categories associated with homesteading has outlasted the pandemic-era boom. Backyard poultry, gardening and homesteading surged as people spent more time at home during COVID. The company said those categories have continued to deliver steady growth.

“We are seeing a fascinating demographic shift,” Kimberley Gardiner, Tractor Supply’s senior VP and chief marketing officer, told Fortune.

Customers shopping those categories tend to skew younger than Tractor Supply’s average shopper, she said. Millennials who moved into the company’s markets during the pandemic have stayed, while the retailer has seen a noticeable uptick in Gen Z consumers who are “stepping back from purely digital lives to seek out tangible, hands-on experiences.”

Canning, the method of preserving food by packing it into airtight containers, has seen especially strong growth among younger customers, according to Tractor Supply, while the average age of a new poultry customer is now more than a full year younger than it was a few years ago.

The company said creators are increasingly introducing consumers to gardening, backyard poultry and other hands-on hobbies. During its Chick Days campaign, Tractor Supply partnered with three creators whose content reached approximately 150,000 potential new customers.

For Gubba, the desire for something tangible helps explain why traditional living has found a receptive audience online. She began sharing food-storage advice around 2020, when empty store shelves and supply disruptions had people thinking differently about where everyday necessities came from. Interest from followers led her to create a food-storage course.

“I actually don’t think people are just drawn to the idea of being a tradwife per se,” she said. “I think they’re more drawn to the more intentional way of living.”

But she said the biggest misconception is that homesteading is about posting photos of the perfect kitchen or apron on Instagram. Social media can make the aesthetic easier to adopt than the work behind it, she said.

Homesteading also means practical skills like fixing fences and caring for sick animals, as well as continuing to work when things go wrong and just becoming more capable.

“People can copy an aesthetic,” Gubba said, “But it’s much harder to copy that way of life.”

This story was originally featured on Fortune.com

This is further complicated by the fact that AI is moving fast. Trillions of dollars are being spent globally on new initiatives, and the competitive landscape is being turned on its head. Every quarter, the list of worthy investments grows longer, and every leader I speak with can make a compelling case for why their initiative matters most.

Here’s what hasn’t changed: capital is finite. Yes, you could raise more money, but there is no inexhaustible pot of gold waiting to be given out. If money is going to one area, you’re making a trade-off and spending less somewhere else.

At ServiceNow, that is not a theoretical exercise. We recently completed our $7.75 billion acquisition of Armis — one of the biggest capital allocation decisions in our history, and a bet that closing the gap between asset visibility and cyber risk mattered more right now than half a dozen other initiatives competing for the same dollars. These are decisions about where we believe enterprise AI is going, what capabilities we need to own, and how much conviction we have before the ROI is obvious to everyone.

As President and CFO, I sit at the intersection of growth and financial discipline. It is my job to make deliberate calls about where to invest, when to wait, and when to say no — and, like many other enterprise leaders right now, I’m aiming at a moving target.

Here are the questions I believe every major investment decision must answer.

1. Does it deepen our competitive moat?

I stress-test every investment decision against a simple question: does it strengthen what is hardest to copy about our business?

Right now, that question carries more weight than ever. When intelligence is cheap and AI can produce functional code in minutes, a meaningful feature advantage can be matched by your competitor in weeks. That raises the bar for what is actually worth funding.

Investment must now balance strategic parity — ensuring you aren’t left behind — with the differentiation required to be a market leader. Increasingly, one path to achieving this is pairing AI with proprietary data, hard-won expertise, and systems built over years.

Take JPMorgan Chase, which built its LLM Suite platform in-house and connected it to the firm’s own data and systems, creating a unified and unique AI resource that others can’t easily duplicate. At ServiceNow, we’re building on a different set of advantages: 20+ years of helping customers execute more than 100 billion workflows, which has given us deep domain expertise, proprietary data, and a massive install base of customers embedded broadly and deeply across our platform.

For every company, the moat will look different. The point is to be honest about the aspects of your business that are genuinely hard to replicate, and to invest in whatever compounds that advantage.

It also means being practical about the path you take to get there. We pride ourselves on being an organic growth and innovation machine. In a market moving this quickly, though, even organizations with a strong build-it-ourselves culture must be open to inorganic plays that bring in critical capabilities and talent faster than they can be developed internally. For many companies, it’s one of the harder shifts this moment requires.

2. Are we funding a real customer need?

The voice that should drive investment decisions is often the one that is not in the room: your customer.

One of my top priorities is making sure we have incredible feet on the street, working with customers to understand their pain points and challenges so we can help them innovate and create value.

One example: we heard from many enterprise customers who were struggling with fragmented AI efforts across their organization. Multiple initiatives were running in parallel with no central visibility or governance. That feedback led directly to an investment in developing what we call the AI Control Tower, a central hub for managing AI across the enterprise.

Some of the most expensive investment mistakes happen when there is a disconnect between what customers need and the products or innovations a company chooses to invest in. If you cannot trace a direct line from a customer insight to a major investment decision, that is a red flag.

3. Are customers adopting what we built and getting measurable business value from it?

An investment decision does not end once an initiative is greenlit — or even when sales are made and customers are onboarded. You have to care about whether customers are actually using what you built, and if it is embedded deeply in their operations and delivering real value.

I believe the teams closest to customers post-sale are often the best early-warning systems in the business. They see friction first and hear where adoption is stalling, or workflows are breaking down.

This is even more critical in this moment of AI adoption, where we know the real challenge lies in execution. According to ServiceNow’s own Enterprise AI Maturity Index, 59% of organizations are using agentic AI, but only 9% have made significant progress in creating autonomous, multistep AI workflows. That means companies are paying for capabilities they haven’t yet unlocked, so they’re not seeing the value they’re hoping for.

Of course, when your customers don’t see value, you’re inviting churn. At enterprise scale, even a single point of revenue retention can be worth hundreds of millions of dollars. This is money that should be driving investments in the right innovations and funding projects that create a competitive edge. Instead, it simply vanishes from the balance sheet.

Balancing bold bets with discipline

In my career, I have led through periods of real pressure. But the pressure companies feel right now to move quickly on AI is at a whole new level. That means companies must stay agile without becoming reactive. As I often tell my team, AI is creating incredible opportunities, but opportunities without prioritization are just noise.

These decisions are also never made in a vacuum. The key to success lies in making sure they are aligned across the business, grounded in what customers actually need, and tied to real value creation, not just experimentation. This is where discipline matters most: when something is not working, you have to be willing to close off the spigot and reallocate capital toward what is.

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