A narrow group of AI enablers now carries nearly half the total value of the S&P 500 — some measures now put the AI-linked share closer to 50%–57%, meaning the concentration is arguably even more extreme than the headline number suggests — and the buildout has become so massive that it is driving the lion’s share of U.S. GDP growth. The overarching consensus in Silicon Valley and on Wall Street is clear: The technology works, and the capability is unprecedented; ergo, mass adoption is an inevitability.

History teaches a different lesson, and the last few months have provided a series of flashing red lights regarding what could lie ahead. In April, a man motivated by anti-AI sentiment attacked OpenAI CEO Sam Altman’s home. In May, college graduates entering a workforce where the technology is the primary reason for job cuts met commencement speakers hailing the AI revolution—including former Google CEO Eric Schmidt at the University of Arizona—with boos.

June began with Anthropic co-founder Jack Clark publicly urging governments to build regulatory mechanisms to slow the technology down and President Donald Trump signing a cybersecurity executive order establishing a voluntary, 30-day national security review for frontier models. And we learned that frontline workers—particularly those from Gen Z—increasingly admit to quietly undermining and sabotaging the efficacy of their employers’ internal AI rollouts.

The tech sector and the broader corporate world rushing to deploy its tools are increasingly suffering from technological determinism: the fallacious assumption that society is a passive operating system that will automatically update to accommodate new capabilities and infrastructure.

Disruptive technologies do not scale based on their abstract, mathematical capability. They scale based on human incentive structures. Human nature doesn’t necessarily limit technological innovation, but it determines the terms of its adoption. Stakeholder alignment will determine how—and where—AI is successfully adopted. When you align the incentives correctly, adoption moves at warp speed. When you ignore them, the human system will build a firewall to freeze the technology in place.

When People Push Back

What happens when an innovation is rolled out in a way that threatens stakeholders while concentrating 100% of the efficiency gains at the top? Look no further than what I call the Jersey Pump Principle.

As anyone who lives in or drives through New Jersey knows, it’s illegal for drivers to pump their own gasoline in the Garden State. In the late 1940s, the invention of the automatic shut-off gas nozzle enabled consumers to safely fill their own tanks without spilling gasoline or sparking a fire. This mechanical breakthrough fundamentally revolutionized the retail fueling model, shifting labor away from the pump and ultimately paving the way for the modern, high-volume convenience store layouts we have today. Yet several states, including New Jersey in 1949, banned drivers from pumping their own gas because independent station owners and labor advocates wanted to keep the status quo and protect jobs. Nearly 80 years later, New Jersey remains a lone holdout, legally locking in thousands of attendant jobs. A superior technology was frozen in time because the human terms of adoption were ignored.

We saw a contemporary, multibillion-dollar version of the Jersey Pump Principle on the streets of San Francisco and Austin with the rollout of autonomous robotaxis. Tech companies secured regulatory approval to expand driverless fleets 24/7. But because the rollout was completely divorced from the human element—marginalizing local drivers, drawing ire from fire departments whose emergency vehicles were repeatedly blocked and offering zero upside to residents—public retaliation followed. Activists discovered they could completely paralyze autonomous vehicles simply by placing cheap, orange traffic cones on their hoods, which tripped the cars’ sensors into permanent safety stalls.

Following highly publicized accidents and immense localized friction, California regulators suspended deployment licenses, with the DMV citing an “unreasonable risk to public safety” in ordering Cruise’s fleet off the road indefinitely. When you offload 100% of the risk onto a community and capture 100% of the reward, human nature will find a low-tech way to break your high-tech machine.

The Efficiency Dividend: When Incentives Align

The inverse of the Jersey Pump Principle is equally true: The right incentive structure can hyperaccelerate adoption with virtually zero friction.

When Henry Ford rolled out the moving assembly line in 1913, it reduced Model T build times from 12.5 hours to 93 minutes each. But for workers, the extreme monotony turned the factory floor into a human disaster. Ford’s labor turnover rate skyrocketed to a staggering 370% in 1913, forcing him to hire more than 50,000 workers just to maintain a baseline workforce of 14,000. The technology was stalling because the human terms were broken. In 1914, Ford reengineered the incentive structure. He doubled the standard industry wage overnight to $5 a day and cut shift length by an hour. The trade-off was clear: Accept the monotony of the assembly line in exchange for immediate financial security. Attrition plummeted, productivity soared and Ford’s profits nearly doubled within two years.

Decades later, Toyota achieved a similar breakthrough with the Toyota Production System. While Western automakers used early robotics as a weapon to slash headcounts, Toyota took the threat of human obsolescence completely off the table. They gave frontline factory workers lifetime employment guarantees and installed the “Andon Cord”—a physical rope that gave any individual worker the power to completely halt the multibillion-dollar assembly line if they spotted an inefficiency. Because automation was tied to worker empowerment rather than displacement, factory workers actively partnered with engineers to optimize the robots, resulting in defect rates lower than those of their American competitors.

The C-Suite Mandate: Designing an Aligned Incentive Structure

History contains a direct mandate for every CEO currently buying enterprise AI licenses. Right now, the overarching incentive structure driving corporate AI deployment is strictly extractive. Executives look to AI to slice headcounts, monitor employee outputs and squeeze margins out of fewer hours. The average employee looks at these expensive software deployments and asks a highly rational question: “Why should I help make this technology successful?” If the only reward for mastering an AI workflow is a higher performance quota or an impending pink slip, workers will quietly slow-roll, sandbag and sabotage the transition.

Technology is a tool, but adoption is a human choice. To navigate this shift successfully, leaders would be well served to conduct a rigorous examination of historical precedents, analyzing exactly how their specific organizations and industries have responded and adapted to past disruptions. The strategic insights needed to solve today’s cultural anxieties are already encoded in industrial history.

A powerful — and more honest — blueprint for this kind of analytical approach is unfolding right now within the automotive sector’s pivot to electric vehicles. Autoworkers initially viewed EVs as an existential threat, fearing that simpler electric drivetrains would mean far fewer assembly jobs. (That fear turned out to be overstated — later plant-level research found BEV assembly actually required comparable or more labor than combustion-engine lines — but the perception of threat was enough to trigger resistance.)

That resistance came to a head in September 2023, when the UAW launched its first-ever simultaneous “Stand-Up Strike” against Ford, GM, and Stellantis. The 46-day strike — not quiet negotiation — is what produced the “just transition” terms now shaping the EV rollout: priority transfer rights, an end to wage tiers, and explicit commitments to bring battery-plant jobs under national union agreements. By removing the threat of human obsolescence, the industry’s retooling is now moving forward with far less friction than it otherwise would have — but the lesson isn’t that incentive alignment avoided a shutdown. It’s that the shutdown is what forced the incentive alignment. For AI executives hoping to skip the messy leverage phase, that’s the less comfortable but more useful precedent.

The lesson for the AI era is clear: If you want your organization to genuinely embrace this transition, you need to move beyond short-term operational consolidation. The path forward requires structural incentive alignment. When employees realize that automated efficiency directly yields a tangible share of the productivity dividend—whether through relief from systemic burnout, clearer paths to high-value work or guaranteed career upskilling—they stop fearing the future and start actively helping you build it.

Trillions of dollars in capital and compute are waiting on the sidelines—but if your people don’t have a clear, personal incentive to build the future alongside you, they will use their collective inertia to ensure your expensive new technology stays underutilized. Just ask drivers in New Jersey.

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

That’s been the work of our team at J.B. Hunt for 65 years. There have been many changes since J.B. and Johnelle Hunt started the company, but the culture and principles that built J.B. Hunt remain the same: take good care of our people, who take good care of our customers, and that takes care of our business. 

Today marks our 65^th anniversary. That milestone reflects more than longevity. It’s a testament to an enduring culture led by trusted people, resilience in navigating changing markets, and the confidence to invest through uncertainty and keep building when the next step is not always clear. 

That mindset has shaped every major chapter of our growth. Intermodal, truckload, dedicated, brokerage, last mile and our technology innovations were all practical solutions to customer problems, strengthening our network and positioning the company for what comes next.

That’s what I mean when I say J.B. Hunt is a growth company. Growth is not only more freight or more revenue. It’s the ability to reimagine what’s possible for our people, our customers and the industry.

Scale must drive outcomes

In transportation, scale only matters if it creates value for customers. A larger fleet, a broader network or a stronger platform must translate into execution that helps their businesses run better.

Our work has to stay focused on practical improvement and disciplined growth. Capacity should be easier to access. Information should help people make better decisions. Technology should reduce friction in the work, bringing customers closer to the people who serve them.

For us, that means using scale to strengthen the parts of the network customers value most: improving service, identifying inefficiencies, creating better visibility and helping our people deliver solutions with more speed and consistency.

Technology is part of that work, but it’s not the goal itself. AI, automation and other tools have to be applied with purpose, tested in real-world conditions and proven through the work that gets done. 

Our work with UP.Labs reflects that mindset. We are co-designing and launching technology startups, bringing J.B. Hunt’s scale and freight expertise to problems that matter across the industry. 

Overroute™, the first startup launched through the initiative, shows how that model is moving from concept to market. Built around real freight workflows, it applies AI to surface exceptions, support coordination and help teams respond more consistently as conditions change. Another startup is expected to launch soon, with two additional concepts currently in development.

That work is part of a broader effort to apply AI and automation across J.B. Hunt to specific parts of the freight lifecycle, including planning, capacity matching, shipment monitoring, billing, cargo protection and safety. These capabilities help our people identify issues earlier, recognize patterns across large amounts of information and focus more time on the decisions and relationships that shape the customer experience.

That is where technology becomes a force multiplier. Not as a separate strategy, but as part of how the work gets done.

Speed without compromise

Speed matters in freight. Our customers and our industry are dynamic and constantly changing. Customers need quick, flexible solutions that fit their business. The industry needs to move with urgency and invest in innovative ways to do that well. 

We can move faster without losing what matters most. Integrity, respect, safety, innovation and excellence, our core values, are not tradeoffs. They are the foundation of our culture, business and relationships. Our safety culture is one of J.B. Hunt’s strongest differentiators. It’s evident in the decisions our people make every day and the trust customers place in us to move their freight with care, consistency and accountability.

Transportation is ultimately about service, and customers count on experienced people who understand their freight and can adapt when conditions change. The right kind of progress makes supply chains more dependable, our teams more effective and our customers more confident in the outcomes we deliver.

That’s why I believe the companies that lead the next era of freight will be the ones that move with purpose, stay close to customers and keep building the capabilities that make service better and safer over time.

After several challenging years, the freight market is beginning to change. Capacity is tightening, and customers are looking for providers who can help them build more flexibility and confidence into their supply chains.

Our recent results reflect the position J.B. Hunt is in today, shaped by disciplined decisions, focused execution and steady preparation for what comes next. In our recently reported second quarter, revenue increased 19% and operating income increased 32% year over year, driven by growth across the business as we continued to leverage investments in our people, technology and capacity to strengthen how our teams serve customers and scale the impact of their work.

After 65 years, I am proud of the company our people have built. I am even more confident in what we are building now: a stronger J.B. Hunt and a more efficient transportation network for the customers and communities who rely on us every day.

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

Will oil prices go up?

It’s impossible to predict the future of oil prices. Several factors determine the movement of oil, but it ultimately boils down to supply and demand. Again, when threats of economic downturn, war, etc. are high, the oil trajectory can turn rapidly.

How oil prices translate to gas pump prices

When you pay for gas at the pump, you’re paying for more than just the crude oil itself; you’re also springing for links along the chain, such as the refineries and wholesalers—not to mention taxes and local gas station markups.

Still, the crude oil aspect affects the final price most dramatically, as it typically accounts for more than half the price per gallon. When oil prices spike, so do gas prices. And frustratingly, when oil prices drop, gas prices tend to take their time drifting down to the lower price (sometimes referred to as “rockets and feathers”).

The role of the U.S. Strategic Petroleum Reserve

In case of emergency, the U.S. has a store of crude oil known as the Strategic Petroleum Reserve. Its primary purpose is energy security in case of disaster (think sanctions, severe storm damage, even war). But it can also go a long way toward softening crippling price hikes during supply shocks.

It’s not a long-term answer—more of an immediate relief to assist the consumer and keep critical parts of the economy running, like key industries, emergency services, public transportation, etc.

How oil and natural gas prices are linked

Oil and natural gas are both major energy fuels. A big change in oil prices can affect natural gas by extension. For example, if oil prices increase, some industries may swap natural gas for some segments of their operations where possible—which increases demand for natural gas.

Historical performance of oil

When examining oil’s performance, there are generally two major benchmarks:

• Brent crude oil is the main global oil benchmark.
• West Texas Intermediate (WTI) is the main benchmark of North America.

Between the two, Brent better represents global oil performance because it prices much of the world’s traded crude. And, it’s often the best way to track historical oil performance. In fact, even the U.S. Energy Information Administration now uses Brent as its primary reference in its Annual Energy Outlook.

Looking at the Brent benchmark across several decades, oil has been anything but steady. It’s seen spikes due to factors such as wars and supply cuts, and it’s also seen crashes from global recessions and an oversupply (called a “glut”). For example:

• The early 1970s brought the first big oil shock when the Middle East cut exports and imposed an embargo on the U.S. and others during the Yom Kippur War.
• Prices dropped in the mid-1980s for reasons such as lower demand and more non-OPEC oil producers entering the industry.
• Prices spiked again in 2008 with increased global demand, but it soon plummeted alongside the global financial crisis.
• During the 2020 COVID lockdown, oil demand collapsed like never before—bringing prices below $20 per barrel.

All to say, oil’s historical performance has been anything but smooth. Again, it’s hugely affected by wars, recessions, OPEC whims, evolving energy initiatives and policies, and much more.

Energy coverage from Fortune

Looking to stay up-to-date regarding the latest energy developments? Check out our recent coverage:

• Trump is betting on the U.S. naval blockade as more oil sneaks out of the Gulf
• Iran rejects U.S. talks for now as wait for Hormuz deal drags on
• Eastern Mediterranean Sea is fast becoming an alternative energy source for Europe

Frequently asked questions

This story was originally featured on Fortune.com

These questions surfaced repeatedly in my conversations last week at the Ai4 conference in Las Vegas, where I spoke with C-suite leaders, academics, and startup founders about how AI is transforming hiring.

Case in point: Kathleen Walch, director of AI engagement and community at the nonprofit Project Management Institute, told me she recently received multiple résumés that were “so overfit” for an open role that the candidates couldn’t plausibly possess all the skills they claimed.

“It’s annoying. And guess what? If you do it, you’re probably not getting calls for interviews,” she said in our onstage conversation, adding: “If you do that, you’re losing your authentic human self. Why are we hiring you? Because you’re a human. If we didn’t want to hire a human, we would create an AI agent.”

Still, the résumé has a role in hiring for now. But AI is putting more pressure on the interview to reveal what a candidate can do. Walch focuses on methodology, asking candidates to walk her through the exact steps they took when managing a complex project, including where they went wrong and how they adjusted.

Aisha Lawrey, Johns Hopkins University’s responsible AI portfolio lead, said she tests candidates’ ethical judgment, probing how they think about a fundamental question: “For all of these different AI tools, just because you can, should you?” 

None of this means candidates should abandon AI in their job search, though. Used smartly, it can help them discover opportunities they might otherwise miss. One person I spoke with last week had spent months manually applying for jobs matching the title he assumed he was most qualified for, without success. Then Indeed’s AI analyzed his résumé, his behavior on the site, and hiring trends in his industry, and suggested a different kind of role. He applied and finally landed a gig.

Indeed CIO and CSO Anthony Moisant told me that 70% of sponsored applications on the platform now come from recommendations its algorithm surfaces to candidates, rather than from searching and applying themselves. Humans, he argues, simply can’t spot every opportunity.

Kristin Stoller
Editorial Director, Fortune Live Media
kristin.stoller@fortune.com

This story was originally featured on Fortune.com

• Exclusive: The inside story of how Demis Hassabis lost the CEO role at DeepMind.
• OpenAI’s revenues are plateauing.
• Profits are not showing up at AI labs.
• Iran and the U.S. wait each other out over the Strait of Hormuz.
• U.S. unemployment would be 5% if people weren’t giving up on work.
• The Fed’s Warsh won a round vs. his critics.
• President Trump’s love of grass.

➡️ Did someone forward you this email? If you would like to receive this information directly, every morning before the markets open in New York, sign up here.

This story was originally featured on Fortune.com

I’m talking about disease, hunger, and poverty. The industry often seems to obsess over white-collar productivity, with tools like ChatGPT, Claude Code, and Gemini promising to basically automate everything we don’t want to do.

But I think I finally found an AI executive who may actually have her head in the right place. 

Fidji Simo, a French-born businesswoman and former CEO of Instacart, left her post as one of OpenAI’s top executives in early July. Now, she’s working to cure chronic diseases, including her own. Simo suffers from Postural Orthostatic Tachycardia Syndrome (POTS), which causes fatigue, fainting, nausea, headaches, and more. It’s considered incurable today, even though it affects millions of people, including the wife of OpenAI president and co-founder Greg Brockman.

In my interview with Simo, she explained how her new startup, ChronicleBio, aims to use AI to analyze blood. The results will help segment POTS patients into more granular cohorts based on the source of their symptoms, rather than lumping them all together as Simo says the medical industry often does today. This can make clinical drug trials more targeted and successful, rather than inconclusive.

“The complexity of these diseases made it that without AI they were incredibly difficult to solve,” she told me. “They’re multisystem, so you need to be looking at the state of the nervous system, the state of the immune system, and how it correlates with your genetics. All of that is a massive data problem that was very hard to get your hands around without AI.” 

After years of reporting on the tech industry, it’s refreshing to see a person in power trying to solve something so universal and personal. I know so many people with chronic conditions that doctors simply cannot solve, or even diagnose.  There is still a massive profit incentive for bringing new, successful drugs to market—the fact that she suffers from the disease herself may mean she’ll at least approach it in an equitable way.

Now, let’s see what she can do. Read my whole story here.

See you tomorrow,

Emily Forlini
X: @EmilyForlini
Email: emily.forlini@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

Ordinarily, hardware wallets like Coldcard provide top-notch security, since they store crypto offline, and a wallet owner is the only one who knows the seed phrase that will unlock it. This should make the wallets virtually unhackable. Unfortunately, the company behind Coldcard made a colossal screw-up by failing to provide a randomized process for creating seed phrases. Instead, that process was based on a predictable pattern, allowing hackers who possessed a sample seed phrase to use trial-and-error to guess others—and clean out Coldcard owners in the process. (You can read a full technical breakdown here.)

The question is how much this hack matters. On one hand, you can say it doesn’t matter much since, compared to Ledger or Trezor, Coldcard is an obscure brand that accounts for less than 2% of the hardware wallet market. And while the hack is a devastating event for those who got robbed, the episode affected only a relative handful of Bitcoin owners, which is likely why the broader crypto market barely moved.

On a symbolic level, though, the Coldcard breach matters a lot. That’s because the vast majority of hacks you read about involve insecure technology like bridges, or befall people slinging weird alt-coins on some platform they shouldn’t have messed with in the first place. This hack was different. The victims in this case were OG Bitcoiners who understood the technology and who kept their stash in cold storage. They behaved like true Satoshi disciples and got mugged nonetheless.

This is why some are viewing the Coldcard debacle not as just another security incident, but as something closer to a crisis of faith. The sentiment is understandable, but also overblown. The Coldcard hack didn’t occur because of any failure in Bitcoin, but because of the negligence of a single manufacturer who had one job and didn’t do it.

And for better or worse, the incident points to an awkward truth: Keeping your coins in an offline hardware wallet stays true to the spirit of Bitcoin, but it is also a nuisance. I love the idea of self-custody and have tinkered with hardware wallets in the past. But I found it’s just easier to rely on a reputable exchange to hold my very modest stash of crypto. Many of you likely feel the same. If the day comes when our political and economic system takes a dramatic turn for the worse, I’ll likely put some Bitcoin on a physical device—after I make sure the device maker is up to the job.

Jeff John Roberts
jeff.roberts@fortune.com
@jeffjohnroberts

This story was originally featured on Fortune.com

“So yeah, I think it’s true, for them it’s slipping.”

Dimon’s own family is an example of the American Dream: The banker’s grandparents immigrated to the U.S. from Greece without high school diplomas, but instilled in their children and grandchildren the need to find a purpose in life and pursue it diligently.

Dimon now leads the biggest bank in the U.S. and is leveraging its might to increase opportunities across the country: J.P. Morgan’s multi-year American Dream Initiative (ADI) focuses on six areas—including support for small businesses and access to skills-based training—to boost economic growth and help people live better lives.

The plan has some way to go if young Americans are to be convinced the historic ideal can be achieved. Only a fifth of adults aged 18 to 29 believe in the possibility of the American Dream, according to a report from AP-NORC, published in June.

While Dimon argues “I’m not a big fan of polls, and most of those people will be wrong,” he added he “completely understands” the anger felt by those at the low end of the wealth scale, and said policymakers from either side of the political divide should be working hard to fix policy that has allowed inequality.

“Whether you’re a Democrat or Republican, acknowledge the problem and then start to say, ‘How can you fix it?’ … That’s what we should be working on, solutions to that problem,” Dimon said. “I think it causes a lot of the polarization today, some of the anger today, that people see that some people are doing very well and they’re not.”

“And even the people who aren’t, by the way, good-hearted Americans, they’re not happy with that either. They don’t like to see their fellow citizens suffering at all. And they want to see that fixed, too.”

Dimon’s capitalist philosophy

Other private market voices have suggested that skepticism about capitalism is also driving societal divides. In his 2026 letter to shareholders, for example, BlackRock CEO Larry Fink wrote that “this is where much of today’s economic anxiety comes from: a deeper feeling that capitalism is working—just not for enough people.”

The rise of democratic socialists such as New York City Mayor Zohran Mamdani lends itself to this theory, and even Dimon (a self-described “full-throated, red-blooded, patriotic, unwoke, capitalist CEO”) says capitalism has its flaws.

However, the banker added that there should be a distinction between imperfect systems and their less desirable alternatives. He explained: “Every system has flaws. There’s never been a system that’s perfect …. The other thing is some of it’s just bad people. There are bad people in capitalist societies. There are bad people in communist societies … And when you listen to people talk about capitalism, they’re confusing all these things at once.

“But I totally agree. It has left behind certain people.”

This story was originally featured on Fortune.com

Good morning. Varun Krishna is running one of the boldest culture experiments in corporate America—and every CEO managing a legacy workforce should be watching. The Silicon Valley veteran has spent three years as CEO of Rocket Companies deliberately fusing West Coast talent density with Detroit grit, a bet he’s now scaling across $16 billion in acquisitions and a 500,000-listing alliance with Compass.

“Silicon Valley places a premium on intellect and intellect alone with the idea that, if you get a group of smart people together, you’re likely to win,” he told me. “Detroit’s success is a function of that, but it is also a function of tenacity, grit, perseverance, and belief.”

Krishna’s goal has been to blend the best of both. He’s come to appreciate the loyalty and commitment that characterizes Rocket’s culture. In Silicon Valley, he says, “you start to realize a lot of the talent is job hunting every 18 months.” But given that he estimates it can take five years to build something that’s truly transformative, loyalty is the virtue he’s trying to cultivate. “People work here because they like the culture. Mortgages are more the how, whereas the what is around home ownership.”

At Rocket, “data and evidence win all arguments,” Krishna says, pointing to Rocket’s acquisitions of Redfin and Mr. Cooper for $16 billion last year as examples of big bets. The company’s recent alliance with Compass to bring more than 500,000 new listings to the Redfin platform and transform the seller experience is another example.

And what can Detroit learn from Silicon Valley? A lot. “The beauty of Silicon Valley is you have this extremely high density of brain power, this mix of startups, venture capital and education,” he says, adding that it’s also the environment of meetups, incubators, late-night dinners, and constant learning. “We have to create that kind of environment, to get more of that sort of youthful entrepreneurial energy here.”

Half of Krishna’s leadership team came from places like Seattle, San Francisco, New York, and Boston. “It’s not just about doing everything in kind of the homegrown way, but it’s about blending it together with outside talent. And that culture is a function of one man.” He’s of course referring to Rocket founder Dan Gilbert who has “invented and reinvented this company.”  

I’m thrilled that Gilbert will be joining us as a speaker at the Fortune 500 Innovation Forum in Detroit on November 16 and 17.  Other speakers include General Motors President Mark Reuss, Chipotle CEO Scott Boatwright, Corning CEO Wendell Weeks, Craig Newmark of craigslist, Penny Pennington of Edward Jones, Macy’s chief Tony Spring, Honeywell CEO Vimal Kapur, NYSE’s Lynn Martin, Dave Bozeman of C.H. Robinson, Ford Foundation’s Heather Gerken and many more. You can learn more here.

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

This story was originally featured on Fortune.com

Hassabis, DeepMind’s co-founder and the lab’s longtime steward, was giving up the CEO title and moving to a chairman role, they said. Minutes later, the news broke that long-time chief scientist Jeff Dean and several other prominent researchers were leaving the company. 

Staff in DeepMind’s London office were shocked. They had questions. Was the move a promotion or a demotion? Would Hassabis retain real influence over DeepMind’s direction? And was Google finalizing a power shift from the London lab to its Mountain View headquarters?

One DeepMind engineer familiar with internal reaction told Fortune employees were “very mixed” on what the reshuffle meant, with some wondering whether the chairman role would ultimately end with Hassabis phased entirely out of the company.

In reality, the shock exit was just the tip of the iceberg of what has been a difficult year for Google DeepMind. Within the lab, three current and one former employee told Fortune tensions were already simmering. Delayed models, high-profile exits of talent, and a very public fight with its own staff over controversial defense contracts had left insiders wondering if Google had lost its way in the AI race.

Once a frontrunner, the company has recently ceded ground to rivals, especially OpenAI and Anthropic. This was most evident in delays around the release of the company’s Gemini 3.5 Pro model, which has missed three release deadlines in recent months. Three of the DeepMind engineers who talked to Fortune blamed the delays on the company failing to prioritize AI coding abilities. 

And now, for the second time in just four years, the company is again playing catch-up. 

Google DeepMind is struggling to retain its technical edge, keep hold of top researchers, and contain an employee revolt over its defense work. Pichai’s reshuffle—which employees say shifts power back to Mountain View and pulls DeepMind closer to Google proper—arrives at a critical moment for the company. But it’s also a gamble. Will new leadership restore Google’s AI momentum, or worsen internal tensions at a crucial moment in the AI race?

A leadership shakeup

In his note to staff on Wednesday, Hassabis said that he was handing over day-to-day operational responsibilities for the lab to allow him the “time and space to focus on the big picture and help influence what is to come to the best of my ability.” 

Instead of selecting a new CEO for DeepMind, however, Pichai announced that DeepMind’s chief technology officer, Koray Kavukcuoglu, was taking over day-to-day operations as senior vice president, reporting directly to him rather than holding a standalone DeepMind CEO title. Kavukcuoglu, unlike London-based Hassabis, works 6,000 miles away, in the California offices in Mountain View. A Google spokesperson said that Hassabis was moving into an expanded role to focus on advancing science and long-term strategy.

Three DeepMind engineers, who asked for anonymity to discuss company matters, told Fortune this represents a broader structural change and a shifting of power to Mountain View, where Google is headquartered, and away from London, where DeepMind was originally founded. 

Losing Hassabis as a CEO and not getting a new one, they said, also means the lab is being somewhat absorbed into the rest of the company and losing some of its independence. The first engineer called it a “small step, but gradually continuing the trend.”

“A lot of people are quite upset by it,” they said. “Both Demis exiting, and losing some of the separation from Alphabet and identity as DeepMind.”

For months, there had been rumors of a significant power shift to California, and concerns about the message that change would send to the London-based staff. Workers describe an ongoing “background pressure” of things moving to Mountain View and several prominent London-based engineers, including AlphaFold veterans John Jumper, Jonas Adler, and Alexander Pritzel, recently exited the company to join Anthropic.

The company had also been slowly moving reporting lines away from Hassabis for a while, according to the second engineer.

A Google spokesperson disputed the characterization that power was shifting to Mountain View and said Google DeepMind remained committed to its London presence. The spokesperson said Hassabis, along with much of Google DeepMind’s leadership, would remain based in the London office. They added that DeepMind would retain separate leadership and research autonomy under Kavukcuoglu.

From a commercial perspective, there may be an argument that things need to change, especially if the company is having issues shipping models in an increasingly competitive environment. Mountain View may well have run out of patience and want a tighter grip on its AI crown jewel.

“It’s very clear that they’re having a big internal debacle,” the second engineer, who spoke on the condition of anonymity, said. “This might have also contributed to speeding up some of these power shifts. Generally in Google, when something is significantly underperforming or not working, their solution is to just do a huge reorg, a reallocation of compute and priorities to try and fix it.”

DeepMind’s journey

This isn’t the first time Google has hit a stumbling block on its AI journey.

Back in late 2022, the company famously declared a “code-red” following the launch of OpenAI’s ChatGPT, fearing the impact on its core search business. Since then, its race to catch up has been somewhat successful. Its Gemini 3 model outperformed all its rivals and pushed OpenAI to declare its own “code-red” in late 2025, as Gemini’s benchmark gains and rapid user growth put pressure on ChatGPT’s lead. 

But since then, things appear to have gone in the wrong direction.

Google’s models are currently not competitive with the bleeding edge of the U.S. frontier on intelligence and coding benchmarks. Even some open-source Chinese efforts are also outshining the tech giant.

Gemini 3.5 Pro was supposed to be Google’s next big answer to advanced models such as Anthropic’s Mythos and OpenAI’s GPT 5.6. Unveiled at the company’s flagship I/O conference in May with a promise to ship “next month,” it missed that June target and another mid-July date. The company headed into August with the model still unreleased, even as Google shipped a trio of smaller Flash models around it and claimed it had already begun training Gemini 4.

Micah Hill-Smith, co-founder and CEO of independent benchmarking company Artificial Analysis, told Fortune that Gemini 3.6 Flash, the strongest model Google has shipped, now sits behind models from Anthropic, OpenAI, one or two leading Chinese labs, xAI, and Meta on his firm’s intelligence index, which measures overall performance of AI models.

“For a brief moment, we actually did have [Google’s] model on top as the smartest model in the world. Right now, it’s fair to say that [model] from Google is less compelling,” he said.

Many of the mistakes that led to this started last summer, when the company failed to prioritize the development of Gemini’s coding abilities, according to a third engineer. Google, they said, was preoccupied with protecting its search traffic and beating OpenAI on chat, while rivals were running ahead on agentic and coding capabilities. Google models now significantly lag rivals on coding agent benchmarks, according to Hill-Smith.

A Google spokesperson said Gemini 3.5 Pro was currently in testing and that the DeepMind team was already working on its next generation of models. The spokesperson acknowledged that the company needed to improve in coding and agentic capabilities, but said its teams were focused on that work.

Google perhaps failed to predict that coding abilities would play such a major role in consumer products just a year later. Developer tools such as Claude Code or OpenAI’s Codex, are now underpinning a new wave of consumer and general-use products. They include Anthropic’s Cowork—an agentic tool built on the same underlying technology as Claude Code that lets non-developers hand off multi-step knowledge work, like organizing files or drafting documents, rather than writing code—and OpenAI’s superapp, positioned as a one-stop entry point for AI tools.

“They dropped the ball a little bit [on coding],” the third DeepMind engineer said. “By the time you realize something like that…a number of months have passed, and now you have to play catch-up. Then it also becomes political.” Coding is now in “big focus” inside the lab, the second engineer added.

However, not winning on pure model performance doesn’t mean Google is down and out of the AI race. The company has other advantages such as vertical integration—at the application, cloud, and chip layers—that are structural advantages that can offset weaker performance at the cutting edge of raw model intelligence.

“For OpenAI and Anthropic, because the model layer is so important to what they do, it’s pretty existential for them to win there,” Hill-Smith said. “Google could end up doing very well in the AI race overall, even if they’re not at the frontier of the model layer,” he said, adding that the company’s unusually deep vertical integration is a structural advantage that could well offset weaker performance at the cutting edge of raw model intelligence.

Falling behind—and jumping ship

The technical difficulties and feeling of playing catch-up twice in less than five years have been a knock to morale, all the current and former DeepMind workers who spoke to Fortune say.

That internal frustration has coincided with one of the most damaging stretches of departures in DeepMind’s history. In a single week in June, Google lost Gemini co-lead Noam Shazeer to OpenAI and Nobel laureate and AlphaFold co-inventor John Jumper to Anthropic.

The employees who spoke to Fortune describe the exits as driven by overlapping forces: aggressive poaching from cash-flush rivals, frustration at being behind in the AI race, and engineers racing to get into a pre-IPO lab before they go public, hoping the equity compensation will make them fabulously rich. “There’s a lot of just hopping over to pre-IPO labs,” one employee said.

A Google spokesperson said talent moved regularly between technology companies and that Google remained focused on attracting people aligned with DeepMind’s mission.

Google has tried to combat some of the exodus, with the second engineer describing big retention and equity bonuses being deployed for some staff, and a prior round of enhanced compensation during a period when Meta was making aggressive counteroffers to AI researchers.

However, employees say, it hasn’t been enough to offset a deeper problem of burnout.

“I’ve been pushing 60-hour weeks for quite a long time because there is so much to do,” the third engineer said. “People are pushing themselves really hard because it is a race, and you have this feeling that you need to catch up… [but] telling people to work 60-hour weeks constantly and that it is constant crunch time usually doesn’t work that well, and people do burn out.”

Google said AI was an intensely competitive field and pointed to researchers who had publicly expressed excitement about their work and the company’s direction.

Employees had issues with leadership too. Even before his exit, workers described Hassabis as absent, especially compared to Google co-founder Sergey Brin, who came out of retirement to help the company’s AI efforts.

“You will see Sergey being incredibly hands-on, working hard on stuff and being involved,” the third DeepMind engineer said. “I can’t say that I’ve ever seen Demis walk around the office in the Gemini area.”

“Demis isn’t a very inspiring leader,” the former employee added. “Certainly not the way people talk about Dario [Amodei] … Demis is more of a guy who is known but not seen.”

A union battle and contentious defense deals

Frustration over the leadership and talent exodus has collided with a separate, more politically charged fight over Google’s defense work.

In April 2026, Google signed a deal with the Pentagon allowing the Department of Defense to use Gemini models on classified networks for any lawful government purpose. More than 580 Google employees, including senior DeepMind researchers, later signed an open letter urging Pichai to reconsider the deal, citing concerns around autonomous weapons and mass surveillance.

Google was not alone in making the controversial deal: OpenAI, xAI, Microsoft, Amazon, and others signed similar classified-use deals. Anthropic was the only outlier, declining to drop its own restrictions on autonomous weapons and mass surveillance—a stance that led the Pentagon to designate the company a supply-chain risk and order the military to stop using its products.

At Google, the Pentagon deal reopened a longer-running fight over the company’s defense and intelligence contracts. Employees have also taken issue with Project Nimbus, a roughly $1.2 billion cloud-computing agreement Google and Amazon struck with the Israeli government in 2021.

For many employees, the Pentagon deal slipped largely under the radar. But for others it was existential.

Alex Turner, who spent more than two years on DeepMind’s AI safety team, resigned in June citing the Pentagon deal as the trigger.

In a public post, he said that his decision followed months of internal advocacy that went nowhere: he even tried to enlist Google chief scientist Dean as an ally. Andreas Kirsch, another DeepMind research scientist, was also a vocal public critic of the deal. Kirsch posted publicly at the time that the contract’s restrictions amounted to little more than “meaningless weasel words” for PR purposes, and said he felt “incredibly ashamed” to be associated with the lab over the decision.

The Pentagon deal was the final straw for a group of London-based DeepMinders, who launched a union bid in May in an attempt to win back some of their waning leverage. The action is the first formal attempt by a frontier AI lab to unionize.

The union drive really traces back to February 2025, when Google removed language from its public AI Principles committing not to build AI for weapons or surveillance that violates international norms. The decision angered many engineers who felt Google was backtracking on its ethical assurances. The company adopted that pledge in 2018 after the intense backlash over Project Maven—a separate defense contract.

Google has declined to voluntarily recognize the union, instead offering talks through the UK’s conciliation service, ACAS. The two sides have already clashed over basic questions, like which staff should be included in the bargaining unit, a person familiar with the situation said. Now employees worry that American leadership may be less inclined to play nice in negotiations.

A Google spokesperson said the company was still in the early stages of the union process and would continue to engage in negotiations in good faith.

DeepMind’s uncertain future 

With Hassabis stepping back and Dean gone, DeepMind appears to be entering a new phase at precisely the moment it can least afford further disruption.

While the lab still has advantages few rivals can match—distribution, cloud business, custom chips, vast research bench, and the ability to embed AI across products used by billions—the reorganization seems to remove a layer between DeepMind and a parent company increasingly focused on turning AI research into commercial—and government—deployments.

For employees, the question is whether new leadership and tighter integration will bring the urgency and clearer direction DeepMind needs to close the gap in coding and agents, or deepen the problems already driving people out: burnout, weakened autonomy, and concerns over how the company will use its technology. Just four years after Google’s first AI “code red,” DeepMind is again trying to prove it can move fast enough to keep pace. But this time, it will be doing so without the two figures most closely associated with its scientific identity.

This story was originally featured on Fortune.com