All News (Page 382)

A striking new story is taking shape in corporate America. Some of the most recognizable Fortune 500 CEOs are stepping down as AI becomes a defining question about what kind of executive is best suited to lead the next phase of the business.

That reflects a meaningful shift. For years, AI was treated as one strategic priority among many. Now, at companies like Walmart, Coca-Cola, and Adobe, it is increasingly looking like the dividing line between one leadership era and the next.

At Coca-Cola, James Quincey explicitly linked his decision to step down to the company’s “next wave of growth,” arguing that the company had made substantial progress under the old playbook but now faced a much larger AI-driven shift. The company’s own reorganization under incoming CEO Henrique Braun reinforces the point. Coca-Cola created a new chief digital officer role reporting directly to Braun and said the change was designed to bring the business closer to consumers and enable faster technology adoption across the enterprise. Braun, for his part, said the company was elevating digital leadership so it could move faster and work smarter across all markets.

At Walmart, Doug McMillon offered a similarly revealing signal. In announcing John Furner as his successor, he described him as uniquely capable of leading Walmart through its next AI-driven transformation. Furner, a longtime operator who rose from hourly associate to lead Walmart U.S., takes over as the company pushes deeper into agentic commerce and AI-enabled retail operations. He also brings leadership experience at Sam’s Club and is closely associated with Walmart’s broader digital acceleration.

Adobe’s situation is somewhat different, but no less telling. Shantanu Narayen’s planned departure comes at a moment when investors are scrutinizing Adobe’s AI positioning and questioning how well its subscription model will hold up against faster-moving generative AI competitors. Adobe has not yet named a successor, and that search is unfolding under unusually intense pressure to prove the company can lead in an AI-defined era. In his message to employees, Narayen wrote that “the next era of creativity is being written right now — shaped by AI, by new workflows and by entirely new forms of expression.”

Taken together, these transitions point to a new leadership test. Boards are not just looking for CEOs who can talk about AI or add tools around the edges. They increasingly want leaders who can reorganize large companies around faster decision-making, AI-enabled workflows, and operating models built for an era of greater autonomy.

Ruth Umoh
ruth.umoh@fortune.com

This story was originally featured on Fortune.com

AI healthcare deals are roaring. 

Digital health startups raised $14.2 billion in 2025, up 35% from 2024, with AI‑enabled companies capturing 54% of that capital and enjoying roughly a 19% premium on average deal size versus non‑AI peers, research from Rock Health details. Investors have rushed into ambient scribes, agentic triage tools, and “doctor‑copilot” platforms—including Abridge (an AI-powered clinical scribe) which raised about $550 million across two mega‑rounds in 2025.

But Zocdoc CEO Oliver Kharraz is spending his time asking a less glamorous question: What actually happens when patients show up in the exam room armed with AI answers? 

Zocdoc—backed by Francisco Partners, Atomico, Baillie Gifford, DST Global, and Goldman Sachs—was valued at roughly $1.8 billion in 2015 after a $130 million round led by Baillie Gifford and Atomico, making it one of New York’s highest‑valued private tech companies at the time. In 2021, Zocdoc raised $150 million in growth financing from Francisco Partners after growing revenue more than 35% year over year pre‑pandemic.

Now pitching itself as “healthcare access infrastructure,” Zocdoc says millions of patients each month use its marketplace to find in‑network doctors. 

The company’s latest survey of 1,186 U.S. adults and 1,000 providers focuses on the interaction between AI and patient care. Zocdoc found that 26% of patients have already asked an AI a health-related question, and 85% of providers say they’re seeing more AI‑informed patients. Yet more than 1 in 5 patients admit they’ve hidden their AI use from their doctor—often out of fear of being judged. 77% of providers say they feel positively about patients using AI and 60% would rather they use AI than Google.

That disconnect is creating friction. Patients come in “anchored” to an AI-generated answer but won’t admit it, Kharraz told Fortune, forcing physicians to “shadow box with an unnamed partner” as they unwind advice that “might not apply to a patient’s specific case.” Zocdoc’s data backs him up: 83% of providers say they have to correct AI information. The most consequential finding: nobody actually wants a robot doctor. Seventy percent of patients say they would prefer to receive medical guidance from a doctor rather than AI, and 65% would rather ask a doctor their medical questions. But AI is filling a very real access gap—65% of patients say they’ve consulted AI because it’s easier than seeing a doctor, with average wait times to see a primary care provider now topping 31 days in the U.S.

Both patients and providers converge on a narrower job description for AI. Their No. 1 use case is the same: preparing better questions for the doctor. Kharraz’s advice to patients: Don’t ask AI for a diagnosis. 

“It’s not as if you can keep AI use a secret. The interesting challenge for organizations like ours is helping mediate that patient-doctor relationship,” Kharraz says.

See you tomorrow,

Lily Mae Lazarus
X: @LilyMaeLazarus
Email: lily.lazarus@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

Good morning. On Fortune‘s radar today:

• The S&P 500 is flirting with correction territory.
• EXCLUSIVE: Don’t hold your breath for artificial general intelligence.
• War in Iran is set to escalate.
• New jobs number incoming.
• Europeans increasingly decline to answer questions about their savings.

This story was originally featured on Fortune.com

In March, Robinhood announced its Platinum credit card, whose perks include generous travel rewards, $250 in annual DoorDash credits, and a free membership to Amazon One Medical. The name of the new card, which has a not-so-low annual fee of $695, is both an homage and a flex: It echoes the card brand made famous by American Express, though Robinhood points out its version is the only one to be “plated in 99.9% pure platinum.”

The offering is the latest splashy option in the fast-expanding world of premium credit cards that are branded not as simple payment tools, but as lifestyles. In this world, “members” enjoy access to concerts and upscale gym memberships, and the opportunity to load up on free goodies from retailers like Lululemon and Apple.

For the well-disciplined, the high-fee cards are a good value thanks to a combination of perks plus rewards for spending that can be cashed in for a host of travel offerings. Even better, all of this comes tax-free, thanks to a legal quirk that treats credit card swag as “redemptions” rather than income.

But not everyone is pleased. In recent months Congress and the White House, mindful of rising credit card debt and growing merchant fees, have renewed a push to pass the Credit Card Competition Act (CCCA), which could make it much harder for card issuers to offer all those perks. That raises a problem for points hunters: Is the go-go era of rewards nearing its end?

Jamie Dimon’s bet pays off

“I wish it was a $400 million loss,” JPMorgan Chase CEO Jamie Dimon famously declared in 2017. He was responding to investor complaints over a $200 million earnings charge the bank had incurred from huge sign-up bonuses tied to its Chase Sapphire Reserve card. Dimon’s comments reflected a bet that the new premium card would, over time, become a big moneymaker.

The calculation proved correct: Today the card is incredibly popular and has helped the bank attract a generation of premium customers to its other services. Indeed, that’s one of the main rationales for banks issuing these lifestyle cards. At the same time, however, JPMorgan has gradually raised its annual fee from $450 to $795, while reducing the redemption value of certain rewards points. American Express, meanwhile, has raised the annual fee for its flagship Platinum card to $895. Changes like these have led some consumers to question whether the potential to capture loot is worth the upfront cost.

Moshe Orenbuch, a managing director at TD Securities, says that JPMorgan Chase and others would argue the card offerings are more generous than ever—they’re just distributed differently. Many top cards, in addition to offering rewards for spending, now provide credits—usually of $5 to $20 a month—for services like Lyft, DoorDash, and Disney+ that can stack up to thousands of dollars a year in value.

“They are trying to create an ecosystem,” notes Sanjay Sakhrani, a card industry expert at KBW. “Ultimately they want to make this not having a card but having an experience.” And for some members of the card issuers’ web of merchant partners, tie-ups with credit issuers translate into big money. Orenbuch notes that Delta Air Lines alone has collected as much as $10 billion from Amex in recent years for supplying seats on its planes to rewards customers.

Chase’s and Amex’s premium cards have been doing such brisk business that new challengers are leaping into the category. In addition to Robinhood’s Platinum card, there is Citi’s $695-per-year Strata Elite, whose debut last year was marred by an application-process bungle that saw the bank freeze thousands of accounts—but which has proved popular nonetheless.

The surge in usage, however, has come with growing pains—most notably at airport lounges. At venues like Amex’s Centurion Lounge and Chase’s Sapphire Lounge, cardholders can enjoy plush seats, chef-made nibbles, and free Chardonnay. But as the cards get more popular, road warriors are increasingly encountering crowds, long queues, and wait times.

The downsides of fat rewards

The glamorous branding of premium cards can also lead some consumers to make foolish mistakes by running up high-interest credit card debt. Sakhrani notes that some premium card customers quickly find themselves carrying monthly balances with interest rates of over 20%—an obligation that can quickly dwarf the value of any rewards they earn.

“Consumer credit is not intuitive. Plenty of people who are otherwise smart can overestimate their own ability to manage credit cards,” says Beverly Harzog, a former CPA and personal finance author who has written about her own experience with card debt. She notes that while some are assiduous about amassing a given card’s full rewards value, many will come to the very reasonable conclusion they can’t risk the costs. In these cases, she suggests people choose a slightly less premium card like the Capital One Venture Rewards card, which can still offer valuable perks but for an annual fee closer to $100. The frugal-minded, meanwhile, may prefer a no-fee, cash-back card like the Citi Double Cash card or the Apple Card.

Merchants, meanwhile, are frustrated by one feature of premium cards: They force businesses to pay higher swipe fees compared with plain-vanilla ones. The CCCA, backed by many of these businesses, would lower the cost of these transactions. President Trump expressed support for the bill early this year, calling for an end to the “out of control Swipe Fee ripoff” and a temporary cap of 10% on monthly interest.

If any of these proposals come to pass, analysts say, banks would be forced to dramatically scale back rewards and turn their “lifestyle” offerings back into ho-hum instruments of credit. For now, though, that appears unlikely. The powerful bank lobby has a growing list of allies—including airlines and hotel chains—that will likely push to preserve the status quo. The good times should continue to roll, letting disciplined consumers pad their incomes with free stuff for the foreseeable future.

This article appears in the April/May 2026 issue of Fortune with the headline “Credit card rewards are more lavish than ever—but you have to work harder to cash in.”

This story was originally featured on Fortune.com

• In today’s CEO Daily: Diane Brady reports on CEOs’ growing frustration with the Trump administration.
• The big leadership story: A yawning workplace ‘design gap’ could stall productivity gains.
• The markets: Mixed globally as the Iran war enters its fifth week
• Plus: All the news and watercooler chat from Fortune.

Good morning. Will a war-induced recession inspire CEOs to speak out against the Trump administration? Economists like Moodys’ Mark Zandi say odds of a recession are now high. We know that most U.S. CEOs disapprove of Trump’s leadership, from his administration’s policies around tariffs and immigration to its approach to science, free speech and rule of law.

While business leaders might not have wanted the U.S. to start a war against Iran right now, they’re divided about when to end it. At the annual CERAWeek gathering in Houston last week, energy leaders from Dow CEO Jim Fitterling to Chevron’s Mike Wirth warned of dire consequences if the Strait of Hormuz isn’t opened to shipping as soon as possible. But JPMorgan’s Jamie Dimon said the war could mean a “better chance” of permanent peace in the Middle East, while BlackRock CEO Larry Fink predicted the war could result in prosperity or a global recession—but not much in-between.

What’s clear is that no one is winning the war at the moment. Oil prices are up more than 50%, forcing Asia to hunt for alternatives. Russia isn’t gaining much, thanks to its war with Ukraine. It’s costing U.S. taxpayers about $1 billion a day, and that doesn’t include the 10,000 jobs lost from the economic impact. The people who’ve paid the steepest price, of course, are the 3,000+ who’ve been killed and more than 4.2 million displaced, according to U.N. estimates.

At some point, there may be too much to ignore. I didn’t see much evidence of high-profile business leaders among the estimated 8 million people attending the 3,300 anti-Trump No Kings protests on Saturday. But I do see signs of mounting concerns: Chubb CEO Evan Greenberg told me last week that “democracy is so fragile,” Citadel’s Ken Griffin revealed that he and his CEO peers find the current government’s favoritism “extremely distasteful,” and more than 60 CEOs, including leaders of 3M, Best Buy, Cargill, General Mills, Land O’Lakes, Target, Xcel Energy and UnitedHealth Group signed that letter of protest against ICE actions in Minnesota. One CEO confessed to me recently that they are “shell-shocked” by the administration’s policies but feel a fiduciary duty to not put their company in harm’s way by speaking up. If the war starts to seriously impact stock prices and profits, that could change.

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

This story was originally featured on Fortune.com

As recently as a few years ago, Dell appeared to be a name destined for the business history books. The stock lost nearly a third of its value in 2022, and it was hard to see the once iconic PC-maker’s place in a post-PC world. Then something extraordinary happened. In the span of two years, Dell quietly built a $25 billion AI infrastructure business from scratch, posted total company record revenues of $113.5 billion, and it is now guiding Wall Street toward $50 billion in AI server sales next year alone.

Recently Fortune visited Dell’s CFO in New York to find out more about how the company pulled off something almost no company of its size has managed to do: Reinvent itself in real time. David Kennedy is a 27-year veteran of Dell, and was confirmed as CFO in November 2025 after serving as interim. In a conference room overlooking the chaos of 34th Street, Kennedy walks me through the recent record performance posted by the Round Rock, Texas-based giant (No. 44 on the Fortune 500).

“If you take the 12 months we’ve just finished, we did $34 billion in AI-optimized server orders in Q4, which tapped us up to $64 billion for the full year, and we exited the year with $43 billion in backlog,” Kennedy said. Fourth-quarter AI server revenue alone surged 342% to $9 billion. “What’s super exciting is our next five-quarter pipeline of opportunities has never been higher.”

For fiscal 2027, Dell has guided for $50 billion in AI-optimized server revenue, representing 103% growth year-over-year. Kennedy attributed demand to global interest across neo-clouds, sovereign AI deployments, and Dell’s enterprise base. “The fear of being left behind,” he said, “is becoming more powerful.”

Bank of America analysts recently raised their forecasts for Dell’s AI-servers, increasing their estimate for the current quarter to about $15 billion and lifting their full-year projection to roughly $60 billion, citing stronger-than-expected demand. Morningstar also increased its fair value estimate, noting that sustained AI demand will be key to long-term upside.

If there’s a cloud over the otherwise sunny outlook, it’s supply. Kennedy was direct: there simply aren’t enough components in the ecosystem to fully satisfy AI infrastructure demand. “I’d love more supply still,” he said. 

But Kennedy argues Dell’s multi-decade supplier relationships give it an edge over competitors in securing what’s available. And unlike some peers, Dell provided a full-year fiscal 2027 guidance — a signal, Kennedy said, that the company has supply commitments to support it.​

On the question of AI server profitability, a topic that has made some investors nervous, Kennedy was unfazed. Dell targets mid-single-digit operating margins on its AI infrastructure business, a figure it has maintained consistently. “Mid-single digits on $50 billion,” he said, “is a lot of dollars.”

At the core of Dell’s strategy is what Kennedy calls an “AI factory,” which is an end-to-end infrastructure stack built around data. That includes GPU-powered servers developed with Nvidia, a large-scale storage business, and networking systems.

“It’s all about data,” Kennedy said. “How do you manage it, store it, use it, deploy it?” He said Dell’s ability to build, deploy, and service systems at 99.9%-plus uptime has helped differentiate it and strengthen customer relationships.

The company now has more than 4,000 enterprise AI factories deployed with customers, including more than 750 added in the fourth quarter alone.

Inside the finance function: Agentic AI

Dell has spent the past two years modernizing and standardizing its systems to prepare for broader AI adoption. That foundation is now enabling the company to scale agentic AI internally, Kennedy said.

The OpEx discipline has come with a workforce reduction. Dell’s total headcount fell roughly 10%, or about 11,000 employees, in fiscal 2026, according to its 10-K filing, which is the third consecutive year of comparable declines. The company has spent $569 million in severance in the most recent fiscal year. Dell said in its 10-K that fiscal 2026 headcount reductions stemmed from employee reorganizations, limits on external hiring, and other cost-alignment measures tied to its business modernization efforts. “Despite these difficult decisions, we continue focused efforts to empower our employees and attract, develop, and retain talent,” the company stated.

Regarding agentic AI, Kennedy has a focus on the finance function. “I’ve started to deploy agents to do reconciliations, do accounting journal entries,” he said. “We’ve launched digital twins in our supply chain and services organizations. We have our own internal sales chat CRM model, which has handed back multiple hours per week to our sales force.”

Kennedy has gone further personally — incubating a team of data scientists within his finance function and building proprietary agents under Dell’s internal governance framework. He uses AI to streamline his calendar, automate emails, and drill down on forecast data by country and segment.

His view on workforce impact is that AI redistributes effort toward higher-value work. “The accountability level is still there,” he said, pointing to relationships with auditors and regulators. “All they’re doing is getting help in getting faster decisions, quicker.”

He also emphasized the importance of data quality and effective prompting: “You’re only as good as the data you have, so you’ve got to make sure that’s clean. And then trying to direct the agent in the right format — because an agent wants to work 24/7.”

This story was originally featured on Fortune.com

Alon Chen joined Google in 2006 at 23, with no marketing experience and no connections at the company. By 28 years old, he was a CMO—overseeing marketing for Israel and Greece, building a $2 billion product line across 30 markets, pulling in a highly six-figure salary and a seven-figure equity package. 

By most people’s standards, he had made it absurdly early—and he says getting there was “easy,” too. Not because of mentors, politics, or any formal promotion track. In an exclusive interview with Fortune, Chen says he just ignored every rule he was given.

“Climbing up was fairly natural and easy,” he tells Fortune, “simply because I just disregarded all the status quo and the rules and realized what’s the right thing to do, and went all the way with it.”

Chen’s not all talk either: When a senior team at HQ blocked his plans to launch Google Partners internationally, Chen launched it anyway—in foreign languages, in foreign markets, without telling anyone in North America. “Once we proved it was extremely successful, then they came and asked us, ‘Oh, can you also launch it in North America?’”

Likewise, getting a promotion was simply a matter of demanding it ahead of schedule. 

Google told him promotions take 2 years—he got his in less than 1

At Google, the general rule of thumb was to wait at least two years before applying for a step up—he says most employees accepted that timeline without question. Chen ignored it entirely, went to his manager within a year, and made the case impossible to refuse.

“I just told my manager, listen, I know this is a year thing. Look what I’ve been able to achieve. It’s way more than anyone else. We’re going to put me up for promotion now.” She did. 

“We have all these rules, we have all these benchmarks, we have all these processes,” Chen says. “That’s what’s going to happen for most of you.”

But for high-achievers, he adds, they’re almost just a formality. Especially when, like him, you’re pulling around 12-hour days and have the results to back up your demands for early progression. “You’re going to be like me, promoted more.” 

“Corporate America can put you in these frames that discourage you,” he adds. But he says the one’s who will be most successful “actually just ignore these and say, “I’m going to do my own thing and take risks, internally.” 

In the end, he took his own career advice literally, opting to become his own boss and do his own thing: With a seven-figure equity package on the table and a career most people would guard with their lives, he handed in his notice—and walked away with zero financial regrets.

Before Google, he was running a thriving business at 15 while in high school

Chen didn’t suddenly wake up one day as a rule‑breaking Google executive. Long before his C‑suite title, he’d already been forced to think like a founder. Growing up in a “low middle-class small town south of Tel Aviv,” his father had a motorbike accident, which left them financially struggling. 

“I used to write code when I was 12, and every year I had to change my computer… the software I used to write was not able to run because it needed more memory,” he recalls. “But he couldn’t afford it.”

So at 15, he went straight to the importers and negotiated for parts so he could upgrade his computer himself. 

“It was my first entrepreneurial adventure,” he adds. “I started selling computers for thousands of different SMBs, throughout my time at high school…  this turned into a very big business.”

His next venture took a different shape entirely. Chen became the digital officer for an LGBT activism nonprofit, building one of the most pioneering advocacy websites across Europe at the time. It was that experience—not a computer science degree, not a corporate internship—that he says caught Google’s eye and landed him his first role there in 2006. “Back then, that was very innovative,” he adds.

Given that background, it’s perhaps less surprising that the golden-ticket job at Google eventually started to feel like a “golden cage.”

When he handed in his notice, his family thought he was “crazy”. His Iraqi-Jewish mother, he recalls, was particularly alarmed—ironically, she inspired the idea for his next venture. 

Financially, he’s worse off as a startup founder—but he has zero regrets

The concept for Tastewise, the AI food and beverage intelligence platform he went on to build, came directly from the family WhatsApp group, where his mom would message every Thursday asking what dietary phase everyone was on before spending a day cooking traditional dishes. 

She saw dinner logistics. He saw a lightbulb moment—and a gap in the market that the world’s biggest food companies hadn’t yet solved: predicting what people actually want to eat before they know it themselves. 

Today, the startup’s technology is used by giants like PepsiCo, Nestlé, Mars, Kraft Heinz, Campbell’s, and Givaudan, and over half its clients are Fortune 100 firms. It has raised more than $71 million in funding.

Financially, he freely admits he’s not ahead of his Google days. “Not yet,” he says. “I’m still building, and I’m all in in the business.”

But given his equity stake, a future Tastewise transaction would likely cement him as a multimillionaire several times over. And he doesn’t waver when asked whether walking away was worth it. “It didn’t matter,” he says of the seven-figure equity he left behind. “It’s almost like it was not a consideration.”

“I used to wake up in the morning, like ‘this is not enough’…. I loved my job. I loved my colleagues. I was extremely happy with my achievements. It was just not mine—not my idea, not my baby. There’s so much satisfaction in creating something out of nothing.”

This story was originally featured on Fortune.com

Last week, Nvidia CEO Jensen Huang made headlines when he told podcaster Lex Fridman that AGI—artificial general intelligence—had already been achieved.

AGI has long been the ultimate goal of many artificial intelligence researchers. That’s been the case even though there is no universally accepted definition of the term. It generally means AI that is as intelligent as humans, but there is a fierce debate over exactly how to define and measure “intelligence.”

In this case, Fridman had offered Huang a very unusual metric for AGI: Could AI start and grow a technology business to the point where it was worth $1 billion? Fridman asked if Huang thought AGI by this definition could be achieved within the next five to 20 years. Huang said he didn’t think that amount of time was necessary. “I think it’s now. I think we’ve achieved AGI,” he said. He then hedged, noting the company didn’t necessarily have to remain that valuable. “You said a billion,” Huang told Fridman, “and you didn’t say forever.”

Few AI researchers agree with the definition of AGI that Fridman offered Huang, which was both more specific (a company worth $1 billion), but also more narrow than most AGI definitions (which tend to refer to matching a vast range of human cognitive skills, not all of which might be needed to build a successful business.) But AI researchers also disagree with one another over what a better definition should be. The term remains stubbornly amorphous despite the fact that several leading AI companies, with collective market valuations of more than $1 trillion, say that AGI is what they are racing towards. Some computer scientists avoid using the term at all precisely because they say it is perpetually undefined and unmeasurable. Others say tech companies like using the term for completely cynical reasons—precisely because it is ill-defined, it’s easy for companies to build hype by claiming big strides towards achieving the fabled milestone. 

The buzz over Huang’s AGI remarks only serves to highlight this quandary at the heart of the AI boom.

Trying to measure AGI

In fact, just days before Fridman dropped his podcast, researchers at Google DeepMind—including DeepMind cofounder Shane Legg, who first helped popularize the term AGI in the early 2000s—published a new research paper that proposed a more scientific way to define and assess whether AI models had achieved general intelligence. The paper, “Measuring Progress Toward AGI: A Cognitive Framework,” draws on decades of research in psychology, neuroscience, and cognitive science to construct what its authors call a “Cognitive Taxonomy.” 

The taxonomy identifies 10 key cognitive faculties—including perception, reasoning, memory, learning, attention, and social cognition—that the researchers argue are essential for general intelligence. The framework then proposes evaluating AI systems across all 10 faculties and comparing their performance to a representative sample of human adults with at least the equivalent of a secondary education.

The paper’s key insight is that today’s AI models have a “jagged” cognitive profile: They may exceed most humans in some areas, like mathematics or factual recall, while dramatically trailing even average people in others, like learning from experience, maintaining long-term memories, or understanding social situations. An AI model would need to at least match median human performance across all 10 areas to be considered AGI, the Google DeepMind researchers suggest.

The researchers also announced a contest with a $200,000 prize pool on the popular machine learning competition site Kaggle for outside researchers to help build evaluations for the five cognitive faculties where existing benchmark tests are weakest.

The DeepMind paper is only the latest in a string of recent attempts to put the measurement of intelligence on more rigorous footing.

Last year, a team led by Dan Hendrycks at the Center for AI Safety, and that included deep learning pioneer Yoshua Bengio, published their own AGI framework and metrics. That paper also divided general intelligence into 10 separate cognitive domains, drawing on a framework for human intelligence developed by three psychologists—Raymond Cattell, John Horn, and John Carroll—that is the most empirically validated model of human cognition. It produced “AGI Scores” for existing AI models; the most capable system tested, OpenAI’s GPT-5, which was released in August 2025, scored just 57%, falling far short of matching a well-educated adult across all the cognitive dimensions.

One of the most ambitious practical attempts to highlight what today’s AI systems still cannot do is the ARC-AGI benchmark, created by well-known machine learning researcher François Chollet. Chollet’s core argument is that intelligence should be measured not by what a system already knows, but by how efficiently it can learn new skills. 

The ARC-AGI benchmark consists of visual puzzle tasks involving grids of colored cells. Each task shows a few examples of an input grid being transformed into an output grid according to a hidden rule, and the test-taker must figure out the rule and apply it to a new input. For a human, grasping the pattern typically takes seconds. For frontier AI models, these puzzles remain surprisingly difficult, because they require the kind of flexible, abstract reasoning—spotting symmetries, understanding spatial relationships, inferring rules from a handful of examples—that current systems struggle with.

This month, Chollet and his collaborators launched ARC-AGI-3, the latest and most demanding version of the benchmark. Unlike earlier editions, which presented static puzzles, ARC-AGI-3 is interactive: AI agents must explore novel environments, acquire goals on the fly, build adaptable world models, and learn continuously over multiple steps—abilities that come naturally to humans but that remain at the frontier of AI research.

Taken together, these new benchmarks represent a growing effort within the AI research community to replace vague definitions about AGI with something closer to scientific measurement. But as these researchers are the first to admit, the difficulty of defining intelligence is as old as the study of thinking itself—and has plagued artificial intelligence as a field from its very earliest days.

Defining intelligence

In 1950, before the term “artificial intelligence” had even been coined and when mathematicians and electrical engineers were just starting to build the first modern computers, the famed British mathematician and computer pioneer Alan Turing wrestled with the fact that it was extremely difficult to formulate a definition of intelligence.

Rather than attempting one, Turing proposed an assessment he called “the Imitation Game,” which later became better known as the Turing Test. It stipulated that a machine should be considered intelligent when it can hold a general conversation with a person, via text, and a second human judge, reading the exchange, cannot reliably determine which participant is the machine and which the human. It was, in essence, an “I’ll know it when I see it” approach to intelligence.

But the Turing Test soon proved problematic too. Eliza, a chatbot developed at MIT in the mid-1960s, was designed to mimic a psychotherapist. Most of its responses followed hard-coded logical rules; Eliza often answered users with questions such as “Why do you think that is?” or “Tell me more” to cover up its weak language understanding. And yet Eliza fooled some people into believing it understood them. Eliza came close to passing the Turing Test even though on almost every other measure it came nowhere close to human cognitive abilities. And, in fact, a more sophisticated chatbot called “Eugene Goostman” officially passed a live Turing Test competition in 2014, again without touching most human cognitive skills.

Today’s large language models converse far more fluently than Eliza ever could, they still cannot match humans across the full spectrum of cognitive abilities—they hallucinate facts, struggle with long-horizon planning, and cannot learn from experience the way a person does.

Compared to the Turing Test, the term “artificial general intelligence” is a relatively recent one. It was first coined in 1997 by Mark Gubrud, then a graduate student at the University of Maryland, who used the neologism in a 1997 paper he presented at a conference on nanotechnology. He used the phrase “advanced artificial general intelligence” to describe AI systems that could “rival or surpass the human brain in complexity and speed, that can acquire, manipulate, and reason with general knowledge, and that are usable in essentially any phase of operations where a human intelligence would otherwise be needed.” But the paper quickly vanished in obscurity.

Then, in the early 2000s, Legg—who would go on to cofound DeepMind—independently coined the same term. He was collaborating with computer scientists Ben Goertzel, Cassio Pennachin, and others on a book about potential ways to create machine learning systems that would be able to address a wide range of problems and tasks. They wanted a term that would distinguish the ambition of these systems from the narrow machine learning algorithms then in vogue, which, once trained, could only tackle a single, narrow task. Goertzel considered calling this more general AI “real AI” or “strong AI,” but Legg suggested “artificial general intelligence” instead, unaware of Gubrud’s earlier usage. He also suggested the term be abbreviated as AGI. This time, AGI took off.

In Goertzel’s book he defined AGI as “AI systems that possess a reasonable degree of self-understanding and autonomous self-control, and have the ability to solve a variety of complex problems in a variety of contexts, and to learn to solve new problems that they didn’t know about at their time of creation.”

The definition was useful for separating work on general AI systems from narrow machine learning ones, but it too contained a fair an unhelpful amount of ambiguity: What did “reasonable degree” mean? Which complex problems in which contexts counted towards the standard?

Legg would later compound this ambiguity by offering a more casual definition of AGI that was in some ways narrower (it didn’t talk about self-understanding, for instance) but equally vague. For instance, he told The Atlantic’s Nick Thompson last year, “I define an AGI to be an artificial agent that can do the kinds of cognitive things that people can typically do. I see this as the natural minimum bar.” But which things? And which people?

Questions like this have continued to swirl around AGI. Does the term mean software that matches the cognitive abilities of an average human? Or the abilities of the humans with the highest IQs? Or the best expert in each individual domain of knowledge? The Hendrycks and Bengio research paper, for instance, defines AGI as matching or exceeding “the cognitive versatility and proficiency of a well-educated adult.” The DeepMind paper proposes measuring against a representative sample of adults. Others have used less precise formulations.

Adding to the confusion, AGI is often conflated in public discussion with a concept AI researchers call “artificial superintelligence,” or ASI—an AI that would be smarter than all humans combined. Most AI researchers consider AGI and ASI to be separate milestones, and very different in degree of sophistication, but in the popular imagination the two frequently blur together.

AGI becomes a corporate goal—and a marketing slogan

If the academic debate over defining AGI has been long and nuanced, the corporate world has introduced definitions that are, to put it charitably, idiosyncratic. DeepMind became the first company to make the pursuit of “artificial general intelligence” a business goal. Legg put the phrase on the front page of the company’s first business plan when he, Demis Hassabis, and Mustafa Suleyman cofounded the company in 2010.

Five years later, OpenAI also made building AGI its explicit mission. Its original 2015 founding principles said that the new lab—at the time a non-profit—was dedicated to ensuring “that artificial general intelligence benefits all of humanity.” Three years later, when the lab first set up a for-profit arm, it published a charter that defined AGI “as highly autonomous systems that outperform humans at most economically valuable work.” Now, for the first time, AGI was being measured by financial metrics, not mere cognitive ones.

And, as it turned out, OpenAI would soon secretly set a highly specific financial threshold for AGI. When Microsoft first invested $1 billion into OpenAI’s for-profit arm in 2019, the tech giant’s agreement with the AI startup made it OpenAI’s preferred commercialization partner for any AI model the lab developed up to, but crucially not including, AGI. At the time, it was reported that the decision of when AGI had been achieved would be at the discretion of OpenAI’s non-profit board.

But, crucially, according to reporting by tech publication The Information in 2024, when Microsoft agreed to invest a further $10 billion into OpenAI in 2023, its contract with OpenAI contained a clause that defined AGI as a technology that could generate at least $100 billion in profits.

OpenAI is nowhere near that mark. The company has reportedly told investors it made $13 billion in revenues last year, but still managed to burn through $8 billion in cash. It does not expect to break even until 2030.

Despite being far short of the financial threshold for AGI in its contract with Microsoft, OpenAI CEO Sam Altman has often made statements that suggest OpenAI is close to achieving the AI milestone as measured by other benchmarks. In a post to his personal blog in January 2025 titled “Reflections,” Altman wrote that OpenAI was “now confident we know how to build AGI as we have traditionally understood it” and that the company was beginning to turn its aim towards superintelligence. In a subsequent essay titled “Three Observations,” he wrote that systems pointing toward AGI were “coming into view.” Yet, at other times, Altman has seemed to acknowledge AGI’s weakness as a concept. Around the same time as his “Reflections” blog post, Altman told a Bloomberg News interviewer that AGI “has become a very sloppy term.”

Microsoft has also chosen to ignore the financial definition of AGI it struck with OpenAI when it suited the company’s marketing purposes. In March 2023, a team of Microsoft researchers published a 154-page paper about GPT-4 provocatively titled “Sparks of Artificial General Intelligence,” arguing the model could “reasonably be viewed as an early (yet still incomplete) version” of AGI.

The paper was widely criticized for hyping the abilities of GPT-4 for commercial purposes. Even Altman distanced himself, calling GPT-4 “still flawed, still limited.”The new research and benchmarks from Google DeepMind and the Hendrycks-Bengio team makes some progress towards establishing a yardstick for AGI, one rooted in decades of study of human intelligence. And what’s clear is that today’s best AI models still don’t measure up to breadth and depth of human cognitive abilities.

Huang, the Nvidia CEO, knows this, just as he was no doubt fully aware of the social media frenzy and headlines he would generate by saying AGI had been achieved. We know Huang knows this because later in the same podcast in which he said “AGI is achieved” he also said that the popular OpenClaw AI agents, which can be powered by any of the top AI models from companies such as Anthropic and OpenAI, could never replicate Nvidia. “Now, the odds of 100,000 of those agents building Nvidia is zero percent,” he said.

Huang is not just Nvidia’s CEO. He is also the company’s founder and the person who has run the company for 33 years, piloting it past near-bankruptcy at one point, to see it now worth more than $4 trillion, making it one of the most valuable companies on the planet. In many ways, Huang is a singular genius. But he’s also a very human one. So maybe we need a new standard, not AGI but AJI—artificial Jensen intelligence. When AI reaches that level, the AI boosters on social media who breathlessly amplified Huang’s AGI claim will really have something to get excited about.

This story was originally featured on Fortune.com