Notice the conversation around AI is sounding a bit predictable?
Workers are told to adapt, reskill, and become “AI-ready” on the double, while business leaders, celebrity entrepreneurs, productivity gurus, and online creators (often with a brand deal somewhere in the background) preach that embracing AI is a personal duty, if not a moral responsibility. Get on the AI train, and you can look forward to a future of less drudgery, more creativity, and maybe even shorter work hours. Miss it, and good luck to you.
But viewed through the lens of shareholder value ideology – the flawed corporate governance system that took over America in the 1980s — the story looks rather different. You see that for the last forty years, every time productivity rises, wealthy executives and other shareholders with massive stock holdings are buying megayachts while workers are told to be grateful they still have jobs.
AI is being launched inside a machine built to send wealth upward. Why would it suddenly start working differently now?
Let’s survey the scene as the AI money starts rolling in.
Consider Microsoft. The exact mix changes from year to year, but whenever the company makes money, you can count on a big pile of it going out the door to shareholders and top execs.
Between 2006 and 2025, Microsoft spent $283 billion on stock buybacks and another $223 billion on dividends — a total equal to 71% of profits. And even as the company bets big on AI, it’s sending enormous sums to shareholders. This is happening while it lays off thousands of workers. The people who power the company are either fired or left wondering whether they’ll get a paycheck next month.
Microsoft CEO Satya Nadella doesn’t have that problem.
In 2025, he pocketed about $123 million – and that’s more than double what he took home the year before. That breathtaking figure comes from William Lazonick and his colleague Matt Hopkins, who count the money executives actually cash in rather than the estimated value of new stock awards, which the media usually report and thus understate the true scale of executive compensation. According to their calculations, Nadella, in his 13 years as CEO, has pocketed nearly $1.2 billion — an average of $91 million a year. About 90% comes from stock-based pay. In 2025 alone, Microsoft’s board awarded him more than $84 million in new stock grants — the largest of his tenure. Those shares won’t all turn into cash right away, but if Nadella can keep Microsoft’s stock price climbing — with help from tens of billions of dollars in stock buybacks — they could eventually put vastly more money in his own bulging bank account.
Nadella’s jackpot-size pay package is the kind that gives other executives every reason to keep Wall Street happy, even if it means leaving workers high and dry.
The future of work looks great – if you’re holding vast amounts of stock awards.
Consider Nvidia, one of the biggest winners of the AI boom. Demand for its chips has thrown off a flood of cash, which gets funneled into research, next-gen AI systems, and hiring engineers at ginormous salaries. Sounds good! But shareholders want to get paid now.
Hence the company’s tens of billions in stock buybacks — with much more to come — even as it competes in one of the most capital-hungry races in tech history. Lazonick reminded me that Nvidia only began doing large-scale buybacks recently as its profits soared because Jensen Huang, who founded the company and remains its CEO, only has about 3% of the stock. “He has to keep hedge fund activists happy or he will lose his job,” notes Lazonick.
That pressure spells potential trouble. Nvidia is trying to build the future while feeding the shareholder value machine, and the machine always wants more.
And who are these hedge fund activists pressuring Huang? They are powerful shareholders who snap up shares in a company and then demand its leaders to make changes they think will juice the stock price. Maybe they push for new executives, or cost-cutting, or bigger payouts to shareholders. They may not run the company directly, or have a clue about how to run it better, but holding stock and convincing other stockholders to back them gives them a loud voice in what happens. These are not “investors” Lazonick emphasizes, but “predatory value extractors who are permitted to steal the value that the real investors – the workers — have created.”
For now, Nvidia execs think they can afford to pour money into the AI race and buy back shares. But Intel serves as a cautionary tale: a company that spent heavily rewarding shareholders while its once-dominant technology advantage decomposed. That system hasn’t gone away.
Even Nvidia’s well-paid engineers can’t hide from shareholder-value logic; they’re just higher up the expense hierarchy. If push comes to shove and margins tighten, they may find themselves on the wrong end of hiring freezes, project cuts, or the dreaded “strategic restructuring.” Yet somehow there’s always room for another buyback.
Executive pay runs on a different trajectory. As Lazonick told me, stock-based compensation has a “ratchet effect” whereby each stupidly large package resets the benchmark for the next. What once looked obscene becomes routine, and over time, billionaire-scale pay is just the going rate.
Companies that haven’t yet gone public aren’t outside this world either. Firms like OpenAI and Anthropic may not do buybacks – yet. But they already face the demands of the shareholder value machine: raise money, chase stratospheric valuations, and keep the people buying the shares convinced that the crazy burn rate is justified. The money pouring into AI may be called “investment,” but Lazonick points out that much of it is really just a bet on who will make money from the speculative boom, and who will be the “greatest fool” left holding the stock when its price plummets from hysteria-driven heights. Do the speculators and manipulators really care if the company is ever productive and profitable? Actually, no.
Here’s a bet: If this perverted system persists, AI won’t pay off for most American workers.
Elon Musk’s new trillionaire status is a vivid example of shareholder value logic at work. When Tesla and SpaceX became valuable companies, guess who captured the gains?
In Musk’s case, the answer starts with his huge equity stake. First, he acquired an enormous number of Tesla shares through compensation packages, which he claimed were necessary to protect the company from people who might try to take it over. Then, when Tesla successfully mass-produced EVs at a moment when consumers wanted them, the surge in the company’s market valuation translated all the shares that Musk had grabbed for himself into mind-blowing personal wealth.
“Tesla’s big shareholders got richer through rising share prices,” explained Lazonick, “even though the company’s productive transformation was driven by engineers, factory workers, suppliers, and the people who designed, built, and scaled Tesla’s vehicles.”
SpaceX shows the same dynamic. Musk maintained a huge equity position, helped establish a private-market valuation at sky-high levels, and accumulated even more paper wealth as that valuation climbed. Now that SpaceX has gone public, many shareholders will likely be focused less on building the company’s productive capabilities than on how high the stock will go and when they can cash out.
Why focus on future innovation when you can get rich squeezing the juice out of the last one?
As Lazonick put it, “Musk has taken the Wall Street playbook and perfected the game — you have a founder with a massive equity stake, genuine innovation, and then traders chasing the stock who can turn technological progress into yet another machine for concentrating wealth.”
The games are getting wilder. Shareholders may leave or stay, but that uncertainty is driving the whole Silicon Valley system. Companies spread hype about what AI will become, and the hype can be worth billions before a single dollar of profit appears.
Economist Servaas Storm argues that AI-promoting CEOs are making “unhinged” and conveniently unverifiable promises about a future of either no work at all or endless abundance and productivity. Meanwhile, some of the biggest AI players are effectively just moving money around in a circle, building a house of cards on the promise that the technology will eventually justify the hype. Microsoft, for example, provides the cloud infrastructure that OpenAI runs on while also committing to buy OpenAI’s products. So some of the demand is baked in from the start: money flows from Microsoft to OpenAI and then back again as OpenAI spends heavily on Microsoft’s services. Both companies can point to growth, but much of it is driven by intertwined deals rather than a fully independent market.
If the real revenue doesn’t show up, it probably won’t be Sam Altman or Satya Nadella who’s exposed.
Workers, on the other hand, are learning in real time how AI is impacting their livelihoods: headcount reduction, disappearing entry-level employment, Orwellian surveillance, algorithmic scheduling, and what Cory Doctorow has aptly termed an all-around “enshittification” of jobs.
How shitty is it getting? Let us count the ways.
Amazon has expanded warehouse and customer-service automation alongside aggressive performance tracking. IBM has openly linked AI adoption to back-office job cuts. Microsoft and Google are slapping AI onto everything they sell, promising a productivity revolution while figuring out how many workers they can dump behind the scenes. Workers at Meta taking leave for pregnancy and disability accuse the company of using AI to target them for layoffs. At Google, employees are demanding layoff protections as the company trims its workforce while spending billions to build out AI.
AI can spit out reports and code in seconds, but somehow, the productivity gains don’t seem to be translating into more free time for workers – often quite the opposite. The boss says, “here are some fancy new tools — now work faster and harder. And don’t look for a raise.”
You might think of AI as the ultimate shareholder value technology: a machine that promises to boost output while further curbing the power of labor, no matter what the stock market is doing. As Lazonick told me, “When the stock prices come down, they lay off workers. They squeeze their pay even more, and they give out massive amounts of stock to themselves. Then, when the stock goes up again, they make massive amounts of money, disproportionately capturing the upside.”
He warns that putting AI development inside this system could amplify inequality on a scale few technologies have ever produced.
“What it’s doing is making a small number of people rich, and a relatively smaller group, super rich, all around the value of the stock,” he explained. “And these people who are getting super rich feel so entitled to it because of how deeply embedded the ideology of shareholder value is and how unchallenged it is.”
Until shareholder value doctrine is challenged, there’s little reason to expect AI, despite its legitimate uses, to truly benefit the people doing most of the work. The gains will be privatized, and the disruptions socialized, just as they’ve been for decades.
The Shareholder Era Was Built, and it Can Be Demolished.
So how do we change this story?
The first step is to be clear that shareholder value system is a fairly recent development.
For much of the postwar era, Americans assumed corporations owed something back to the society that granted them the privilege to exist. They were expected to make useful products, provide stable, long-term jobs, invest for the future, and share prosperity with the workers, communities and taxpayers that made their success possible. As Lazonick likes to point out, Americans once believed public corporations owed shareholders remarkably little. That changed in the 1980s, but there’s no reason to assume that it can’t change again.
You could see the difference the earlier mindset made in places like Detroit’s auto plants and the steel mills of the Midwest. When workers produced more, they got higher paychecks, pensions, and a genuine, if imperfect, sense that each generation could do a little better than the last. Unions played a critical role, and so did corporate norms that placed more emphasis on investment, workforce development, and sharing the gains. The public didn’t expect nearly all of the benefits of economic growth to pool at the top. Why should we accept that now?
As Lazonick told me: “Shareholder value was constrained by the power of unions and managers who knew they needed the cooperation of unionized workers to create value. Even though there were shareholder-elected boards, norms were such that trying to squeeze workers the way they are squeezed today wasn’t tolerated.”
Good news: there are signs that norms may be on their way to shifting. Unions are more popular than they’ve been since the 50s, and more Americans are fed up with corporations having outsized power in our society.
The second step is busting through myths that have become so familiar they take on the ring of facts.
Myth #1: Corporate law requires executives to maximize shareholder value. Wrong: it doesn’t. As legal scholar Lynn Stout and others have shown, corporate law gives managers plenty of room to do something other than focus on the stock price. “If managers had a legal duty to maximize shareholder value, they must have been breaking the law for most of the 20th century,” Lazonick wryly notes. “The post-WWII growth of the US economy to a position of global leadership and upward socioeconomic mobility for most Americans must be evidence of their crimes.”
Myth #2: Shareholder value is simply good business. The evidence of decades suggests otherwise. Buybacks, layoffs, and financial engineering may produce fast gains for people holding lots of stock, but they can also make companies weaker over time. A system built around extracting value for shareholders leaves workers less secure, and the broader economy less capable of turning productivity gains into prosperity for everyone. During the shareholder value era, U.S. based firms have declined relative to foreign competitors in a growing range of critical technologies.
Myth #3: It’s just the way capitalism works. Wrong again. Maximizing shareholder value is the product of specific changes in corporate governance, financial markets, and public policy. Americans didn’t accept this idea because they debated it or even understood it. Rather, it got embedded through laws, incentives, executive pay, financial markets, and business schools. All this can be challenged.
Even many of its champions acknowledge that the system isn’t written in stone. In 2019, the Business Roundtable – basically the nation’s CEO class speaking in unison — abruptly declared that corporations should serve workers, customers, suppliers, and communities, not just shareholders. The timing wasn’t accidental. Elizabeth Warren’s proposed Accountable Capitalism Act threatened to take on the shareholder value system by requiring large corporations to consider stakeholder interests and give workers representation on corporate boards. The Roundtable’s statement might have been convenient PR, but it revealed that corporate leaders understood that the political winds could shift.
The third thing to recognize is that individual workers can’t solve the AI challenge on their own.
Workers are constantly told to reskill, embrace AI, and become more productive. But, as Thomas Ferguson put it, “What we have is a distribution problem, and no amount of reskilling fixes a system where productivity gains are pre-assigned to shareholders.”
That observation changes how we should think about the technology. If AI doubles productivity, we need to be asking if that translates into better wages, shorter workweeks, greater job security, and more investment in people. Or will does simply produce another surge in stock buybacks, ridiculous executive compensation, and more megayachts?
The answer depends on the rules governing corporations, which is why the task is to change the incentives and rules that have redirected corporate wealth upward for the last four decades.
Some ideas to consider:
- End the buyback machine. Lazonick argues stock buybacks should once again be treated as market manipulation, as they largely were before the SEC’s 1982 Rule 10b-18 opened the floodgates. They should be banned altogether, not merely taxed.
- Rethink executive pay. It’s time to reward executives for building stronger companies and expanding productive capabilities instead of just jacking up stock prices. Compensation should reflect investment in workers, innovation, productive capacity, and the creation of durable value, rather than payouts for shareholders and executives. Nobody should get gobs of cash just for one or two years worth of results. Payouts need to be for long term success.
- Change who gets a seat at the table. Workers and taxpayers bear enormous risks when companies succeed or fail. They deserve meaningful representation on corporate boards alongside shareholders who are concerned with building productive companies.
- Rebuild worker power. AI should not become another tool for surveillance, deskilling, or union avoidance. Workers need stronger bargaining rights and a real voice in how new technologies are deployed.
- Don’t mistake an AI dividend for economic reform. Proposals like Bernie Sanders’ for a sovereign AI fund do something good in recognizing that the public should benefit from the enormous wealth AI may create. But a dividend alone risks treating the symptoms without getting at the disease. It lets companies and politicians capturing – and stealing — the vast majority of gains keep on with business as usual. A $1,000 check in your mailbox may provide a little relief, but remember, that’s a tiny drop in an ocean of wealth that’s already been concentrated at the top. What we need is to make sure the people creating the wealth share in it from the start.
- Invest in people’s ability to use AI. Retraining programs are important, but it’s much more important to make sure people have the education, skills, healthcare, security, and opportunities required to use new technologies to make their lives better. That means investing in workers and strengthening the schools, healthcare systems, and public institutions that allow everyone to share in the benefits of AI-assisted innovation that improves the quality of life.
- Limit corporations’ spending on politics and lobbying. Getting the big money out of politics has to be a priority. AI companies shouldn’t be able to use their wealth to write the rules that govern them. That may ultimately require a constitutional amendment to undo court decisions that give corporations broad political spending rights. Corporations are legal creations, not people. They exist because society grants them special privileges, so they should serve the public interest — not use their wealth to drown out the voices of the people they affect.
Notice that none of these proposals asks companies to stop making profits or make the government the real owner. They simply ask that corporations once again balance the interests of the people who actually create value instead of focusing solely on those who happen to hold shares.
Politically, none of this is easy. But neither was winning the eight-hour day, legalizing collective bargaining, or building the social protections that many now take for granted.
Norms change when people figure out that the story they’re told doesn’t quite square with reality. For decades, Americans were assured that corporations putting shareholders first would make everyone better off. But millions have lived through the results of that experiment — more wealth at the top while their own economic lives barely budged or crumbled. Eventually, ideas once dismissed as radical can start looking like much-needed corrections.
If a company promises to treat its workers well, that’s a nod toward corporate social responsibility. But what we really want to hear is companies supporting rules that raise the floor for everyone: stronger collective-bargaining rights and worker representation in corporate governance; real limits on workplace AI surveillance and bossware; changes to the tax and corporate rules governing buybacks and executive compensation; tougher antitrust enforcement; more public investment in worker transition and education; and rules to keep corporations from using their political spending to undermine the very reforms meant to hold them accountable. That’s what transformation looks like.
We don’t have to let the AI age become another great wealth transfer upward. Properly governed, corporations can once again serve the people who make their success possible and help build prosperity for all Americans. Other successful economies already do this. So did America not all that long ago.
We created corporations to serve the public. Somewhere along the way, we forgot who they’re supposed to be working for: us.
