The old robber barons understood that enormous wealth created a social debt. Today’s technology oligarchs increasingly behave as though their wealth has relieved them of one.
Elon Musk built much of his fortune persuading the world that fossil fuels were the past and electric power was the future. Tesla’s mission was to accelerate the transition to sustainable energy. Musk once called burning fossil fuels “the dumbest experiment in history.”
Now he needs more gas.
According to The Wall Street Journal, SpaceX plans to build a natural-gas power plant for a massive new chip-manufacturing complex in Texas. Musk’s AI facilities in Tennessee and Mississippi have relied on dozens of gas turbines. Earlier this year, he personally acquired a company that manufactures gas and diesel turbines. Meanwhile, Tesla quietly removed “sustainable” from its tagline.
There is a practical reality here. Artificial intelligence requires enormous quantities of dependable electricity. Utilities cannot build generating capacity, transmission lines and grid connections at the speed Musk wants to build AI infrastructure. Natural gas can bring power online much faster.
The problem is not that Musk confronted a difficult engineering tradeoff. It is how easily a principle that helped create his fortune became negotiable when it interfered with creating his next one.
That has become a defining characteristic of the technology oligarchs. They do not necessarily lack principles. Their principles are fungible, changing with the political winds, the commercial opportunity and the location of the next dollar. What makes it so grating is that these are not entrepreneurs struggling to survive. They have accumulated more money than they or several generations of their families could reasonably spend. Yet every additional billion seems to justify another abandoned conviction.
We have seen this concentration of wealth before. Rockefeller, Carnegie and J.P. Morgan built monopolies, crushed competitors, fought organized labor and wielded political influence unavailable to ordinary Americans. Their philanthropy did not erase how they made their fortunes.
But the robber barons understood something today’s technology oligarchs seem unable, or unwilling, to grasp. Extraordinary wealth creates a legitimacy problem. It also creates a social debt.
John D. Rockefeller became famous in his later years for handing shiny dimes to children and people he encountered. It was calculated public relations, but the dimes were only the visible symbol of something larger. Rockefeller funded universities, medical research and public-health programs.
Andrew Carnegie argued in “The Gospel of Wealth” that the rich were trustees of their fortunes and had a responsibility to use them for the public good. He financed libraries, universities, museums and cultural institutions. J.P. Morgan supported museums and libraries and used his money and influence to help stabilize the financial system during the Panic of 1907.
None of this made them heroes. Their motives included reputation, legacy and self-interest. But they recognized that enormous power required some public accounting. They attempted to purchase redemption because they understood they needed it.
Today’s technology oligarchs do give money away. Jeff Bezos created the Earth Fund. Mark Zuckerberg and Priscilla Chan established the Chan Zuckerberg Initiative. Musk operates a private foundation. Billions have been committed to legitimate causes.
The difference is that their philanthropy rarely seems to restrain their commercial conduct or define their relationship with the public. It often resembles another founder-controlled venture. The billionaire selects the problem, determines the solution, appoints the people in charge and remains accountable primarily to himself.
Carnegie built libraries that became community institutions. The tech oligarchs build platforms they continue to own.
The disconnect reaches far beyond climate policy.
Marc Andreessen wrote an influential essay titled “It’s Time to Build.” He correctly identified America’s inability to build enough housing in economically successful cities as a major cause of skyrocketing prices. He complained that ordinary people could no longer afford to move to the places where jobs were being created.
Then someone tried to build near his house.
When Atherton, California, considered allowing multifamily housing, Andreessen and his wife registered their “IMMENSE objection.” They warned that the proposed zoning would reduce property values, increase traffic and diminish their quality of life. The Atlantic’s reporting documented the contradiction.
Andreessen believed in abundance until abundance reached the gates of Atherton.
Mark Zuckerberg has demonstrated similar flexibility around free expression. Meta spent years building content-moderation systems and defending its relationships with professional fact-checking organizations. The company presented those programs as necessary responses to misinformation, election interference and harmful content.
After Donald Trump returned to power, Meta eliminated its U.S. fact-checking program, reduced enforcement around politically sensitive speech, elevated Republican policy executive Joel Kaplan, added Trump ally Dana White to its board and donated $1 million to Trump’s inaugural fund. Zuckerberg described the election as a “cultural tipping point” toward free expression. Reuters documented the reversal and its timing.
Perhaps Zuckerberg sincerely concluded that Meta’s moderation policies had gone too far. Companies should be allowed to correct policies that are not working. But Meta did not make the same changes everywhere. In Europe, where regulation made retreat more expensive, it maintained a different posture.
Free expression, it seems, can be calibrated according to regulatory exposure.
Jeff Bezos offers another example. He owns The Washington Post, whose identity rests on holding powerful people and institutions accountable. Immediately before the 2024 presidential election, Bezos stopped the Post from publishing its planned endorsement and ended the newspaper’s tradition of presidential endorsements.
Bezos called the decision principled and argued that endorsements damaged public trust in journalism. He denied any quid pro quo, and there is no proof that he acted to protect his other businesses. But Amazon and Blue Origin have enormous interests before the federal government. A Blue Origin executive met with Trump on the same day the decision became public. Reuters reported the decision, the response and Bezos’s denial.
Bezos created precisely the appearance of compromised independence that the owner of a major newspaper should have anticipated. “Democracy Dies in Darkness” loses some force when the owner can reach over and switch off the light.
The emerging AI industry follows the same pattern. In his 2023 Senate testimony, Sam Altman called for regulation of powerful AI systems. OpenAI later opposed California’s SB 1047, arguing that regulation should come from Washington. In 2026, Altman opposed proposals that would require government approval before companies release new models.
Not every regulation is wise, and Altman can legitimately support some rules while opposing others. But “please regulate us” increasingly appears to mean regulation the industry can help design, not regulation capable of delaying its next product.
Google took a similar path. After employee protests over Project Maven, the company published AI principles restricting work involving weapons and certain surveillance applications. In 2025, Google removed those prohibitions and replaced them with broader language about national security and democratic leadership. The Washington Post documented the change.
Google did not suddenly discover that weapons existed. It discovered that defense AI had become a strategically important market.
The defining image of this new era may be the technology billionaires and CEOs assembled around Trump at his second inauguration. Amazon, Meta and Google each contributed $1 million to the inaugural fund. Amazon provided another million dollars of value by streaming the ceremony. Sam Altman and Tim Cook made personal contributions. Musk, Bezos, Zuckerberg and Google CEO Sundar Pichai received extraordinarily prominent seats. Reuters reported the corporate donations, while the Associated Press captured the scene.
Business leaders must work with elected governments. That is not the objection. The objection is the speed with which people who had criticized Trump, suspended him from their platforms or positioned their companies as guardians of democratic norms reorganized themselves around the new center of power.
They did not merely read the room. They rushed to buy the best seats in it.
This matters beyond whether these individuals are personally likable. They control communications platforms, AI systems, cloud infrastructure, media organizations, satellites, defense technology, data centers and increasingly their own power generation. Their companies have become part of the infrastructure on which modern society depends.
When the public distrusts the people building AI, that distrust transfers to the technology. Workers hearing promises of abundance while watching companies eliminate jobs are not irrational. Communities being asked to accept power plants, water consumption and enormous data centers are not necessarily resisting progress. They are responding to leaders whose commitments appear to last only until they become expensive.
Changing one’s mind is not hypocrisy. Circumstances change, new evidence appears and responsible people revise their positions. The problem is the remarkable consistency of the direction in which these revisions occur. The principle yields to the commercial interest.
The original robber barons understood that money could buy influence, institutions and improved reputations, but that concentrated power still required public legitimacy. Today’s oligarchs appear to believe ownership of the infrastructure makes legitimacy optional.
It does not.
They may own the platforms, the models, the satellites and the data centers. They do not own the public’s consent. One day, probably at the worst possible moment, they may discover the difference.


