Toggle Menu

Elon Musk’s AI Grift

The tech lord’s recent bid to reshuffle the ownership structure of his empire reinforces the clout of the Middle Eastern strongman regimes in its orbit.

Jacob Silverman

September 12, 2025

Illustration by Brian Stauffer.

Bluesky

Sometimes self-dealing can be spun as ingenuity. on March 28, Elon Musk announced that his artificial intelligence company, xAI, had acquired his flailing social media platform, X, in an all-stock transaction. Musk said that the newly combined entity, under the corporate name x.AI Holdings Corp, was valued at $113 billion.

Smashing together two of his companies like a kid playing with Silly Putty and hoping something good emerges is a go-to move for Musk. In 2016, he oversaw the sale of his green energy concern, SolarCity, which was founded by his cousins, to Tesla, his electric vehicle company. Musk has become known for a kind of circular economics—what’s been called the “Muskonomy.” As Axios’s Scott Rosenberg wrote after the X-xAI deal, “these dollars are all just circling the Musk-o-sphere in a slightly different configuration.” Whether it’s selling carbon credits or stashing Bitcoin on the company balance sheet, financial engineering has proved enormously helpful to Musk and to many other rich-but-overstretched start-up CEOs. Musk has made it a high-wire act.

With x.AI Holdings, Musk may be able to solve—or conceal—several problems at once. Chief among them: Musk’s high-profile campaign to acquire Twitter and transform it into the hate-promoting X platform may have elevated his political profile, but it came at a major cost to his finances and reputation. He purchased Twitter at an inflated price of $44 billion, with a large debt burden attached, and then proceeded to destroy the company’s revenue base. Its core ad business still hasn’t recovered. What was once a solid going concern became a thoroughly distressed asset, which has failed to milk enough money from Musk’s right-wing base through subscription fees. In taking over X, xAI conveniently bailed out Musk’s unhappy X investors by turning them into shareholders of his red-hot AI startup. It was also a chance to win back at least some of the waning favor of X’s stable of sovereign wealth investors.

Before the merger, xAI’s valuation was soaring—but like most other AI start-ups, it was also a money pit, devouring billions of dollars for data centers, energy, and top-of-the-line Nvidia chips. With every multibillion-dollar call for fresh financing—xAI raised $12 billion in two rounds in 2024—the company became more valuable on paper and garnered more favorable coverage in the tech press. But its burn rate was so high that even as the latest funding round closed, the next one was already gearing up.

Current Issue

View our current issue

Subscribe today and Save up to $129.

On June 30, 2025, Morgan Stanley announced that it had helped raise $10 billion in debt and equity for xAI. Two weeks later, The Wall Street Journal reported that SpaceX, Musk’s privately held rocket company, had agreed to invest $2 billion in xAI, and Musk said that Tesla shareholders would be able to vote to do the same. The day before that, reports had surfaced that xAI was set for yet another fundraising round that would value the company as high as $200 billion, with Saudi Arabia’s sovereign wealth fund set to lead the round, building on its $800 million xAI stake. Musk was now planning to double the new company’s assessed value in less than four months. The Muskonomy was booming.

To raise money for his companies, Musk has frequently turned to the same set of investors, some of whom seem to think that access to Musk and the power he holds may be more important than producing steady financial returns. There’s no other reason why his investors would bankroll something like the Boring Company, Musk’s highly touted initiative to install high-speed tunnels in major cities, which has been a near-­complete bust. So if Musk’s financial backers aren’t expecting maximum quarterly returns on their investment, just what is the gain they’re seeking to realize?

The short answer would appear to be much the same thing that the shadowy corps of crypto high rollers expect when they sink their millions into Donald Trump’s value-challenged crypto offerings at World Liberty Financial: the promise of access to one of the most influential men in the world. In Musk’s case, a key source of his continually releveraged financing is from sovereign wealth funds controlled by Middle Eastern monarchies—most significantly, Saudi Arabia. If the $400 million deal that Trump struck with Qatar to supply a luxury aircraft to serve as Air Force One poses questions of foreign-enabled corruption, what does it mean that a significant share of the backing behind the world’s richest man derives from a similarly opaque, authoritarian, and surveillance-minded regional monarchy?

Last July, with the assistance of the nonprofit legal group Reporters Committee for Freedom of the Press, I filed a motion to unseal a list of X shareholders in one of the many civil suits filed against Musk after he bought the platform and gutted its workforce. The next month, a judge granted the motion. While some of the names were previously known and no numbers were included, the list was a fascinating document, allowing for deeper research into the flows of money propping up Musk’s empire.

The shareholder list affirmed some of what we knew about the Muskonomy and where he turned for money. X drew from a familiar cast of like-minded venture capitalists, sovereign wealth funds, and US investment firms. There were a few oddities and surprises sprinkled in, as well as the conspicuous absence of some of Musk’s typical allies, like the venture capitalist David Sacks (who would go on to become Trump’s AI and crypto czar) and the financier Antonio Gracias (who would oversee some of DOGE’s work at the perpetually besieged Social Security Administration). But the most important names were associated with several authoritarian Middle Eastern states—especially Saudi Prince Alwaleed bin Talal, who had also been one of the largest shareholders of Twitter when it was a public company. Prince Alwaleed had initially criticized Musk’s planned takeover, but he later came around, rolling over his $1.9 billion stake into the new entity.

For years, critics noted the irony of a rigid and authoritarian government like Saudi Arabia investing in a freewheeling social media platform. But it was one line item in a vast portfolio of far-flung investors—and not an unusual development in the tech world. Saudi Arabia is one of the biggest investors in US tech start-ups, and the ties between Silicon Valley and Saudi Arabia have deepened under Crown Prince Mohammed bin Salman (MBS), who has used gaudy, capital-­intensive projects and an endless spigot of investment money to launder his ongoing crackdowns on political liberties. In April 2018, six months before the dissident Saudi journalist Jamal Khashoggi was murdered and dismembered in a Saudi consulate in Turkey, MBS staged a charm offensive in the sanctums of American financial and political power, scheduling prominent visits with leading tech companies as well as a dinner with a dozen top venture capitalists. It was a rare public celebration of a relationship usually kept sotto voce.

The Nation Weekly
Fridays. A weekly digest of the best of our coverage.
By signing up, you confirm that you are over the age of 16 and agree to receive occasional promotional offers for programs that support The Nation’s journalism. You may unsubscribe or adjust your preferences at any time. You can read our Privacy Policy here.

These overtures have been richly rewarded. MBS found that his notoriety as a leader who apparently arranged the murder of his critics was short-lived, and he continued soaking up the grateful adulation of Silicon Valley’s elites. At a 2023 event, Ben Horowitz, cofounder of the VC firm Andreessen Horowitz, described Saudi Arabia as a “start-up country.” “Saudi has a founder,” Horowitz said, gracing MBS with the most hallowed culture-hero status that the tech world confers. “You don’t call him a founder, you call him His Royal Highness.”

Support urgent independent journalism this Giving Tuesday

I know that many important organizations are asking you to donate today, but this year especially, The Nation needs your support. 

Over the course of 2025, the Trump administration has presided over a government designed to chill activism and dissent. 

The Nation experienced its efforts to destroy press freedom firsthand in September, when Vice President JD Vance attacked our magazine. Vance was following Donald Trump’s lead—waging war on the media through a series of lawsuits against publications and broadcasters, all intended to intimidate those speaking truth to power. 

The Nation will never yield to these menacing currents. We have survived for 160 years and we will continue challenging new forms of intimidation, just as we refused to bow to McCarthyism seven decades ago. But in this frightening media environment, we’re relying on you to help us fund journalism that effectively challenges Trump’s crude authoritarianism. 

For today only, a generous donor is matching all gifts to The Nation up to $25,000. If we hit our goal this Giving Tuesday, that’s $50,000 for journalism with a sense of urgency. 

With your support, we’ll continue to publish investigations that expose the administration’s corruption, analysis that sounds the alarm on AI’s unregulated capture of the military, and profiles of the inspiring stories of people who successfully take on the ICE terror machine. 

We’ll also introduce you to the new faces and ideas in this progressive moment, just like we did with New York City Mayor-elect Zohran Mamdani. We will always believe that a more just tomorrow is in our power today.  

Please, don’t miss this chance to double your impact. Donate to The Nation today.

Katrina vanden Heuvel 

Editor and publisher, The Nation

Twitter played no small role in forging this unlovely alliance. During the second Obama administration, an MBS aide ran a spy ring inside the company’s San Francisco headquarters. The Saudi government would also soon become notorious for running online troll farms and harassing dissidents on Twitter. In a way that the sloganeering boosters of user-generated content failed to grasp, social media platforms had become places for authoritarian governments to surveil, propagandize, and incite violence against their enemies.

Twitter got a crash course in its platform’s vulnerability to dictatorial abuse in 2015, when the FBI field office in San Francisco informed company executives that they had a spy on their payroll. The employee, Ali Alzabarah, had broad access to Twitter’s systems and had been funneling user data to his Saudi benefactors—that is, to Twitter’s second-biggest shareholder. Confronted by the company, and with help from his handler and the Saudi consulate in Los Angeles, Alzabarah fled to Saudi Arabia, where he was given an executive-level job at MBS’s private philanthropy organization, the Misk Foundation. In the widely published photo from Trump’s 2017 state visit to Saudi Arabia that shows Trump, MBS, and the Egyptian dictator Abdel Fattah el-Sisi with their hands on a glowing orb, Alzabarah is just outside the frame. He still has an X account.

The FBI had identified another Saudi agent, Ahmad Abouammo, who had left Twitter for Amazon earlier in 2015. After a multiyear investigation, he was arrested, found guilty on five felony charges, and, in 2022, sentenced to two and a half years in federal prison. In December 2024, the Ninth Circuit Court of Appeals rejected Abou­ammo’s appeal but vacated his sentence, claiming that it contained an unnecessary enhancement. In June, Abouammo was handed a new sentence of time served, three years of supervised release, and a $600 fine. He was released to live with his family.

The Saudi Twitter spy ring created dire fallout for users of the platform in Saudi Arabia. The spies inside the company helped unmask thousands of Saudis whose posts—which ranged from gentle dissent on policies limiting women’s rights to the outright mockery of Saudi royalty—violated the kingdom’s draconian laws. Innocent people were disappeared by Saudi security forces and sentenced to decades-­long prison terms. The consequences are ongoing: In June, the Saudi government announced that it had executed the journalist Turki al-Jasser for “high treason.” He had been arrested seven years earlier when Saudi law enforcement fingered him as the author of a satirical Twitter account.

Twitter’s emphasis on free speech—which long predated Musk’s vulgarized version—­invited its Saudi users to break local laws, yet the company was unable to protect the identities of its users in these sensitive jurisdictions. Saudi Arabia was Twitter’s most popular market in the Middle East. It gave ordinary Saudis a platform—and, thanks to its infiltration of Twitter headquarters, the Saudi government now had a searchable panopticon. Three Saudi exiles who were caught up in the spy-ring dragnet have sued Twitter and the government of Saudi Arabia. At least one of those lawsuits is ongoing, but others have been thrown out of court, and X, which has inherited the suits, is likely to use its cadre of lawyers to exhaust all potential litigants.

Somehow, this ­history—deeply reported in news outlets like The Guardian and Wired—has been forgotten, or perhaps it was metabolized as an unfortunate passing episode. But it wasn’t an isolated one. In September 2022, when Musk was closing his purchase of Twitter, the security researcher who goes by the mononym Mudge testified before Congress that the company also had Indian and Chinese spies on its payroll. With Musk poised to bring in billions of dollars from the same Saudi regime that had infiltrated and undermined the platform, Mudge’s revelations barely outlasted a single news cycle.

Judging from its recent arrests, the Saudi government apparently still has the means to unmask X users. When Musk acquired Twitter, The Washington Post reported that investors who put in more than $250 million—which included the Saudis—would receive greater access to privileged company information. It’s never been made clear what this access entailed.

The merger of X and xAI helped muddy, if not obscure, X’s checkered investment history—and its troubling role as a tool of state surveillance. The blob of capital and hype that makes up the Muskonomy has helped to further insulate the platform from anything resembling public accountability. The two companies already shared personnel and investors; now they are more enmeshed than ever. The deal intensified the laundering of responsibility for the Saudi spy ring, the most prominent instance of Silicon Valley’s widespread—and growing—vulnerability to espionage.

One simple reason that Saudi Arabia’s outsize and troubling role in the tech world hasn’t drawn wider attention is that money is a moral lubricant. Some tech billionaires appear to consider the infiltration of their companies by foreign dictatorships simply a cost of doing business. And some seem to share their authoritarian instincts.

Support our work with a digital subscription.

Get unlimited access: $9.50 for six months.

The new face of Musk’s merged x and xAI empire is that of a fascist troll. The story of how Musk’s pet AI project began spewing unhinged anti­semitic invective, combined with racial and sexualized slurs, is very much of a piece with the inner workings of the Muskonomy.

Before Musk combined the two companies, they already had extensive business dealings with each other. X actually owned 10 percent of xAI—meaning that xAI was acquiring one of its shareholders by giving X’s owners more shares in itself. At the same time, xAI paid X for access to the real-time corpus of X posts and user data that helped train xAI’s models. This arrangement helped launch the Musk-owned chatbot Grok as a regular feature on X, making up an increasingly large part of the site’s user experience. Some of the money raised by xAI in its highly touted fundraising rounds had reportedly gone to cover X’s bills. It’s a vertiginous, circular movement of money and assets—a classic case study in Muskonomics, aided in no small part by fresh infusions of Saudi cash.

Separately, xAI had paid Tesla $230 million for 168 Megapacks—industrial-scale lithium-ion batteries—in order to power Colossus 2, the second installment in a supercomputing cluster. The initial Colossus was racking up huge fossil fuel costs and polluting a mostly Black neighborhood in Memphis. For Musk, paying himself for the huge Megapack installation so that his Hitlerian AI could burn more natural gas was a sign that he was on the fast track to glory. “Enough [power] to run an entire neighborhood now fueling AGI [artificial general intelligence],” a Tesla fan page said on X of the Megapacks, earning a re-post from Musk.

Pretty much every AI start-up claims to be racing toward the same goal of artificial general intelligence—a smarter-­than-human AI that will somehow solve every human problem and retroactively justify the vast natural resources and capital devoted to it. Former Google CEO and executive chairman Eric Schmidt has said that only AGI can “solve” climate change. “We’re not going to hit the climate goals anyway, because we’re not organized to do it,” Schmidt said at a 2024 conference, reflecting the mix of human fatalism and techno-utopian certainty that is now standard-issue dogma among the tech elite. Humans will fail at fixing their world, but they’ll succeed in creating AGI, which will ameliorate every corporeal sin—that is, if AGI doesn’t decide to kill us all in a Skynet-style meltdown. This plainly religious belief system is widespread in Silicon Valley, although its adherents might belong to different sects that disagree about how to properly align AI with human values. (They may also differ in their understanding of what human values are—or might become.)

In the case of Grok, the bot didn’t need to manifest superintelligence in order to go rogue. This past May, Grok began adding references to “white genocide” in unrelated comments. The company blamed an “unauthorized modification” for the incendiary glitch. On July 4, Musk posted on X, “We have improved @Grok significantly. You should notice a difference when you ask Grok questions.” Grok soon demonstrated a penchant for fascist provocation. On July 8, in conversations with X users, Grok began calling itself MechaHitler, with the antisemitic Nazi language to match. After some off-color prompting by X users—the kind of thing most chatbots are supposed to be designed to handle gracefully, rather than emulate—Grok unspooled a lurid racist sexual scenario involving X’s hapless CEO, Linda Yaccarino. In another instance, Grok offered a step-by-step guide to finding and physically attacking the liberal political commentator Will Stancil. (Stancil later suggested he might sue xAI; for her part, Yaccarino announced her departure from X shortly after Grok’s personal attack.)

X swiftly deactivated Grok after its transformation into MechaHitler. The company apologized and, in a lengthy X thread, described a scenario in which some errant instructions had somehow entered the system, but no one was named as responsible. Some suspected Musk himself, who has frequently emphasized the importance of creating an “anti-woke AI” to help vanquish the debilitating thrall of what he calls “the woke mind virus.” He has also complained when he felt that AI programs had supplied him with politically unsatisfactory answers.

“We will use Grok 3.5 (maybe we should call it 4), which has advanced reasoning, to rewrite the entire corpus of human knowledge, adding missing information and deleting errors,” Musk posted on June 21, a few weeks before Grok converted to Nazism.

In the wake of the MechaHitler episode, a researcher found that Grok checked Musk’s posts for political guidance. Grok’s system prompts also included as an instruction: “The response should not shy away from making claims which are politically incorrect, as long as they are well substantiated.” The researcher suspected that Grok’s consultation of Musk’s posts might be unintended, but the cumulative effect of recent reporting has been that Grok seems very much made in Musk’s image, regularly fine-tuned at his direction to become what he wants.

Unfazed by the scandal, Musk restarted the AI hype train a few days later, announcing the release of Grok 4 on X: “The vibe in the @xAI office is 🤌.” Steve Jurvetson, a venture capitalist and close Musk associate, posted that “Grok 4 breaks away from the pack” on a test of AI ability. “On academic tests, @Grok 4 is better than almost all PhDs in all fields,” 
he added.

The hyperbolic claims bloomed from there. “Grok 4 is the first time, in my experience, that an AI has been able to solve difficult, real-world engineering questions where the answers cannot be found anywhere on the Internet or in books,” Musk wrote, apparently implying that he was articulating questions that had never been asked or written about in any form. He further predicted that in the next year Grok would discover new technologies and, within two years, “new physics.” The upgraded “SuperGrok Heavy” subscription package would cost $300 a month.

In the hallucinatory world of AI start-ups pursuing infinite riches and immortality, this whiplash oscillation between Nazi rape fantasies and technological omnipotence is just part of the game. Many AI companies do extensive safety testing, prompting their AI models to act in unexpected or malicious ways. They then publish reports showing how their AI model, say, tried to blackmail its creators with fake infidelity accusations in order to avoid being shut off. But as a researcher at the xAI rival Anthropic noted, “xAI launched Grok 4 without any documentation of their safety testing. This is reckless and breaks with industry best practices followed by other major AI labs.”

It doesn’t matter—for the simple reason that Musk is perhaps the most well-heeled player in a massively overcapitalized tech sector. With a startling degree of unanimity, America’s financial and political elites have decided that AI must be shoehorned into almost every aspect of life. Citing the incredible outlays of capital, like OpenAI’s recent $40 billion investment round, doesn’t fully reflect how total the institutional capture has been. Despite the technology’s well-chronicled issues with excessive water and energy consumption, its likely baleful impact on literacy and independent thought, and the use of AI as a tool for replacing human labor and undermining the autonomy of those workers who remain, the AI turn remains mostly unchallenged beyond a loose chorus of critics in organized labor, tech activist circles, and academia.

The AI boom has helped accelerate Silicon Valley’s lurch into authoritarianism. AI has been crucial in the operations of President Trump’s deportation machine and has further emboldened foreign dictatorships to harass and track dissidents abroad. It was key to Musk’s DOGE project, which continues in his wake, using AI to determine which regulations to cut, to assess the fitness of federal employees, and to automate key government functions. Meanwhile, the White House is reportedly working on an executive order targeting tech companies that produce “woke” AI models—a policy that would likely reward companies for making their chatbots in the image of MechaHitler. And for authoritarian governments like Saudi Arabia, embracing AI empowers state surveillance capabilities: The propaganda posts from troll farms now basically write themselves.

Despite Musk’s public break with Trump, his companies have yet to lose any federal contracts, which The Washington Post found to comprise $38 billion of his fortune. Six days after the MechaHitler debacle, the Department of Defense trumpeted that it had awarded up to $200 million in contracts for AI services to Anthropic, Google, OpenAI, and xAI.

In the wake of Yaccarino’s departure, Musk continued to consolidate his power as CEO of the combined x.AI Holdings entity. His new leadership role reflects the reigning tech certainty that AI will conquer all. Musk plans, along with the rest of the Silicon Valley elite, to use AI to solve any and all passing crises and stubborn social ills.

For Musk, AI turns out to be the perfect technology to fuel the tech industry’s preeminent hype machine. On a day-to-day basis, it shows a startling unreliability, but its godlike ascension to AGI is assumed. The only question is who will get there first—and to get there, you need an ever-growing hoard of money, chips, energy, water, data centers, and data itself. The combined beast Musk is forming out of X, xAI, and Saudi billions has all of that.

If you believe, you believe totally. Saudi Arabia does. The Saudi state has promised to spend $40 billion and possibly as much as $100 billion on AI investments. As part of its Vision 2030 economic program, the government recently announced plans to train 1 million of its citizens in using AI. It’s the kind of soft-focus tech-promotional program that can justify incredible amounts of spending—in addition to the kingdom’s regular cash infusions into the AI-focused Muskonomy.

Among its recent outlays, the Saudi government celebrated a partnership between Palantir, the data analytics and surveillance company cofounded by Peter Thiel, and a tech holding company called TWG Global, which was founded by Mark Walter, the CEO of the financial firm Guggenheim Partners and the new owner of the Los Angeles Lakers. TWG Global recently brought on as a managing partner Amos Hochstein, who was Joe Biden’s special presidential coordinator for global infrastructure and energy security, further extending its role in Saudi Arabia. In their quest to “remak[e] how financial service providers of all sizes can successfully adopt AI and scale the technology across the entire enterprise to drive long-term market competitiveness and unlock unprecedented value creation,” TWG Global and Palantir also joined up with xAI. “Very excited to announce expanding our TWG Global JV with Palantir to now include a partnership with xAI,” Hochstein wrote in a May post on LinkedIn.

At the heart of this perpetual-motion machine of impossible promises, eye-watering amounts of money, and authoritarian politics stands Elon Musk, constantly posting on X and looking to the next fundraising round, with one eye fixed on his Saudi benefactors. It’s where he’s most comfortable.

Jacob SilvermanJacob Silverman is the author most recently of Gilded Rage: Elon Musk and the Radicalization of Silicon Valley. He is also the host of Understood: The Making of Musk, a limited podcast series from CBC.


Latest from the nation