What It Will Take to Change the Regime in Iran
The U.S. Military Must Go Big—and Then Let Iranians Do the Rest
By Behnam Ben Taleblu
Foreign Affairs — February 27, 2026
Behnam Ben Taleblu is Senior Director of the Iran Program and a Senior Fellow at the Foundation for Defense of Democracies.
Key Takeaways
• The Islamic Republic is at its weakest point since 1979, but limited strikes or decapitation operations (the “Venezuela model”) will not bring it down—the regime is too institutionalized and would close ranks rather than fracture.
• The U.S. should launch a comprehensive air campaign targeting Iran’s air defenses, missile arsenal, nuclear sites, political institutions, and the full chain of command of its security forces (IRGC, Basij, police) across all provinces.
• The goal is not a U.S. ground invasion but to degrade the regime’s repressive apparatus enough for the Iranian people—who have already demonstrated their willingness to sacrifice—to finish the job and reclaim their country.
The Islamic Republic of Iran is, quite possibly, at its weakest point since its founding, in 1979. In June, Israeli and U.S. attacks destroyed its uranium enrichment capacity and many of its air defense systems. In December and January, the country experienced the most widespread domestic uprising since the birth of the Islamic Republic. Throughout, it has faced spiraling economic and environmental crises that it cannot fix. None of these events has knocked out the Islamic Republic. But there is no doubt it is down.
Now, U.S. President Donald Trump is threatening to attack the country. He has made it clear he has little tolerance for the regime’s efforts to rebuild its nuclear program or the extraordinarily brutal way it cracked down on protests. “If Iran violently kills peaceful protesters, which is their custom, the United States of America will come to their rescue,” he said last month. “We are locked and loaded and ready to go.” The president has since amassed U.S. air and naval assets in the region, and he is considering a variety of strike options.
But that doesn’t mean major attacks are guaranteed. In fact, thus far, the administration’s decisions raise more questions than they answer about what Washington aims to achieve and how. Right now, Trump is practicing gunboat diplomacy, hoping that the threat of force will coerce the Islamic Republic into making a nuclear deal better than the one he left in 2018. If that fails, he is mulling decapitation operations or limited strikes in order to get the regime to bend.
It is easy to see why the Trump administration is prioritizing diplomacy and limited strikes. The Islamic Republic may be weak, but it is still lethal and capable of harming U.S. forces and civilian targets throughout its region. The president, meanwhile, has repeatedly proved reluctant to start a protracted military campaign. But the reality is that after decades of trying and failing to change Tehran’s behavior with sanctions, sabotage, and, more recently, one-off strikes, the time has come to go big. The regime is simply too ideological to be cowed by a few rounds of bombing. The Iranian people, meanwhile, have made it more than clear that they are ready to transform their country. The United States can and should help them by using its military power to neutralize the Islamic Republic’s military capabilities and degrade its repressive domestic apparatus.
Such measures could inspire the masses of Iranians who took to the streets in December and January to do so again. Just this week, Iran witnessed smaller-scale campus protests, showing that animosity against the regime very much remains. If regular protests resume, American military power could level the playing field between the street and the state, giving the country’s demonstrators a chance to succeed.
DEAL OR NO DEAL
The Trump administration may have threatened Iran with major military action. But there are reasons to think that, for now at least, it might have other ideas. For starters, the president’s comments on Iran this year have oscillated between threats of war and the imperative of a nuclear agreement. “Hopefully Iran will quickly ‘Come to the Table’ and negotiate a fair and equitable deal,” Trump wrote on Truth Social in late January. “I would rather have a Deal than not but, if we don’t make a Deal, it will be a very bad day for that Country,” he posted a month later. During the February 24 State of the Union address, Trump declared that although his “preference is to solve this problem through diplomacy,” he would “never allow the number one state sponsor of terror … to have a nuclear weapon.”
Tehran claims the two sides have made progress in their talks. But thus far, Iranian officials have, expectedly, refused to give up core elements of their nuclear program, so there is reason to think that Trump will be compelled to attack—even as the two countries continue to negotiate indirectly. If past is prologue, he will keep such action short and sharp. In his first term, for example, the president ordered the assassination of Qasem Soleimani, a prominent commander in Iran’s Islamic Revolutionary Guard Corps (IRGC), with a drone strike in January 2020. In June 2025, he directed the use of massive ordnance penetrator bombs against Iranian enrichment facilities. And the recent extraction of Venezuelan president Nicolás Maduro from Caracas by U.S. military forces took place in the span of one evening. Notably, Trump has cited the Maduro operation while threatening Iranian Supreme Leader Ali Khamenei. “Like with Venezuela,” the president said in a Truth Social post describing Washington’s naval buildup around Iran, the U.S. military is “ready, willing, and able to rapidly fulfill its mission, with speed and violence, if necessary.”
But a quick, one-and-done operation is extremely unlikely to down this regime, even if it succeeds in killing Khamenei. The Islamic Republic may have once been a sultanistic state built on the personal cult of its founding supreme leader, Ayatollah Ruhollah Khomeini. But over the last three decades, Khamenei (Khomeini’s successor) has institutionalized his rule and his regime both by empowering loyalists throughout a considerably larger state bureaucracy and by supporting competing power centers. As a result, the Islamic Republic looks more like a series of pillars than it does a pyramid, with a powerful deep state made up of security officials with their own vested interests in maintaining the regime.
Seen in this light, many of the leaders and veterans of the IRGC are less Khamenei’s subordinates than his partners. Indeed, it is unclear to what extent the 86-year-old Khamenei is actually running the regime day-to-day. The Islamic Republic, for example, made remarkably quick military decisions during the 12-day war with Israel in June, when Khamenei was likely in a bunker and, according to a New York Times report, not using electronic communications. It did so even after many senior IRGC officers were killed in just one night by Israeli attacks.
The Venezuela model, in other words, will not work in Iran. In fact, a singular and spectacular strike on Khamenei might have the opposite of its intended effect. Rather than fostering division and thus jeopardizing the regime, remaining officials would be more likely to close ranks, at least initially. They would keep the system running and try to seek revenge.
GOING BIG
To some analysts, the fact that limited strikes will fail is reason enough to think twice about employing the military. “Iran will not cave to major demands simply because of a bombing campaign,” wrote Nate Swanson, a former White House adviser on Iran, in a cautionary Foreign Affairs piece. “An attack by the United States on Iran could result in unexpectedly deadly retaliation—and a much longer and potentially damaging conflict.” Unless Iran is bluffing, this analysis is correct. But it is not a compelling reason to avoid using the military. Iran is the world’s foremost state sponsor of terrorism, one of the most outright anti-American governments in the world, and the country with the largest ballistic missile arsenal in the Middle East. Nearly a half century of experience has shown that the Islamic Republic will not meaningfully moderate any of this behavior or treat its citizens any better. Washington now has a historic opportunity to bring down the regime, and it cannot pass it up out of fear. Indeed, the fact that Iran would almost certainly escalate in response to a limited U.S. strike is all the more reason to go big from the beginning and avoid settling into the kind of gradual escalation that turns wars into quagmires. Washington cannot let Tehran dictate the pace or terms of the conflict.
None of this means that the United States needs to launch a massive ground invasion and single-handedly topple the regime. In fact, part of why Washington should strike now is that Iran’s own people are prepared to do much of the work themselves. For the past decade, while the United States has treated “regime change” as a pejorative term, unarmed Iranians have increasingly taken to the streets to seek systemic change in what has become a national uprising. Today, it is clear that Iranians are willing to make tremendous sacrifices to get rid of their leaders. That is why the regime had to kill at least 30,000 people, according to the estimates of human rights groups, to put down the most recent protests, and why some Iranians have nonetheless continued to demonstrate in the weeks since. Iranians therefore do not need U.S. troops to march into Tehran. What they need is for the American military to weaken the regime enough for them to succeed.
A one-and-done operation is extremely unlikely to down the Islamic Republic.
Trump can start by having the U.S. intelligence community carry out covert operations designed to pave the way for kinetic activity. Intelligence operatives, for instance, should surge secure communications technology, including satellite Internet devices, into the country so that Iranians can still have Internet access even if their government cuts off domestic service. The administration should also authorize information operations aimed at undermining the resolve and cohesion of Iran’s security forces. Finally, the administration should direct the intelligence community to identify and help Iranian security forces that are willing to defect.
Then, the United States can proceed with airstrikes. It should begin by suppressing and destroying the regime’s air defenses to gain aerial superiority. Doing so shouldn’t be all that complicated, given Israel’s success in knocking them out in past rounds of fighting. But there are platforms the Israelis did not suppress or destroy, and others that Iran has repaired. Afterward, Washington should move against the long pole in the tent of Tehran’s deterrence: its formidable ballistic missile arsenal. Iran can use these weapons to complicate the United States’ approach by inflicting damage on military assets, regional energy infrastructure, maritime shipping, and even on civilian targets in Israel. Rather than wait for Iran to launch large volleys of these projectiles in hopes of overwhelming U.S. and allied defenses, Washington must collapse the network of subterranean bases where medium- and short-range ballistic missiles are kept, which Iranian officials have long hailed as “missile cities.”
To be clear, the United States and, potentially, its partners will sustain losses from Iranian missiles. But by preemptively striking and destroying these bases, Washington can help limit the damage from any Iranian counterattack. If Tehran disperses its mobile launchers after bringing them above ground, the United States should dedicate real-time intelligence, surveillance, and reconnaissance assets to identify their locations and strike them, as Israel did during its 12-day war. (Israeli estimates claim that the country’s attacks on these launchers reduced the regime’s firing capacity by at least 33 percent.) From there, the United States can handicap Iran’s future missile capacity by targeting production facilities, which open-source intelligence suggests are located in or near the cities of Isfahan, Khojir, Parchin, Semnan, and Shahroud.
Degrading Iran’s missile infrastructure will do more than make it easier for Washington to conduct the rest of its military campaign. It could also help ordinary Iranians. Drastic as it might seem, some regime supporters have floated using missiles against Iran’s own people. Likewise, if the Islamic Republic’s missile and nuclear infrastructure is destroyed and its contents entombed, the regime’s representatives will have less leverage when negotiating with U.S. officials or protesters over a political transition. For that reason, the United States would be wise to attack Iran’s remaining nuclear sites, especially as the regime is moving to harden or rebuild them.
PAVING THE WAY
Still, hitting Iran’s missile and nuclear program is unlikely to help Iranian demonstrators in the near term. To do that, Washington should also target the regime’s political institutions and security installations, both as a means of inspiring protesters and to make it harder for the Islamic Republic to coordinate and effectively suppress demonstrations.
Some American analysts might object to targeting Iran’s political institutions as a violation of the country’s sovereignty. But Iranians would likely welcome such a campaign. The country’s people are highly nationalistic, but they are nonetheless looking abroad for help in their fight against the regime. During the latest round of protests, for example, Iranians made English-language videos describing their plight. Others named streets in Trump’s honor, in hopes of getting the president’s attention. When Trump subsequently promised that help “is on the way,” he raised their expectations and likely fostered continued protests. As one demonstrator told The Wall Street Journal in a text message, “We’re all staring at the sky, hoping Trump will bomb us, just to destroy Khamenei and his regime.”
Trump failed Iranians then. But he can correct course now. He could begin to help them by striking state institutions that have ordered or supported the crackdown on Iranians such as the office of the supreme leader, the Ministry of Intelligence, the IRGC counterintelligence directorate, as well as institutions that enabled the regime to shut off the Internet. Washington should also use overt and covert means to blind, impede, and militarily pick apart Iran’s wartime command-and-control structures—including those on the supreme national security council, the defense council, and the armed forces’ general staff. The United States could then move to targeting bases and command centers housing other parts of the IRGC, the Basij paramilitary (whose members patrol Iran’s streets and put down protests), and the so-called special units of the country’s law-enforcement forces, who handle crowd suppression. Finally, if Washington determines that Iran’s proxy foreign militias are entering the country to slaughter demonstrators—as appeared to happen during the last round of protests—it should not hesitate to use force against them. These militias, after all, are already designated by the U.S. State Department as foreign terrorist organizations and already have the blood of Americans and U.S. allies on their hands.
Such actions will probably eliminate many key regime officials, including even Khamenei. In striking this wide variety of targets, the United States would show Iranians that it wants to change Iran for the better, not just limit the damage the regime can do to foreign adversaries. Washington should hit these sites even if key officials and leaders are not there. Doing so will still reduce the regime’s repressive capacity, impede continuity of government operations, and send a strong signal that can motivate the Iranian people.
Washington must fracture the Iranian regime.
As it carries out strikes, the United States will need to reduce the regime’s capabilities all across the country, not just in Tehran. That means U.S. officials should target every IRGC and Basij provincial command and battalion headquarters, as well as the command centers belonging to Iranian police forces. Doing so would inspire further protests while making it harder for the regime to suppress them. Nothing is likely to galvanize Iranian protesters more than seeing local security forces come under attack.
Finally, the United States must be prepared to use not just bombers and advanced fighter jets but lower- and slower-flying drones and aircraft, which could provide close air support to crowds squaring off against regime militants. Such attacks would help the United States take out mid- and lower-level commanders, drilling down on the chain of command of Iran’s security forces. It might also prompt these officials to retreat. As rank-and-file Iranian forces begin worrying about their survival, their instinct for self-preservation might kick in and override lingering loyalty to their units and commanders. Iran’s security services, in turn, might finally fracture.
This fracturing would be the key to how Iranians finish off the regime. As the Islamic Republic’s forces are sandwiched between U.S. airstrikes and popular pressure, they will have every incentive to either lay down their weapons or flip and join with the protesters. In the latter case, they might bring their guns with them, transferring coercive power to the street. Both outcomes could embolden demonstrators to press their advantage. They might take over police stations and occupy government and municipal buildings. With the security forces out of their way or actively helping them, protesters could seize state television, radio, and other communications platforms and broadcast the end of the Islamic Republic. Such an outcome would have parallels to the 1979 revolution: the Islamic Republic celebrates February 11 as “Islamic Revolutionary Victory Day” because it is the day when the Iranian armed forces declared their neutrality, effectively abandoning the existing government and handing the country over to its revolutionary masses.
The collapse of the Islamic Republic would, of course, be turbulent. Many analysts fear that Iran would simply get a new strongman, possibly someone who rises up from the ashes of the IRGC. But that is far from foreordained. Newly empowered protesters could use their new platforms to call for the country’s large civilian bureaucracy to keep working in order to maintain government functions. They might also bring in Iran’s exiled opposition leaders, who have been planning for and could help lead a transition. For Washington to eschew military options out of fear of instability would thus be to make the perfect the enemy of the good. The real source of destabilization in and around Iran, after all, is not the prospect of regime change. It is the Islamic Republic itself.
ONE WAY OUT
When Trump was asked about what an acceptable “endgame” with Iran would be, he replied: “to win.” After decades of Iran’s unabated anti-Americanism and hostility, it should be clear to all that winning means toppling the Islamic Republic. The regime is the arsonist behind many of the Middle East’s fires and a terror to its own citizens. Washington should employ sustained military force in order to break it and thus pave the way for Iranians to take back their country.
That doesn’t mean helping take down the Islamic Republic will be easy. The United States will run into significant challenges. It will encounter unknown variables, including the resilience of Iranian security services when facing American firepower and the Trump administration’s own risk tolerance once it encounters difficulty. But at this point, bold action is the only way to break the current dynamic. The United States has both the capacity and the capability to weaken the Iranian government while mitigating the resulting dangers. The Iranian people have the drive and determination needed to finish off the regime. Together, the two have everything they need to win and make a new Iran.
The Week’s 10 Biggest Funding Rounds: OpenAI Takes The Spotlight With Record-Setting $110B Round
It was going to be a fairly business as usual top 10 list this week until OpenAI decided to disrupt our Friday with news that it raised $110 billion in new funding. Yes, $110 billion. That is so much money, and so record-setting as a private company funding round, that it makes all those other $100 million and $200 million rounds we usually write about look very paltry by comparison.
That said, we did nonetheless see a number of these kinds of rounds, in sectors including semiconductors, AI, healthcare and biotech.
1. OpenAI, $110B, artificial intelligence: Generative AI giant OpenAI announced that it has raised $110 billion in new investment at a valuation of $730 billion pre-money, or $840 billion post-money. The deal includes $50 billion from Amazon, $30 billion from SoftBank, and $30 billion from Nvidia. San Francisco-based OpenAI says more investors are expected to join as the round progresses.
2. (tied) MatX, $500M, semiconductors: MatX, a startup that designs custom chips and hardware architectures to support large language models, secured $500 million in Series B funding as it prepares to scale manufacturing. Jane Street Capital and Situational Awareness led the financing for the Mountain View, California-based company.
2. (tied) Vero Networks, $500M, broadband: Boulder, Colorado-based Vero Networks, a fiber infrastructure and broadband internet provider, picked up $500 million in a growth funding round backed by Braemont Capital, Hamilton Lane and Delta-v Capital.
4. Shine Technologies, $240M, fusion: Janesville, Wisconsin-based Shine Technologies, a developer of fusion technologies with applications in the medicine and energy sectors, raised $240 million in equity funding led by NantWorks.
5. Revel, $150M, hardware testing tools: Revel, developer of a software platform for hardware test and control, closed on $150 million in Series B funding. Index Ventures led the financing for the Los Angeles-based company, which plans to expand its offerings across aerospace, defense, robotics and industrial sectors.
6. Honest Health, $140M, healthcare: Nashville, Tennessee-based Honest Health, a provider of tech-enabled tools for primary care providers, secured $140 million in a new financing led by NewSpring.
7. Slate Medicines, $130M, biotech: Slate Medicines, a startup working on therapeutics for headache disorders, announced its launch alongside $130 million in Series A financing. RA Capital Management, Forbion Capital Partners and Foresite Capital led the investment for the Raleigh, North Carolina-based company.
8. Ubicquia, $106M, smart infrastructure: Fort Lauderdale, Florida-based Ubicquia, provider of an analytics platform for smart lighting, grid monitoring and public safety applications, raised $106 million in Series D funding. 67 Capital and Marunouchi Innovation Partners led the financing for the 12-year-old company.
9. (tied) Basis, $100M, AI-enabled accounting: Basis, an AI agent platform for accountants, closed on $100 million in Series B funding at a $1.15 billion valuation. Accel led the round for the New York-based startup, along with Google Ventures, Lloyd Blankfein and Khosla Ventures.
9. (tied) Aalyria, $100M, satellite and network communication: Google spinout Aalyria, a developer of software that configures communications satellites to meet demand, secured $100 million in Series B funding. Battery Ventures and J2 Ventures led the financing for the Livermore, California-based company.
9. (tied) Viture, $100M, smart glasses: Viture, a San Francisco-based maker of extended reality (XR) smart glasses and accessories, says it raised $100 million in a financing led by Legend Capital.
Pavel Durov & Telegram — Key Takeaways
Durov, the 41-year-old sole owner of Telegram (1B+ users), was detained in France in August 2024 over the platform's failure to moderate criminal activity. He's under formal investigation on a dozen charges and barred from leaving France, though the travel ban was later relaxed. A trial isn't expected before year-end.
The bigger picture: His case has become a flashpoint for CEO accountability in Europe. France raided X's Paris offices this month, Spain is introducing personal liability laws for tech execs, and Russia has simultaneously opened its own criminal case against Durov for "aiding terrorism" — while pushing users toward a state-controlled messenger.
Business model concerns: Telegram runs with ~60 employees and is funded through $3B+ in corporate bonds. Nearly a third of its H1 2025 revenue (~$300M) came from Toncoin-related deals — a crypto deeply intertwined with the platform. Despite ~$400M operating profit in H1 2025, net losses were $222M due to Toncoin writedowns (the token is down ~80% from peak). Critics call the crypto-Telegram interdependence "a house of cards." A planned xAI/Grok integration ($300M deal) hasn't materialized.
Geopolitical tension: Durov denies ties to Moscow, but reports surfaced of 50+ trips to Russia between 2014–2021. He was in Baku overlapping with a Putin state visit just before his arrest. He's pivoted toward US libertarian audiences (Tucker Carlson, Lex Fridman) and cultivated relationships with Musk and French billionaire Xavier Niel.
Bottom line: Durov is squeezed from all sides — French prosecutors, Russian authorities, bondholders wanting an IPO, and critics questioning both Telegram's encryption claims and its crypto-dependent financials. His stance remains defiant: "I would rather die than become anyone's asset."
Telegram, Pavel Durov and the shaky future for tech’s libertarian princelings
The messaging app’s founder has become a poster boy for the backlash in Europe over CEO accountability — and a target for Russia
Pavel Durov, the billionaire boss of messaging app Telegram, steps off his plane in France and is swarmed by a Swat team of police officers. He pulls off his shirt to reveal a torso so sculpted that it momentarily confuses his opponents before they come to their senses, cuff him and march him off to a cell. “Privacy is over!” a faceless goon shouts, zapping him with an electric current. Durov awakes to find a chip implanted in his head and the task of breaking out of a high-security prison before him.
This is the plot of Total Glitch, a game hosted on Telegram that dramatises a fictitious French prison break by a hero who recovers from bouts of intense combat with ice baths and yoga. Durov, a 41-year-old wellness evangelist who in real life recovers from bouts of intense exercise with ice baths and yoga, says he had no role in the game’s development, but he has promoted it on his channel. “Only on Telegram could independent developers pull this off . . . and the results are wild,” he wrote last October in a Telegram post viewed more than five million times.
The scenario the game imagines is not entirely far-fetched. In August 2024, Durov, who was born in Russia but has French and United Arab Emirates citizenship, was detained by French police within minutes of stepping off a private jet at Paris-Le Bourget Airport. He was held over alleged failures to moderate criminal activity on Telegram, from drug trafficking to terrorism and the dissemination of child sexual abuse material. After four days in custody, he was released and placed under formal investigation on a dozen preliminary charges and barred from leaving France. A trial, which is unlikely to come before the end of the year, could determine whether Durov will be convicted and face fines, or even a long prison sentence.
One of the biggest messaging platforms in the world, with more than a billion users, Telegram is marketed as a safe haven for communication, as well as a one-stop shop for broadcasting, file sharing, shopping, gaming and more. But its claims to be ideologically and technologically unassailable from the prying eyes of authorities, combined with a hands-off approach to moderation, have made it the default channel for anyone operating at the margins of legality, according to critics. It’s where you might find gruesome videos of explosions and beheadings alongside semi-official communications from militaries and intelligence agencies.
The French charges against Durov have serious implications for every other social media platform and the liability of their top executives. “Arresting a CEO of a major platform over the actions of its users was not only unprecedented — it was legally and logically absurd,” Durov wrote in a Telegram post, a year after his brush with police.
Since then, European lawmakers have shown ever more interest in holding tech CEOs accountable for alleged abuses on their platforms. This month, French and European investigators raided the offices of Elon Musk’s X in Paris, as part of a wide-ranging investigation into its algorithms as well as the spread of AI-generated sexual abuse material. Spain is introducing laws to ban under-16s from social media and hold executives personally liable for hateful content. Last week, Russia stepped up its disruption of Telegram, throttling some features as part of an effort to force users on to a state-backed “national messenger” built for surveillance, and accusing Durov of pursuing his own self-interest while enabling terrorism and crimes against children.
For the libertarian tech titans of the US, and their political allies, Durov has become a martyr to European censorship. In the 18 months since he was released from custody, he has displayed increasing disdain for French authorities, accusing them of pressuring him to silence certain European conservative voices on Telegram and suggesting his detention was politically motivated. (French officials point to a strict separation between the judicial and political systems. The prosecutor’s office leading the investigation declined to comment for this story.) After the recent raid on Musk’s Paris office, Durov posted that France was “the only country in the world that is criminally persecuting all social networks that give people some degree of freedom… Don’t be mistaken: this is not a free country.”
To some observers, the investigation looks like a responsible attempt to force a secretive leader to shed some light on the allegiances and inner workings of his multibillion-dollar empire. Unlike the bosses of rival platforms such as Meta or xAI, Durov wholly owns Telegram, funding it through a personal fortune estimated at $17bn, which comes from his ownership stake, as well as personal bitcoin investments. And although he has frequently dangled the idea of an initial public offering, it remains unclear if he is ready to loosen his grip. In early 2024, in an interview with the FT, he claimed to have rebuffed offers of equity investments that valued the company at more than $30bn. According to analysis of documents obtained by the FT, interviews with former insiders and publicly available information, he has tied Telegram’s revenue prospects in part to a niche cryptocurrency whose underlying technology was originally developed by his team, rather than relying on mainstream advertisers who might demand more disclosure.
As the months pass, the circle of people wanting answers from Durov is growing. Corporate bondholders, who he has tapped for more than $3bn, including a $1.7bn offering last May, are asking whether the company will ever make good on the plan to float on public markets, now further delayed by the investigation. Some had been attracted by an option to buy shares in any future flotation at a discount of as much as 20 per cent. One bondholder raised concerns that the company’s IPO promises represent an “investment trap”.
Durov is now caught between two fires. In the west, the scant information around Telegram and its owner has done little to quell speculation that Durov still has ties to Moscow, despite his repeated denials. In Russia, a criminal case was opened this week against him for “aiding terrorism” and allowing Telegram to become a tool of western intelligence. In response, Durov claimed that Russian authorities were “fabricating new pretexts to restrict Russia’s access to Telegram as they seek to suppress the right to privacy and free speech”. Amid the conflicting narratives, Durov’s critics argue, it is hard to believe in anything other than his singular ruthlessness.
“Durov lives in a world of spin where he will decide on a story then grow a back-story and it will become the truth,” said Axel Neff, a Telegram co-founder who is now estranged from Durov. “He views himself as a man of the people. But how does one person become the sole shareholder of a company probably worth around $50bn? You’ve got to step on some people. To the victor, go the spoils.”
The “Mark Zuckerberg of Russia”, as Durov became known in his twenties after founding the country’s answer to Facebook, VKontakte, cast himself as an impish, anti-establishment folk hero from the start. Born in St Petersburg to a Ukrainian mother and Russian father who was a prominent professor of philology, Durov always intended to build “a one-man show and to do it alone, and that hasn’t changed”, said Nikolay Kononov, author of a 2025 book about Durov and Telegram, The Populist. “He’s a one-man state . . . the king of the outlaws.”
Durov seemed unafraid to publicly troll authorities, even while the Kremlin tightened its grip on the internet sector during the early 2000s. In 2011, in response to demands from Moscow that he shut down the accounts of Russian opposition leaders on VK, as the platform became known, Durov posted a picture of a “resistance dog” in a hoodie with its tongue out. When Russian police raided VK’s offices in 2013 over an alleged hit-and-run incident with a policeman, Durov posted, “When you run over a policeman, it is important to drive back and forth — so that all the pulp comes out.” Telegram has denied he had any involvement in the hit-and-run and called the accusation an attempt to discredit Durov during a period of corporate upheaval at VK.
It was in 2012, Durov has said, that he and his older brother, Nikolai, a maths prodigy, started developing Telegram as a “private tool” for the pair to communicate securely. In his telling, Durov left Russia in early 2014 after more pressure from Moscow to share Ukrainian user data with the government, selling his VK stake under duress to oligarchs. He eventually settled in Dubai.
Telegram ‘is more like a religion than a company to’ Durov
Shortly after his exit, Durov was sued by United Capital Partners, a Russian investment group that had bought a 48 per cent stake in VK. It argued he had “embezzled corporate funds, breached his fiduciary duties to shareholders and diverted corporate opportunities to himself” by setting up Telegram using VK resources. Durov was “attempting to present himself as a political dissident in order to divert attention from his serious legal challenges,” UCP said. Durov countersued in a US court, alleging UCP was trying to coerce and defraud him out of his ownership of Telegram. As part of a 2014 settlement, all parties agreed to drop their claims. But Durov got his prize: the reins of Telegram.
These days, there is no question of who is in control at Telegram. To work there is to join the cult of Pavel. The boss sports a custom-designed, all-black ensemble, inspired by Keanu Reeves’ character in The Matrix. He enjoys a lifestyle of ascetic ritual: at least 200 push-ups and 100 sit-ups each morning. No alcohol, tobacco or coffee.
Where Silicon Valley rivals such as Meta and X have thousands of staffers, Durov has around 60 full-time underlings. The identities of his staff are shrouded in secrecy. On its own LinkedIn page, the company states: “Telegram has a No LinkedIn policy. Accounts claiming to be past or present employees never worked at Telegram.” But, according to people familiar with the matter, many are young, brilliant software engineers plucked from relative obscurity by Durov and his brother from coding championships. These hires are paid lavishly. They spend summers working away from Dubai to escape the heat, once at a castle in Umbria.
“There is practically zero churn. They share the same values and believe in the mission of the company,” Durov told the FT, in the 2024 interview several months before his detainment in France. He claimed to be “deeply” involved in all Telegram’s processes. “It’s more like a religion than a company to him,” said one Dubai associate. When Durov missed the twice-weekly company meeting during his detention, it was one of only two times he had ever done so in 12 years, according to people familiar with the matter.
Dated mid-2025, previously unreported documents revealing Telegram’s leadership team support the claim that Durov’s top lieutenants have stuck around. Igor Diakonov, chief infrastructure officer and head of infrastructure management, joined in 2013. So did Mike Ravdonikas, who, as chief operations officer, runs Telegram’s external comms, regulatory affairs and moderation, as well as maintaining a website showcasing his own poetry. Durov’s three top engineering leaders also joined in 2013, all from VK. No executive has left since 2014, according to one person familiar with the matter.
While Durov is the public face of Telegram, the intellectual power is his brother Nikolai, according to multiple people who know the brothers. Those who know the chief scientist describe him as a mathematical savant: shy, dishevelled, often lost in his thoughts. As a student, he won prestigious maths competitions, before twice winning the International Collegiate Programming Contest, the Olympic Games of programming. (Neither French prosecutors nor Telegram would comment on a Politico report that there is a warrant out for Nikolai.)
“They are like two hemispheres of one brain,” said Kononov. “Nikolai is about tech inventions, and Pavel is about translating these into products, marketing and fulfilling his political dreams.” Neff agreed: “When you have someone like [Nikolai] in your back pocket, you can basically do anything in the tech world. Pavel generates the demand and the spotlight.”
With Nikolai at his side, Durov has made a roaring success out of Telegram. Surpassing one billion monthly users in March 2025, the platform has become a key communications channel for officials and the military, particularly in Russia, Ukraine and other eastern European countries, and a democratic lifeline during moments of civil unrest, from Hong Kong to Iran and Belarus.
Insiders say the app’s superpower has been rolling out product updates at breakneck speed, expanding beyond private messaging into a one-stop shop with public groups, broadcast channels, gaming, shopping and payments services. “What Durov is good at is lateral arbitrage — taking something that works in one market and launching it into a new market and improving on that,” said Neff.
In 2021, cyber security experts argued that despite Telegram’s promises of absolute privacy in its private messages (not even the company is supposed to be able to access them), its end-to-end encryption displayed cryptographic weaknesses. As the app broadened its services to include group-chat capabilities and broadcast channels, critics noted that users might mistakenly assume these features were also end-to-end encrypted by default, and that Telegram had failed to correct the record.
Last week, a Russian minister claimed that foreign intelligence services were able to see Telegram messages sent by Russian soldiers fighting in Ukraine. “Telegram’s not a private messenger,” Signal founder Moxie Marlinspike said on a podcast in February. “It’s the opposite.”
Telegram told the FT that its employees “cannot access” any user messages. Even those that are not fully end-to-end encrypted and instead stored in the cloud are protected because its decryption keys are split into parts and never kept in the same place as the data they protect, it has said. “No breaches of Telegram’s encryption have ever been found,” it said. “The Russian government’s allegation that our encryption has been compromised is a deliberate fabrication intended to justify outlawing Telegram and forcing citizens on to a state-controlled messaging platform engineered for mass surveillance and censorship.”
In 2018, Nikolai Durov developed Gram, Telegram’s own cryptocurrency and underlying blockchain, while Pavel raised $1.7bn in private sales of rights to Gram ahead of its launch. But the following year, the US Securities and Exchange Commission sued Telegram and blocked it from distributing the tokens, alleging Gram was an unregistered security. Telegram settled the case, paying an $18.5mn fine and agreeing to return $1.2bn to investors, including Silicon Valley venture capital firms such as Benchmark and Kleiner Perkins.
Gram’s underlying code was revived by the open-source community, with a cryptocurrency called Ton emerging as the most popular token. Durov still has close affiliations with the Ton community. Andrew Rogozov, a former VK executive and a Durov associate, was a founding member of The Ton Foundation, an open-source group driving Toncoin’s development. He now runs The Open Platform, a start-up that builds Toncoin tools for Telegram. Meanwhile, Durov has publicly championed the coin and integrated numerous Ton features into Telegram. Developers are encouraged to build games and experiences into the app that use the currency; advertising on Telegram is paid for using Toncoin; a Toncoin wallet has been built into the app for peer-to-peer and ecommerce payments; creators are able to sell and earn in the currency.
In 2025, Toncoin began to gain mainstream popularity. In March, VC firms invested more than $400mn in the currency, according to Durov, who cited top investor groups Sequoia Capital, Benchmark, Ribbit, Draper and Vy. Shaun Maguire, a Sequoia partner known for his pro-Trump stance, said in a statement that the open-source team working on Ton was “best in the world at the intersection of consumer product thinking and crypto infrastructure”. He added: “When you combine this with the global distribution of Telegram, we’re very excited to see where they go.” The coin has now been listed on major cryptocurrency exchanges such as Kraken, Gemini and Coinbase, as Durov makes inroads in Silicon Valley.
The Toncoin ecosystem has propped up what would otherwise be modest revenues and growth for Telegram, according to financial filings obtained by the FT. In theory, as demand for the crypto rises due to its integration on the platform, so does the price of Toncoin, which Telegram holds in plentiful supply. In the first half of 2025, nearly one-third of Telegram’s revenue, or $300mn, came from Toncoin-related deals, according to documents and people familiar with the matter. Premium subscription nearly doubled to $223mn sales in the first half of 2025, as the company added news features for power users, while advertising revenue ticked up only 5 per cent to $125mn over the period.
On a call with bondholders late last year, a transcript of which was seen by the FT, Telegram’s chief investment officer John Hyman said “the bulk” of crypto revenues were through a long-term contract with the Ton Foundation that would continue until 2029. “We see huge potential for Ton, for the position we still have in Ton,” said Hyman, a former Morgan Stanley banker who single-handedly oversees Telegram’s debt financing.
Yet the cryptomarket is notoriously volatile, and Toncoin’s price has been weighed down by wider market bearishness. Despite an operating profit of nearly $400mn in the first half of 2025, Telegram’s net losses were $222mn, compared with a $334mn net profit in the same period of 2024. One person familiar with the matter said this was because the company had to write down the value of its holdings of Toncoin, the price of which is down by about 80 per cent since its peak at $7.88 in July 2024.
“There’s very little organic demand [for Toncoin] outside of Telegram’s own internal mechanisms,” said Elies Campo, who helped work on Telegram’s growth, partnerships and business development between 2015 and 2020. “The circularity is the core problem . . . If either side weakens, the other follows.”
Last summer, Nasdaq-listed Verb Technologies, a marketing video software group, agreed to a $558mn investment from a vehicle run by Manuel Stotz, then president of the Ton Foundation. In return, Verb switched to become a digital assets treasury company holding Toncoin. Verb, since renamed Ton Strategy Co, has invested more than $700mn in the cryptocurrency. Announcing the deal, Stotz, who is now Ton Strategy’s executive chairman, described Durov as the “GOAT (“Greatest of All Time”) — in the domain of social media & product”, and posted a topless photo of the founder: “His proof-of-work [is], well, quite visible😁”
Telegram told the FT that its “long-term business model is not dependent on cryptocurrencies”, adding that its “main focus” is selling premium subscriptions, as well as advertising. The company called Campo’s analysis: “uneducated, unreliable and just plain wrong”.
Some bondholders remain bullish about Telegram’s business prospects. One told the FT that the platform had opportunities for swift revenue growth given it was highly undermonetised. Another added, “What makes these guys different is they are better in connecting the whole ecosystem to the messaging platform, and they seem more capable than other platforms in finding ways to monetise the ecosystem.”
Campo was not convinced: “Strip out the Ton sales and crypto deals and you have roughly $350mn in real product revenue” in the first half of 2025, he said. “It’s a house of cards . . . Pavel has navigated impossible situations before. The margin for error is just extremely thin.”
It was mid-May 2025, at the Cannes Film Festival, and Durov was attending a charity auction in the resplendent Hotel du Cap-Eden-Roc with his girlfriend, the glamorous crypto influencer and aspiring actress Yulia Vavilova. As the auctioneer reeled off the lots, Durov casually bid $400,000 to buy her a role in an upcoming Spike Lee movie. The auctioneer slammed down a gavel and the ballroom erupted into cheers of delight. Durov did not so much as flinch.
It was more than an extravagant romantic gesture. The theatrics served as a reminder to anyone watching — French authorities included — that nine months after his detainment Durov would not be cowed. Since the previous summer, he had been living in luxury lockdown in one of Paris’s palace hotels. Vavilova had posted pictures of five-star spas and horse riding near the palace of Versailles. That year “felt like a survival race straight out of Squid Game,” she wrote on Telegram, “while other moments were more like an enchanting Emily in Paris story.” The travel ban was relaxed in July 2025.
A Telegram bondholder put it differently. “He’s so rich that he feels he owns the world. In some ways, it’s true . . . Poor Pavel, he’s forced to live in some Parisian palace. I’m sure he’s enjoying it and living like Putin does in Russia.”
Durov has appeared “increasingly paranoid” since his detainment and “seems to mostly be talking to a very narrow circle of people”, said Aleksandra Urman, a researcher at the University of Zurich who has studied extremism on Telegram. She noted he had also strategically embraced American libertarians, granting video interviews to conservative commentator Tucker Carlson, and meeting in Paris with podcaster Lex Fridman. “He’s trying to pivot to a certain part of his western audience. In the west, he knows his main user groups have been extreme-right or right-wing groups. It lands well.” Telegram disputed this characterisation, adding that its users “don’t belong to a certain political party, what unites them is a belief in free speech”.
Durov is also close to Xavier Niel, the maverick French tech and telecoms billionaire who was among the first people Durov called after his detainment, according to people familiar with the matter. Niel has taken regular walks with Durov around Paris and advised him on how to navigate the French legal system. Niel, who spent a short stint in prison early in his career related to his early investments in sex shops in France, told the FT in November 2024: “When I went to prison, everyone disappeared on me. So when a friend runs into a problem in France, I’m not the kind of person who doesn’t pick up the phone.”
Durov has also courted Elon Musk, who associates say he long admired. In December 2024, Durov met Musk for lunch when the X boss was visiting Paris for the reopening of the Notre-Dame cathedral. Shortly after a second meeting in May, Durov announced a one-year deal with xAI to integrate the start-up’s controversial Grok chatbot into Telegram. Telegram was to receive $300mn from xAI, as well as half the revenue from xAI subscriptions sold via the messaging app, he wrote on X.
That partnership has not materialised. Instead, Telegram recently launched Cocoon, a “decentralised AI network” designed by Nikolai and others, whereby individual owners of the expensive GPU chips that power AI models rent them out to developers in return for Toncoin.
There have also been PR crises. In July 2024, a woman named Irina Bolgar told Forbes Russia that Durov was father to her three children, providing proof of his paternity. Days later, Durov posted on social media about his widespread sperm donation projects, saying it was possible he had “over 100 biological kids” and he planned to “open source” his DNA so that his children could find one another. In August that year, it emerged that Bolgar had filed a criminal lawsuit against Durov in a Swiss court in 2023 alleging he had been violent towards one of their children and had stopped supporting the family financially, according to court documents. Telegram rejected the allegations: “These events, which never occurred, were only ‘remembered’ by Ms Bolgar . . . as she was preparing to initiate legal proceedings for personal financial gain. The suspicious timing raises doubts about the credibility of these accusations.”
He’s so rich that he feels he owns the world. In some ways, it’s true . . . Poor Pavel
One of Durov’s preoccupations, as he has gone after western investment and audiences, has been distancing himself from his native country. In the past, Durov has refused to take sides on the Russia-Ukraine war or comment on Vladimir Putin: “Let’s not go there,” he told the FT in 2024. But he continues to deny any ties to Moscow. “A sad spectacle of a state afraid of its own people,” he said on Tuesday after the news of the Russian criminal investigation broke.
In the wake of the French investigation, conjecture about these ties has grown in the west. Upon his detainment in Paris, Moscow raced to offer Durov assistance, demanding consular access to him and suggesting the case was politically motivated. Independent Russian news website iStories reported it had obtained records of Durov having travelled in and out of Russia more than 50 times between 2014 and 2021.
Other critics noted that immediately before his detainment, Durov had been travelling around Central Asia with Vavilova, staying at the Four Seasons in Baku, Azerbaijan, where they drove racing cars, took helicopter rides and sharpened their marksmanship at shooting ranges. The trip appears to have overlapped with Putin’s state visit to Baku on August 18 and 19. The Kremlin has publicly denied there was any meeting between the pair.
Even so, said one European official, if Durov had “problems with Russia . . . I struggle with the messaging. Durov was in Baku, that’s practically Putin’s backyard”. Telegram said that Durov made infrequent trips to St Petersburg to visit family, which he posted about publicly on his Instagram account, noting he has not returned since the start of the war in Ukraine. The company rejected the idea of any meeting between Durov and Putin in Baku. It said: “Mr Durov was in a city of 2.4 million people outside of Russia, not at a backyard barbecue across from the Kremlin.”
Evaluations differ on the strength of the investigation against Durov. His allies believe the case against him is weak, the result of a net cast as wide as possible by French prosecutors who, they claim, lack expertise in digital platforms. “I have never seen so many accusations with so little legal precision,” one of Durov’s lawyers, Christophe Ingrain, told French magazine Le Point.
The prosecutors leading the investigation say they acted after receiving complaints about Telegram’s lack of responsiveness to requests from law enforcement in France and other European countries. Johanna Brousse, the magistrate who leads the cyber crimes division at the Paris prosecutor’s office, is a particularly formidable opponent. Her peers describe her as a determined campaigner unafraid to push boundaries to prosecute across borders and at the cutting edge of technology. “She’s like Joan of Arc . . . she’s impassioned about her work with a strong sense of the collective,” one said.
Durov’s patience with the French justice system seems to have worn thin. Last year, he accused top French intelligence officials of pressuring him to censor certain conservative Moldovan and Romanian accounts on Telegram ahead of fraught elections in each country. “I refused. We didn’t block protesters in Russia, Belarus, or Iran. We won’t start doing it in Europe,” he wrote on X.
In a rare public statement France’s intelligence agency DGSE said it “strongly refutes allegations that requests to ban accounts linked to any electoral process were made”. Instead, it had “been obliged, on several occasions” to contact Durov “to remind him firmly of his company’s responsibilities, and his own personal responsibilities, in terms of preventing terrorist and child pornography threats”.
In the coming weeks, Telegram is due to repay the principal on some of its bonds. Whether investors will be as forthcoming in the future as they have been in the past remains to be seen. For its part, Telegram insists on its independence in the near term. “Following next month’s bond payment, we will remain highly liquid, with cash on hand that materially exceeds our outstanding bond obligations,” it said. “Our balance sheet is strong, we are generating substantial cash flow, and we have neither the need nor the intention to seek external financing.”
One person familiar with Durov’s thinking claims the drawn out investigation has begun to weigh on him. “He feels all this is very unfair and misguided,” they said. Publicly, though, Durov has remained steadfast. “Up until this day we don’t understand what we did wrong . . . the investigation is ongoing and it’s been frustratingly slow,” he said, via video link to the Oslo Freedom Forum in May 2025. His appearance was introduced by a projected image of Durov dressed as a Greek deity, oiled chest barely covered by a white toga, with a goat slung over his shoulder.
Asked by the interviewer if he considered himself a hostage in France, or if French intelligence were trying to pressure him to “do what they want”, Durov was solemn. “I seem to be held to higher standards than most other platforms,” he replied. “I would rather die than become anyone’s asset”.
JPMorgan cuts services for Citadel Securities in clash over roles
Move underscores tensions between Wall Street banks and clients as businesses converge
JPMorgan Chase cut some trading functions it provided for Citadel Securities after the firm launched a service for clients that muscled in on the bank’s business, according to people familiar with the matter.
JPMorgan told Citadel Securities last year it would no longer offer the group so-called high-touch equity trading services, such as handling trades that are not done purely electronically and pitching research for trade ideas, the people said.
The move came after Citadel Securities announced plans for its own high-touch equity business that would compete with JPMorgan. To lead its effort, Citadel recruited the head of JPMorgan’s high-touch equities trading service, Elan Luger.
Citadel Securities is a market-maker that matches buyers and sellers on trades. It was founded by Ken Griffin, who also runs the separate hedge fund firm Citadel.
JPMorgan is still offering prime brokerage services to Citadel Securities and so-called programmatic trading services. The bank’s relationship with Citadel is unaffected.
The moves underscore the occasionally fraught relationship between banks and trading firms as their roles on Wall Street evolve. For a long time, banks viewed them as clients but are increasingly considering them competitors too as they expand into new areas.
JPMorgan and Citadel Securities declined to comment.
Citadel Securities has grown into a market giant by electronically facilitating billions of trades, often acting as an unseen processor of order flows from retail investors. Banks have shied away from spending the huge sums of money required to maintain technology infrastructure for such vast trading volumes.
But Citadel Securities’ move into high-touch equities trading has put them more directly in competition with investment banks such as JPMorgan.
For this, companies like JPMorgan and Citadel Securities seek to execute large, often complex trades for big asset managers and hedge funds such as BlackRock and Millennium Management.
With the start of its new equities business, Citadel Securities had been sourcing block trades, or large tranches of a company’s shares, through investors looking to sell shares, rather than investment banks, a person familiar with the matter said.
JPMorgan co-head of commercial and investment banking Troy Rohrbaugh on Monday told a shareholder event that the bank has “a long track record” of relationships where they both compete and service firms such as Citadel Securities.
“I feel very comfortable that we can hold our own and gain share,” Rohrbaugh said. “They may gain share as well but it will arguably, in our view, be at the expense of someone else.”
Miami-based Citadel Securities beta tested its offering for high-touch equity trading last year and formally launched the business at the start of 2026.
Equities trading has been a booming business for Wall Street in the past few years, benefiting from extreme bouts of market volatility around geopolitical and monetary policy swings.
JPMorgan’s reported equities trading revenue rose 33 per cent in 2025 to more than $13bn. Privately held Citadel Securities does not disclose financial results, but the FT reported last year that its profits jumped nearly 70 per cent in the first quarter of 2025 to $1.7bn.
Activists Step Up Pressure on Struggling Stocks. Beware, Long-Term Investors.
- Appaloosa’s David Tepper criticized Whirlpool’s stock issuance.
- Activist campaigns reached an all-time high of 255 last year, fueled by M&A resurgence and equity market volatility.
- A Harvard study found 60% of companies underperformed the market by nearly 10% after activist settlements.
Whirlpool stock continues to go down the drain, and David Tepper of Appaloosa Management, the company’s third-largest investor, isn’t happy. He’s joining the growing list of prominent activists pressuring underperforming companies to get their financial house in order.
Tepper sent a strongly worded letter to Whirlpool management on Wednesday, arguing that the company’s decision to issue more stock “resulted in a large, unnecessary dilution of shareholders.” Shares plunged 14% Tuesday on the news of Whirlpool’s $963 million offering, and have tumbled more than 30% over the past year.
Whirlpool isn’t the only major company that has come under attack from activist investors. Norwegian Cruise Line Holdings and Tripadvisor are in the crosshairs of Paul Singer’s Elliott Management and Jeff Smith’s Starboard Value, respectively.
In fact, activist shareholders have been having a field day lately. According to a report from Barclays, there were 255 activist campaigns announced last year, a new all-time high.
“The resurgence of M&A and equity market volatility created favorable conditions for activism,” said Jim Rossman, global head of shareholder advisory for Barclays, in the report. KeyCorp, Barrick Mining, Lululemon, and Keurig Dr Pepper were also notable activist targets toward the end of 2025, he added.
But with the trend likely to continue this year, regular investors should be wary. After all, activists often target weaker companies in need of change.
Activist investors have also targeted underperforming stocks overseas. The CEOs of U.K.-based oil giant BP and British consumer products conglomerate Unilever were both forced out last year after pressure from Elliott and Nelson Peltz’s Trian, respectively. (Elliott was the most active activist, if you will, launching 18 campaigns last year, according to Barclays). And just this year, Japanese toilet maker Toto has come under pressure from activist firm Palliser Capital to focus more on its ceramics business, a key supplier of components to semiconductor equipment firms.
But investors may want to think twice before buying a stock that’s come under fire from activists. Researchers at the Harvard Law School Forum on Corporate Governance pointed out in a report last April that stocks often trail the broader market once activists get what they want.
The Harvard team found that about 60% of the 634 companies they looked at between 2010 and the end of 2024 underperformed the market after activists settled with a company. And the average lag behind the market was nearly 10%.
“Companies targeted by activists may have fundamental issues that drew the activist in the first place and remain post-settlement,” the researchers said. “Another explanation is that settling with an activist and prioritizing their agenda in the boardroom is inherently disruptive, ultimately weighing down the company’s performance.”
The Harvard team added that “settlements do not generate benefits for investors with long-term investment horizons.” But there is one notable exception: when activists successfully pressure a company to sell itself. Those stocks outperformed the market by more than 15%.
Still, that’s a risky bet.
But what if an investor wants to follow the lead of activist investors? One possible approach is buying the stock when the agitator is first making noise and then selling if the target company finally gives in to an activist’s demands. Investors in Whirlpool, Norwegian, Tripadvisor, and other activist targets need to keep that in mind.
Insider Trading Is Shaping Prediction Markets. Polymarket Sees an Edge.
Polymarket’s social-media account has repeatedly identified signs of insider trading on its prediction-market platform. It isn’t clear what happens next.
On Dec. 16, a new anonymous account on prediction-market platform Polymarket made a $68,000 bet that Kevin Hassett would be nominated as the next Federal Reserve chair. An official Polymarket social-media account posted a screenshot of the trade, seeming to suggest it came from a person in the know: “Either @novice99 can read Trump’s mind, or they’re serving coffee at Hassett’s breakfast meetings.”
On a stock exchange, any evidence of insider trading would quickly draw the attention of corporate counsels and regulators. On prediction markets, it could be an unavoidable part of the business model.
Stand CEO Edward Ridgely, who designs software for prediction-market traders, says there are times when “insiders are obviously shaping the market.” He pointed to the spike in Polymarket odds that Taylor Swift would get engaged hours before the pop star’s official announcement.
In the end, Hassett didn’t get the nod for Fed chair. The Polymarket social-media account has made more than two dozen posts flagging other potential insider transactions happening on the company’s prediction market. The posts allow other Polymarket users to play along. About half of the bets paid off, Barron’s found.
Polymarket didn’t respond to questions about the social-media posts. “Having an edge to the market is a good thing,” Polymarket CEO Shayne Coplan told 60 Minutes in November, in response to a question about inside information on the platform. Coplan said the company has focused on the ethics of insider transactions, before adding that it was “sort of an inevitability that this will happen, and there’s a lot of benefits from it.”
Both Polymarket and rival Kalshi tout the learnings that come from people putting money on future events. For the predictions to be accurate, or efficient, the platforms rely on a pool of traders to distill the wisdom of the crowd.
The prediction markets are under growing scrutiny as a steady drip of questionable trades make headlines. In early January, one trader made a $30,000 bet on the ouster of Venezuelan President Nicolás Maduro hours before he was captured by U.S. forces, resulting in a $400,000 payout. The trade spurred U.S. Rep. Ritchie Torres (D., N.Y.) to introduce a bill barring federal workers from prediction markets.
This month, an Israel Defense Force reservist was arrested and charged for allegedly trading on military operations through Polymarket. “In carrying out such bets, relying on secret and classified information, there is a real security risk to IDF activities and state security,” the Israel Police said in a statement.
How to handle inside information is already generating a schism in the prediction-market world. Kalshi’s rulebook explicitly bans “insiders” from trading on its markets; Polymarket’s rulebook does not.
This past week, Kalshi announced that it had suspended two accounts for suspicious trading: a former California gubernatorial candidate who bet on himself to win the 2026 governor race, and an employee of YouTube star MrBeast who traded event contracts related to his boss’ videos. The Commodity Futures Trading Commission said in a follow-up press release that both instances were potential violations of insider trading laws.
The California candidate didn’t respond to a request for comment, and the MrBeast employee couldn’t be reached.
The Coalition for Prediction Markets, a Washington, D.C.–based trade group, has recognized the potential damage from insider trading. In January, it took out a full-page ad in the Washington Post emphasizing that “regulated prediction markets (already) ban insider trading.”
Kalshi is a member of the coalition; Polymarket isn’t.
In January, a Polymarket trader with the username “HD2” bet nearly $100,000 that BlackRock executive Rick Rieder would be nominated as the next Federal Reserve chair. Polymarket Money, an X account run by Polymarket, posted a screenshot of the trader’s account. Polymarket’s own annotation emphasized that HD2 was new to the platform. “Who is @HD2?” Polymarket Money asked.
Market watchers say that accounts making large, first-time bets can be a signal of insider activity.
In another post, Polymarket Money said that a trader on its platform either “sees the future, or they’re writing the script.”
Since its first X post on Dec. 9, Polymarket Money has shared a total of 25 posts that identify possible signals of insider trading on Polymarket, Barron’s found. In each example, Polymarket Money posted a screenshot of a trade and asked followers what the trader might know.
Polymarket didn’t make executives available for comment, and it didn’t respond to questions about whether its system took additional steps to monitor the trades.
Advertisement - Scroll to Continue
The CFTC requires prediction-market operators to have trading surveillance systems on par with those at other licensed exchanges, like CME Group’s Chicago Mercantile Exchange.
Eventus, which provides market-surveillance software, lists Polymarket as a client on its website.
Polymarket has said it complies with applicable laws, and CEO Shayne Coplan told The Wall Street Journal that “the moment there’s a suspected insider, it’s pointed out on X, and it’s visible on Polymarket immediately.”
Polymarket has a data-sharing partnership with Dow Jones, the publisher of Barron’s.
Kalshi told Barron’s that if the HD2 trade were made on its platform, it would be flagged by their monitoring system.
Prediction markets sell event contracts, a financial instrument regulated by the Commodity Futures Trading Commission, whose insider trading rules are modeled after those of the Securities Exchange Commission, the stock market regulator.
The CFTC says it has “full authority to police illegal trading practices.” Those illegal practices include the “misappropriation of confidential information in breach of a pre-existing duty of trust and confidence to the source of the information.”
When it comes to stock trading, companies often have their policies in place to protect against the chance of insider trades.
“It is illegal to buy or sell shares in Walmart or our business partners based on important or sensitive inside information,” Walmart says in its employee code of conduct. There’s no mention of prediction markets.
It’s unclear how many companies have taken the step of explicitly banning employees from using work-related information to trade on prediction markets. Barron’s asked all 30 companies in the Dow Jones Industrial Average if they had issued new guidance around trading on prediction markets since their mainstream emergence last year. One declined to comment; 29, including Walmart, haven’t responded.
Not all trading by insiders is illegal. Company executives, for instance, can trade their stock according to certain rules and trading windows. Those trades are often closely watched by market participants seeking to track the so-called smart money.
Prediction-market traders are now doing the same thing. Some have made big money mimicking the trades of potential insiders with specialized software that can instantly copy the moves of any trader on Polymarket, which posts all of its transactions to the blockchain.
Advertisement - Scroll to Continue
Stand, Ridgely’s firm, has facilitated more than $200 million in trading volume across Polymarket and Kalshi.
As with most exchanges, trade monitoring largely falls on the prediction-market platforms, which are tasked with identifying suspicious patterns such as new accounts making large trades and quick withdrawals after a winning bet.
The CFTC steps in when the platforms refer cases for investigation, or when it deems surveillance systems to be insufficient. The agency’s capacity to regulate prediction markets and other exchanges has been curtailed over the past year, however.
Barron’s reported in early February that the regulator’s storied Chicago office had zero enforcement attorneys and progress on enforcement actions had been slowed by layoffs and attrition.
“Several employees left before my tenure, but we have appropriate resources throughout the country,” CFTC Chair Mike Selig recently told Barron’s.
The rapid ascent and large variety of prediction markets has presented new challenges to regulators and complicated the surveillance of suspicious trades.
Joe Schifano, head of regulatory affairs at Eventus, says his company’s surveillance tools take in all kinds of data to develop scenarios consistent with insider trading and manipulation. When a user makes a trade or a series of trades that line up with an established scenario, red flags go up.
Kalshi is constantly writing contracts with new language, says Nicole Kagan, who drafts many of those contracts for Kalshi.
The speed at which new contracts are listed along with their sheer number and variety—Kagan says Kalshi has “thousands”—makes it “theoretically impossible to create a scenario for surveilling each one that’s tailored to each contract,” says Dorothy DeWitt, the director of the CFTC’s division of market oversight from 2019 to 2021, when the regulator first granted a U.S. license to Kalshi.
Monitoring tools are built on decades of observing activity in commodities markets, DeWitt says. Prediction-market oversight doesn’t get the benefit of that history.
Asaf Meir, CEO of Solidus Labs, Kalshi’s monitoring partner, says the most sophisticated market abusers know that those limitations in surveillance systems exist, and are constantly testing the boundaries set by prediction markets. The role of the exchanges is to stay one step ahead, Meir says. “It’s a bit of a cat-and-mouse game.”
Prediction markets are raising new questions in an age-old debate around information and free markets. In fact, some see inside information as key to prediction markets’ accuracy, which has been touted in academic research.
“I think it’s actually incredibly important to have some form of insider information in these markets,” says Polygon Labs CEO Marc Boiron, whose firm operates the blockchain used by Polymarket. “When nobody has information that is actually useful to determine the outcome of that market, you’re just plain gambling.”
Still, Boiron and Kalshi emphasize the need for fair markets and insider-trading laws.
Robert DeNault, Kalshi’s head of enforcement, says insider trading can reduce market participation over the long term. “If a user can’t trust that they’re going to come onto our platform and not be in an unfair information asymmetry with another user, then they’re not going to keep using our platform,” he says.
A few traders who spoke to Barron’s believe they’ve been on the losing side of an insider trade. Kyle Baker once had positions on a Saturday Night Live mention market, betting on what words the host would say in their opening monologue. A few hours before the show went on the air, he saw a raft of new activity on Kalshi’s order book. The inside information? Potentially someone who watched the dress rehearsal of the show just a few hours earlier.
Tech Giants Are Paying Up to Power AI. These Utilities Will Be Big Winners.
Electric companies are gaining the upper hand in deals with Alphabet, Amazon, and others.
No AI magic trick is quite as impressive as the one it’s doing in financial markets—giving some utilities an upper hand with tech giants that are 100 times their size. It’s a trick that investors can cash in on.
A handful of utilities are poised to profit from the rush of spending by big tech firms, locking in a decade or more of cash flows to provide power to data centers.
It’s all the result of a potent new dynamic in the U.S. market for electricity. Tech companies are desperate to plug in their AI data centers, each of which can use 50 to 100 times as much power as the Empire State Building to run and cool their servers. In the past, utilities would hook data centers right into the grid like any other customer, and charge them a standard rate. But a political backlash against the warehouses sparked by soaring electricity prices is making that option untenable. So, tech companies are being forced to pick up more of the tab, signing deals that make them pay up even if AI proves to be a bust.
The issue even made it into President Donald Trump’s State of the Union speech on Tuesday. “We’re telling the major tech companies that they have the obligation to provide for their own power needs,” he said. The goal is that “no one’s prices will go up, and in many cases prices will go down for the community, and very substantially down,” he said.
The utilities that have already made deals with tech companies say that consumers will receive a kind of data-center dividend, $100 or more in annual savings on their electricity bills. But the utilities themselves may be even bigger winners, as they lock in years of cash flow.
It’s no secret by now that AI data centers use massive amounts of electricity, and they are only getting more power hungry. The data centers that are already under construction will need as much electricity as all of Italy, according to JLL, a commercial real estate and investment firm that tracks the industry and operates some data centers. The existing grid can supply some of that juice, but there’s a limit. Eight of America’s 13 regional electric grids are already at or below critical spare capacity levels, according to Goldman Sachs, raising the risk of blackouts if too many big customers plug in at once. As much as a quarter of expected data-center developments could end up short of electricity by 2028, according to Morgan Stanley analyst Stephen Byrd.
Tech companies are going to extraordinary measures to get that electricity, including using converted jet engines as power turbines and making special deals with Bitcoin mining companies to divert the crypto power to AI. But even those won’t be enough, analysts say.
In general, it’s good to own a commodity that’s in short supply. Just look at Nvidia. Its chips have only gotten more valuable as demand has risen. But electricity is different. Utilities have an obligation to serve local customers at a reasonable price. Surging demand from other kinds of users can cause more problems than opportunities.
The biggest problem now is that electricity bills are on the rise as utilities build out the grid and add new generation to serve data centers. Average U.S. residential electricity bills are up more than 30% since 2020. In Virginia, known for its warehouse-filled “Data Center Alley,” costs for all users are up an average of 17% in just the past year. Bills may continue to increase. The Edison Electric Institute, a trade group that represents investor-owned electric companies, projects that its members will spend $1.1 trillion from 2025 to 2029 on capital projects, nearly double the annualized rate at which they were spending in the prior decade. Someone has to pay for all of that investment.
Americans are increasingly blaming data-center operators for their rising bills, because it’s expensive to build the power plants and transmission lines to serve them, and costs are often shared by everyone who uses the grid. At least a half-dozen state legislatures are debating bills that could stop new data centers from being built. The anxiety cuts across partisan lines. Sen. Bernie Sanders has called for a national moratorium on data centers, but they’ve also drawn fire from Florida Gov. Ron DeSantis, who wrote on X that “there isn’t even close to enough grid capacity for these data centers. Consumers will see electricity rates spike—and to what end?”
Trump himself has started paying much more attention to electricity costs, which are shaping up as a political liability ahead of the 2026 midterm elections. He had promised in 2024 to cut electricity costs in half in the first year of his term. Instead, average bills rose 6.3%.
On its face, this kind of attention looks like bad news for utilities. Rising costs and decreasing reliability are a dangerous mix for an industry that isn’t particularly popular even in good times. But the political pushback against data centers has lately been falling more on the tech companies than on the utilities. Data centers have become “the villain,” Jefferies analyst Julien Dumoulin-Smith argues.
It isn’t hard to see why. Compared with such past buildouts as America’s railroads, data centers do little for local economies. They bring a fraction of the jobs to the towns where they’re built, and consume resources like power and water. More than $150 billion worth of data-center projects have been delayed or canceled since 2023 because of community pushback, according to Data Center Watch, a research firm backed by AI security company 10a Labs.
The tech giants are starting to realize this is a vulnerability. After previously fighting utility plans to make them pay more for power, tech executives have said this year that they’re prepared to take on more of the burden. In January, Microsoft President Brad Smith announced a “Community-First AI Infrastructure” plan meant to dampen the pushback. The first tenet of the plan is that “we’ll pay our way to ensure our data centers don’t increase your electricity prices.”
“Especially when tech companies are so profitable, we believe that it’s both unfair and politically unrealistic for our industry to ask the public to shoulder added electricity costs for AI,” he wrote.
Amazon.com and other tech giants made similar commitments to pay their own way. If the tech firms hold true to these pledges, the benefits will flow through to consumers. But they’ll also help utilities, whose earnings growth rates already are turning higher.
“If you’re at the utility, you’re able to benefit from this moment of political pressure that fell onto the hyperscalers,” said Rodney Rebello, a portfolio manager focusing on utilities at Reaves Asset Management. “They have to pay more. And now you can effectively run behind that.”
The utilities that are benefiting the most tend to be in unexpected places, outside the historical data-center growth areas like northern Virginia. In fact, 64% of data centers under construction are in so-called “frontier” markets, from the Midwest to the South, according to JLL. Tech companies are drawn to these markets by friendly regulations, available power, and tax breaks.
NiSource, a publicly traded electric and natural-gas utility that operates in Indiana and five other states, may be the best example of a utility that’s embracing the moment.
In 2024, the fast-moving world of data centers showed up at the doorstep of a sleepy NiSource subsidiary called Nipsco in northern Indiana. Nipsco serves about 500,000 electricity customers including a few industrial plants along Lake Michigan. All of its customers combined use about 2.4 gigawatts of power capacity, or as much as is provided by two to three large nuclear reactors. But two years ago, it started getting inquiries from data-center developers asking for much more power.
Added together, the prospective customers were likely to more than double the area’s power demand in just a few years. At the high end, the new customers could have added as much as eight gigawatts of new demand, says Nipsco President Vince Parisi. Tech companies were attracted to the area’s friendly regulatory environment, proximity to Chicago, abundant water supplies, and fiberoptic network.
That kind of sudden, enormous demand surge is a mixed blessing. Adding eight gigawatts of new demand to a 2.4 gigawatt system is like chaining an M1 Abrams tank to a Toyota Camry and expecting the two vehicles to drive calmly down the highway together. In theory, data-center deals could supercharge Nipsco’s earnings growth, with the company building large new power plants and earning regulated returns on the value of their infrastructure. But it could also leave Nipsco on the hook for billions of dollars in liabilities if the data-center developers walked away after the infrastructure was already built. When a power plant is abandoned, the financial burden tends to fall on consumer bills.
Nipsco didn’t want to turn these new customers away. So, it created a new subsidiary called GenCo, whose sole purpose would be to serve the biggest customers, keeping the impact away from their normal mom-and-pop customers.
In November, GenCo inked its first deal with Amazon, a plan to build 2.6 gigawatts worth of natural-gas power plants and 400 megawatts worth of battery storage—more than Nipsco’s entire existing system. Amazon’s data centers are expected to draw about 2.4 gigawatts worth of power, with the excess generation being sent to the rest of the grid. Amazon is expected to pay for the full cost of the generation, plus its share of transmission costs and other “shared system charges.”
The buildout is expected to cost nearly $7 billion, and the contract lasts 15 years—though Amazon will have an option to back out of part of it by 2029. Nipsco told regulators that it expects to be able to send $1 billion back to customers over the course of the contract, amounting to a discount on their bills of $7 to $9 a month.
Nipsco is in “active negotiations” with another one to three gigawatts worth of data-center customers “and then potentially behind that, there could be another three, as well,” Parisi said.
The deals won’t just help customers; they’ll also flow to the company’s bottom line. NiSource expects its adjusted earnings growth rate to accelerate to 8% to 9% from 2026 to 2033, above its prior growth rate of 6% to 8%. It’s a big reason Rebello likes the stock, which is a holding in the Virtus Reaves Utilities exchange-traded fund, which he co-manages. Dumoulin-Smith thinks it can get to $53, about 15% above current levels.
Other utilities are also in line to gain, and benefit customers.
Advertisement - Scroll to Continue
Louisiana’s Entergy, for instance, is building an enormous data-center campus for Meta Platforms in rural Louisiana, and working on one in Mississippi for Amazon. For the Louisiana center, Meta agreed to fund power plant equipment in advance. “For our residential customers, we estimate the data-center contracts in place today will generate approximately $5 billion in rate offsets from their fair share of contributions to fixed costs during the life of the contracts,” said Entergy CEO Andrew Marsh on a company earnings call. The company expects to grow earnings by over 8% annually through 2029.
Idacorp, owner of Idaho Power, has also been signing big data-center deals with companies like Meta, and is expanding its asset base faster than any other company in the industry, according to BTIG analyst Alex Kania, who rates the stock at Buy and expects it to rise another 10%. Idaho Power says its bills are 20% to 30% below the national average, “in part through the design of pricing and contractual provisions for new large load customers.”
To be sure, there’s some skepticism that these plans will work out in the way that utilities say they will. Some electricity experts and consumer advocates say the companies aren’t showing their math. Nipsco’s state regulatory filings list the Amazon power contract as “highly confidential.” Utilities often don’t detail how costs for added equipment like transmission lines will be paid for, argues Ari Peskoe, the director of the Electricity Law Initiative at the Harvard Law School Environmental and Energy Law Program.
Others note that the agreements between utilities and tech companies run for only 12 to 15 years, whereas power infrastructure like gas plants tends to last much longer. If tech companies don’t extend their contracts, utilities could be forced to find other ways to pay for that infrastructure. “You could sign a 15-year contract. But what does that mean for the other 45 or 55 years of the asset?” said Charles Hua, the executive director of PowerLines, a nonprofit focused on consumer energy issues.
Rebello said he believes that the tech companies are essentially paying for the full capital costs of the plants like the ones Nipsco is building, so the “leftover” costs after 15 years won’t be debilitating to existing users. Parisi of Nipsco said that’s an accurate assumption.
The other unknown here is how these agreements will affect the tech companies themselves over the long term. The five largest data-center players— Alphabet, Meta, Amazon, Microsoft, and Oracle —are on track to spend $710 billion in capital expenses this year. Most of that money will go toward pricey equipment like Nvidia chips.
Electricity costs are a minor part of the capital budget, and shouldn’t make much of a dent on the companies’ balance sheets. But the longer-term electricity contracts that tech companies are now signing with utilities will become significant repeating costs in another section of their financial results: the income statement.
“The new generation is actually less than 10% of the total cost of what it is to get that data center up and running,” said Moody’s Ratings analyst John Medina, who covers infrastructure finance. “But from an operating cost, like two-thirds of the annual operating cost is the power.”
The tech companies don’t make it easy to track exactly how much they’re spending, said Moody’s Ratings analyst Raj Joshi, who covers tech. “You can’t even get to the actual businesses,” he said. “They lump hundreds of billions of dollars in single lines. The SEC doesn’t require them to disclose” the electricity costs.
There’s a back-of-the-envelope way to calculate the potential costs. Electricity produced by a plant with a gigawatt of power capacity costs roughly $1 billion a year, depending on the source and the location. Assuming tech companies will add 40 gigawatts by 2030—and assuming they contract for it directly and pay for all of it—they could collectively be adding about $40 billion in annual electricity-related costs, a small but still noticeable piece of their $700 billion or so in total expenses.
Joshi isn’t yet treating it as a major credit problem, but “it does introduce this additional element of risk,” he said.
Caroline Golin, who led Google’s energy-market development until last year, says that tech companies have been buying power for years but for much different purposes than they are today. Google itself was a pioneer in the electricity market, creating contracts that allowed it to buy renewable power in long-term deals. But the purpose of their power purchases was to shift their energy use to cleaner sources as part of decarbonization goals. It was possible to do so using their existing cash flows, rather than taking on debt.
“These energy teams, they were built to meet a climate goal,” she said. “They weren’t built to produce tens of gigawatts of capacity in a short amount of time. And now they’re going to have to.”
The new kind of power purchases they’re making are mission critical, much more expensive, and almost certainly dirtier from an environmental perspective, she said.
The tech companies have said they would offset their fossil fuel-generated power purchases with renewables like solar. If so, their costs for all kinds of power are certain to keep rising.
The path for utility profits in the data-center boom is becoming clearer. Tech companies are ready to foot more of the cost. The bill will eventually come due.