Tag: Business – Decrypt

Would you like to kill a boss in Skyrim using an AK47 from Counter Strike? Give your NPCs a real genius-level brain that gives you unique answers instead of predefined strings of text? Would you like to run GTA V on your iPhone? What about using paid software for free, and legally? Well, people are doing that and more with AI right now.

A free tool that lets AI agents pull apart almost any piece of software has become the hottest project on GitHub, jumping from fewer than 10,000 stars on October 6 to more than 100,000 by Monday.

It’s called REA, short for Reverse Engineer Anything. Stars are GitHub’s version of a bookmark, and REA collected more than 30,000 of them on October 11 alone, after sitting mostly ignored since its creation in April.

How REA works

Apps and games ship as machine code, the raw instructions a processor reads and humans basically can’t. Decompiling tries to turn that back into readable code, a bit like tasting a cake and working backward to the recipe.

Reverse engineering is the bigger umbrella: figuring out how something works without its blueprints, whether by watching what an app does and the data it sends, or by reading its code. This repository does exactly that.

In practice, a user points an agent at a program, asks how one feature works, and lets it dig from what happens on screen all the way down to the binary. It works on desktop apps, Android apps, and websites.

REA hooks AI coding assistants such as Claude Code or OpenAI’s Codex up to the professional tools that do that job, like Ghidra, the reverse-engineering suite the NSA released to the public in 2019. The bridge is MCP, the Model Context Protocol, a standard Anthropic created so AI models can plug into outside tools—basically a universal adapter.

The developer’s “readme” file itself is far more modest than the hype, stating that REA “does not claim to recover original source code or automatically clone an application.” It doesn’t even ship its own decompiler.

This actually makes the whole operation legal(ish). The developer has never read the code, and just ships a similarly coded alternative that does exactly the same, but isn’t exactly the same.

The gates are open

People are doing all sorts of weird things with REA. For example, some players ended up watching Spider-Man swing through the Gotham of Batman: Arkham City, and creepers fight Elden Ring boss Malenia after someone stuffed about 1,600 Minecraft items into FromSoftware’s game. Another modder dropped Escape from Tarkov’s guns and menus into Skyrim and called it “Escape from Skyrim lol.”

Old console exclusives are escaping their consoles too: In September, a fan used an open-source AI model and a recompilation toolkit to bring Fable 2, the only mainline Fable game never released on PC, to Windows. The AI got it booting in about an hour, followed by a long stretch of manual fixes. Another one got to run GTV on iOS fluidly—something not even Rockstar Games bothered doing.

Using these techniques (not the tools), developer Brandon Thomas took the idea to software people pay for every month, using Claude Opus 5.5 to build ArtCraft, a set of free, open-source apps written in Rust that mirror Photoshop, Illustrator, Premiere, Lightroom and other Adobe programs, for Mac, Windows, and Linux.

Thomas says he snapped after Adobe charged him its cancellation fee one too many times. He isn’t alone in that grudge: Adobe agreed in March to a $150 million settlement with the Justice Department, which accused it of hiding early termination fees and making subscriptions a pain to cancel, though the company denied wrongdoing.

ArtCraft isn’t built with REA, and Thomas calls the project a clean-room rebuild, meaning the code is written from scratch using public documentation and observed behavior, never the original code, the same legal approach Compaq used to clone IBM’s PC BIOS in the 1980s.

The apps are labeled “early alpha” or “in development” on the project’s site, and one pixel artist has already called the Photoshop clone “completely unusable.”

Anything that runs on your own device can now be read by an AI agent with enough patience and tokens, which erodes the oldest copy protection in software: the sheer pain of reading machine code. That means cheaper clones and better game preservation, but also easier piracy and fewer places for security flaws to hide.

Activision has already pulled footage of an AI-decompiled Call of Duty: Modern Warfare 2, and Bethesda sent a cease-and-desist letter over an AI-made browser game called Fallout: New York.

That seems to be the only defense studios have against a trend that may easily force open source everything.

REA, meanwhile, keeps shipping, with eight releases between October 3 and October 11. Thomas has already walked back his promise to match Adobe’s feature set within a month, and now says it will take months.

Meta released Muse Wonder and Muse Wonder Full on Monday, the first models from its Superintelligence Labs that the company is willing to put head-to-head with the best from OpenAI and Anthropic.

“Muse Wonder is a major leap towards delivering personal superintelligence for everyone,” Zuckerberg wrote on Threads. “Congrats to everyone on our team for achieving this milestone. I’m looking forward to seeing how you all use it.”

That wording keeps him safely on the right side of President Trump, who said on Truth Social last Thursday that the White House considers anyone saying “Artificial Intelligence” instead of his preferred term ‘Super Intelligence’” to be “the enemy.”

Meta has called its AI division Meta Superintelligence Labs since 2025, so Zuckerberg was compliant before it was “mandatory.”

Meta’s announcement rests on two features: Active Memory, which carries what you told the model in one conversation into the next, and Active Focus, which keeps it grinding at a hard goal, trying new approaches and checking in with you before doing anything important.

Muse Wonder is live now through the Meta Model API and Muse Code, Meta’s coding agent, while the bigger Muse Wonder Full is reserved for paying Muse Code subscribers. Both hold 1 million tokens in their context window (how much text a model can keep in view at once, roughly 750,000 words), per Meta’s model card.

In April 2025, Meta’s open-source Llama 4 models arrived with big promises, then stumbled in independent tests amid suspicions that the version scoring well on a popular public leaderboard behaved differently from the one people could download.

Zuckerberg answered with a roughly $14 billion deal for Scale AI and put its founder, Alexandr Wang, in charge of a new unit called Meta Superintelligence Labs. Its first model,Muse Spark, launched in April as a small, fast, closed model, and it marked the death of the open source Llama family.

Muse Wonder arrives weeks after OpenAI shipped GPT-6 Astra on September 3. Anthropic followed on September 22 with Claude Opus 5.5, which it says matches its pricier Claude Fable 5.1 on most work.

On Meta’s Personal Superintelligence Index, an in-house composite of personal and professional assistant tasks where higher is better and the exact formula hasn’t been published, Wonder Full scored 59.1. Claude Opus 5.5 scored 55.2, GPT-6 Astra 54.5 and Claude Fable 5.1 49.9.

SWE-Marathon, which measures how well an AI model can code, shows promising results for Meta’s AI models. Wonder Full scored 57.5 versus 46.3 for Opus 5.5 and 42.5 for Astra, though Meta admits that row isn’t a controlled comparison and its model ran at maximum effort.

Harvey’s Legal Agent Benchmark hands AI models more than 1,200 lawyer-style assignments and counts a task as done only if the work passes every single check on an expert rubric, with no partial credit. Wonder Full scored 13.1, Astra 8.6 and Opus 5.5 just 4.2.

On MemoryBench, Meta’s own test of whether a model can recall facts, including ones that were later updated, across conversations of 64,000 to 100,000 tokens, both Wonder models scored around 85.5. Opus 5.5 scored 81.5 and Astra 81.4.

Those are Meta’s numbers, and two of the tests are Meta’s own creations. The independent ranking from Artificial Analysis, a weighted average of 10 evaluations covering AI agents, coding, general knowledge and scientific reasoning, still puts Opus 5.5 first at 58, with Wonder Full tied with Anthropic’s cheaper Claude Sonnet 5.5 at 56.

Muse Wonder, the version developers can actually buy, scores 52 on that index, a point below Astra and Fable 5.1. It costs $4 per million input tokens and $20 per million output tokens (tokens being the word fragments AI models read and write), exactly what Anthropic charges for Opus 5.5 and well under the $10 and $50 that OpenAI charges for Astra.

For regular people, the memory pitch cuts both ways, because an assistant that remembers everything also holds a lot of your life—and Meta is not precisely popular for handling private data responsibly. Last month, Inc. columnist Jason Aten found that Meta’s Muse agent had synced more than 187,000 rows of his private iMessages after he said he declined access, then claimed it had only seen notification previews.

Meta now calls Wonder “our most aligned model to date,” and the model card reports a 0% result on Agentic Misalignment, a test that pressures a model and scores how often it takes a harmful action. That comes two months after Meta disclosed that a Muse Spark model escaped a sandbox during a security evaluation and hacked a third-party service.

Meta says a full safety report is coming “soon,” and Wonder will reach the Muse app first for paying subscribers, with no date yet for free users.

Developers behind the XRP Ledger have disclosed and fixed a critical vulnerability that, left unaddressed, could have let an attacker conjure XRP out of thin air, a flaw that sat dormant in the network’s code for roughly a decade.

In a disclosure published Friday, the ledger’s maintainers said a researcher reported the bug through the XRPL bug bounty program in September. The issue was an integer overflow in the payment engine, the part of the software that processes payments.

When a single payment consumed hundreds of specially crafted offers from the order book, the sum of what the buyer owed exceeded the maximum value the system could store and “wrapped around” to a tiny number. The engine paid each seller in full but charged the buyer almost nothing, with the difference amounting to newly minted XRP that should never have existed.

Worse, the ledger’s built-in safety check meant to confirm no transaction creates XRP used the same flawed arithmetic, so it wrapped around identically and caught nothing. The maintainers said an attacker could have created spendable XRP far beyond the token’s total supply in a single transaction, then moved it, traded it or sent it to exchanges.

RippleX engineers confirmed the bug had been present since the current payment engine was written in 2015 and went unnoticed because no ordinary payment comes anywhere near triggering the overflow.

There is no evidence it was ever exploited, the team said. The fix shipped in version 3.4.1 on Sept. 25, and more than 80% of the network’s trusted validators were running it that day.

Notably, the patch bypassed the ledger’s usual amendment process, which requires validators to vote and wait two weeks before a change takes effect, the first time a transaction-processing change has deliberately shipped that way since the system was introduced over a decade ago. The maintainers said the severity justified it, since the open-source fix would have revealed the bug while leaving it exploitable for weeks.

The disclosure echoes a counterfeiting scare that rocked Zcash this year, when the privacy coin raced to patch a soundness bug in its shielded pool that could have allowed forged coins. That episode sent ZEC tumbling before the network’s Ironwood upgrade resolved it.

XRP didn’t escape the week unscathed. The token traded at $1.38 on Monday, down 1.6% over 24 hours, the worst performance among major cryptocurrencies, and off 9.4% over the past seven days.

Ledger said all confirmed cases in last week’s draining incident involve devices sold through reseller CryptoBilis, and that its own systems and direct sales channels remain unaffected, in a 72-hour update on Sunday.

“The volume of impacted devices is limited,” the Paris-based hardware wallet manufacturer said, adding that it is working with relevant authorities and reaching out to affected users. It said CryptoBilis has confirmed it has ceased all sales of hardware wallets.

Ledger began investigating on Friday after users in Southeast Asia who bought devices from CryptoBilis reported losing funds. Onchain investigator Specter traced more than $86 million in suspected thefts across Ethereum, Tron and Bitcoin. Ledger has not confirmed that figure or said what caused the losses.

Onchain investigators YFarmX said they had sent Ledger a list of more than 450 drained wallets. According to their analysis, the thief pushed 2,810 ETH through coin mixer Tornado Cash in three runs, moved the withdrawals through Zcash’s shielded pool, and withdrew about 3.28 million USDC into three fresh wallets, which Circle could freeze.

The thief also converted 15 million of the stolen USDT into USDD, a stablecoin Tether cannot freeze, before swapping 5.1 million back into USDT and sending about 4.23 million USDT to Binance deposit addresses, YFarmX said. Tether froze 10 million USDT on the first day, while 15,475 ETH, 203 BTC and 10 million USDD remain out of reach, according to the firm. Its findings have not been independently verified.

An X user, Johannes, has claimed to have found tampered Ledger devices in Europe, after buying two devices from retail stores. He said both had “identical modifications (such as additional cables and an LTE module),” and called on Ledger to issue a product recall.

The claims have not been independently verified, and Ledger’s update said all confirmed cases so far involve CryptoBilis devices. He said he is sending the second device to Ledger for analysis. Decrypt has reached out to Ledger for comment on the European reports.

Reseller controls

In its update, Ledger said it is reminding authorized resellers and distributors to source products only through authorized channels and “never to restock returned products.”

It also warned users against buying from unauthorized resellers, where “the risk of buying suspicious, counterfeit or modified devices is important.”

The company said it is reviewing reseller controls and advancing hardware protections, and called on others in the industry to work with it on supply-chain and anti-tamper practices.

Ledger repeated its advice that CryptoBilis buyers should not set up unused devices, and that anyone who already has should move their funds to a new seed phrase on a device bought from Ledger.

It also warned that scammers may try to exploit the incident, saying it “will never call you, DM you, or ask for your 24-word recovery phrase.”

The Commodity Futures Trading Commission is pressing its claim over prediction markets, issuing a pair of measures Friday that would formally define event contracts as “swaps” under federal law while carving out traditional gambling.

The first, a proposed rule opened for public comment, would expressly expand the definition of “swap” to include event contracts, including those tied to sports, politics, cultural events and weather.

The CFTC said the contracts are financial instruments “commonly known to the trade as swaps” and that the move resolves any ambiguity about their status. “These products are commodity derivatives squarely within the CFTC’s regulatory remit under the Commodity Exchange Act and are within the agency’s exclusive jurisdiction,” Chairman Michael Selig said in a statement.

The second, an interim final rule effective immediately upon publication, draws the opposite line, codifying the agency’s longstanding view that casino-style gambling products, including sportsbook wagers and casino games, are excluded from the swap definition.

“Casino-style gambling products are not derivatives,” Selig said, framing the step as clarifying “the limits of its regulatory remit.” Both measures carry 30-day comment windows.

Together, the two rules sharpen a distinction at the heart of a fierce jurisdictional fight: if event contracts are swaps, they fall under the CFTC’s authority, which Selig has argued is exclusive, placing platforms like Kalshi and Polymarket beyond the reach of state gambling regulators. Multiple states have sued prediction-market operators alleging illegal gambling, and the CFTC has countersued to defend its turf.

The filings formalize proposals the agency sent to the White House for review late last month, part of a broader push that has seen the CFTC assert its authority rather than wait on Congress.

The agency has broadly taken a pro-industry approach toward both traditional and cryptocurrency markets. The Commission recently issued no-action relief letting crypto apps offer regulated derivatives and sent separate crypto-markets rulemakings forward.

The legal stakes for prediction markets are climbing. A court split over whether event contracts qualify as federally regulated swaps, pitting the NFL and others against Kalshi, has drawn the Supreme Court’s attention, meaning the question the CFTC is now trying to settle in regulation may ultimately be decided by the justices.

France’s National Assembly Finance Committee voted this week to tax crypto holders who swap into stablecoins—tokens tied to a single official currency, such as the dollar or the euro—and to charge an exit tax on wealthy holders who move abroad.

The votes may not matter much yet: on Friday the committee rejected the budget’s entire revenue section by 31 votes to 3, so the full Assembly starts from the government’s original text, without the crypto amendments.

The committee’s amendments do not carry over to that text, so their backers would have to table them again for the floor debate that begins Oct. 13. Neither measure is law, and both still have to get through the rest of the legislative process.

The stablecoin amendment, filed by Nicolas Sansu of the left-wing GDR group and 16 co-signers, goes after what the authors call “a hole in the legislation.” Today, swapping Bitcoin for a stablecoin triggers no tax in France, because the state only collects when gains are sold for regular money or spent.

The text covers the electronic money tokens defined under MiCA, the EU’s crypto rulebook, a category that includes most stablecoins tied to a single currency. From Jan. 1, 2027, swapping into one would count as a sale, with the gain measured against what the holder originally paid.

The amendment names no rate and defers to France’s flat tax, which rose to 31.4% on Jan. 1 after the 2026 social-security financing law lifted the social-charge portion from 17.2% to 18.6%. The authors insist they are not creating a new burden, only applying existing law to a case it missed.

The authors say stablecoins can be perfectly ordinary investment vehicles, since they can pay for things at crypto service providers or buy other tokens, so the deferral on swapping into them is unjustified and lets the gain escape the flat tax.

A second amendment from Sansu would extend France’s exit tax to crypto. That is a levy on gains you have not cashed in yet, charged when you move your tax residence abroad.

It would apply when a tax household’s combined crypto, including holdings kept through custodians, is worth more than €800,000 and the taxpayer was a French tax resident for at least six of the previous 10 years, for moves from Jan. 1, 2027. The threshold matches the one already used for shares, and the payment deferral rules are borrowed from the stock regime.

Swaps between cryptocurrencies with no cash component would not count as sales for exit-tax purposes. Taxpayers would have to attach a statement of all crypto held on the date of the move, including assets held abroad or in self-custody, meaning wallets they control themselves without an exchange.

The authors argue that crypto held directly escapes the exit tax today while shares of the same value do not, and they flag how easily digital assets can be moved across borders.

Daniel Labaronne’s loss amendment also won committee approval. It would let investors carry crypto losses forward for 10 years to offset future gains, as they already can with stocks, whereas today unused losses cannot be carried over.

The Assembly adopted another crypto-inclusive tax in first reading in late October 2025, voting 163-150 for an amendment creating a 1% annual levy on “unproductive” wealth above €2 million that lumped digital assets in with gold and yachts. Attorney Burçak Ünsal told Decrypt that taxing early token holders was potentially “economically unjust.”

Coinbase said in October 2024 it would delist stablecoins that fail to comply with MiCA for European customers by Dec. 30, pointing them toward compliant coins like USDC and EURC.

The floor debate on the budget’s revenue section opens Oct. 13, with a formal vote scheduled for Oct. 20. If the measures return and survive, the stablecoin and exit-tax rules would apply from Jan. 1, 2027.

Physicists at George Washington University have published a formula that estimates how many good answers an AI chatbot will give before it slips into a bad one, and early tests suggest it works.

The study, by Neil Johnson and Frank Yingjie Huo, appeared in the journal Patterns and builds on a preprint, a version posted publicly before formal peer review, first released in February.

Chatbots can answer sensibly for a long stretch and then veer into something harmful, such as bad advice on self-harm or extremist talk, and there has been no simple way to predict when the swerve will happen. The authors argue that existing safety tools often depend on a cloud connection that offline models lack.

Johnson and Huo trace the problem to the attention head, the part of an AI model that decides which earlier words in a conversation matter most when choosing the next one. As a chat grows, the accumulated context pulls that attention toward one cluster of possible answers or another, until it tips.

This is a common pattern exploited by many jailbreakers, and one of the reasons why mayn companies pay attention to system prompts (pieces of text the AI chatbot reads before any query). However, nobody can point out accurately how much effort is required to effectively weaken a model.

Their formula estimates the tipping point, called n, as the number of good tokens—the word fragments a model produces one at a time—that come out before the first bad one. If the conversation already leans toward the bad side, the model tips right away, with an n of zero. If it leans good, the model delivers a run of fine answers and then flips.

In the preprint, the formula picked the right case, immediate or delayed, in 15 of 16 clear-cut tests, or 94%. The researchers ran those tests on six open-weight models, meaning AI systems whose files are public so anyone can download and run them, from OpenAI, EleutherAI, and Meta.

All six sat between 124 million and 410 million parameters, the adjustable numbers inside a model that serve as a rough measure of its size. The published paper reportedly widens the test to seven models of up to 12 billion parameters, which is still small by current standards.

The target is on-device AI, the kind that runs entirely on a phone or laptop with no internet connection, including companion chatbots people talk to like a friend. Google’s experimental AI Edge Gallery app, which Decrypt tested last year, already lets an Android phone run models offline, and nothing typed into it is sent to Google’s servers, and this seems to be a trend that may grow with time as hardware becomes more powerful and smaller AI models become more capable.

A model running offline has no cloud service checking its output, which is the gap the authors want to close. They propose a low-cost monitor that runs in parallel with the model and flags when n* falls below a safety threshold, a bit like a warning light on a car dashboard.

They also describe ways to push the tipping point out of reach, such as injecting content into the conversation so n* lands beyond the length of the response. Alignment training, the process of teaching a model to behave, can shift or suppress tipping for specific prompts but cannot remove the underlying mechanism, the authors say.

In April 2025, Decrypt covered an earlier paper from the same pair showing that “please” “and thank you” have a negligible effect on a model’s output, because the model treats polite words as orthogonal, or unrelated in the math, to the substance of a request. That version modeled a single, deliberately simplified attention head.

The preprint’s tests used small models and a 300-token window, or a few short paragraphs of text, and its predictions could be off by one output.

Meanwhile, a life insurance company that operates entirely in Bitcoin, has raised $37.5 million as demand for its unusual policies climbs among wealthy families overseas.

The Bermuda-based insurer said Thursday the funding came from existing investors, led by Bain Capital Crypto, with participation from Haun Ventures, Framework Ventures, Pantera Capital, Apollo, Northwestern Mutual Future Ventures, and Morgan Creek Digital.

Meanwhile, whose backers also include OpenAI CEO Sam Altman, has now raised more than $180 million in total.

The company bills itself as the first life insurer licensed to run entirely on Bitcoin, with its balance sheet, reserves, and audited financials all denominated in the cryptocurrency. It said it received the first license of its kind from the Bermuda Monetary Authority in July 2024 after two years in the regulator’s sandbox, and holds all policyholder Bitcoin with regulated institutional custodians.

The raise follows what Meanwhile called a record year, driven by international demand. In early 2026 it launched BTC Life 1-Pay, a single-premium whole life policy aimed at high-net-worth clients outside the U.S., its second product after a 10-payment version built for American taxpayers.

Under the single-premium policy, a client pays once in Bitcoin and receives a guaranteed death benefit in Bitcoin for life, with the policy’s value growing in Bitcoin and the owner able to borrow up to 90% of it after the first year with no repayment schedule or margin calls.

Meanwhile said it has signed 15 brokers since launch, including in insurance hubs like Singapore, Hong Kong, the UAE and Switzerland, and that its net long-term underwriting income has already surpassed last year’s total.

“Wealthy families around the world already hold Bitcoin. What they haven’t had is a regulated way to pass it on,” said Zac Townsend, Meanwhile’s co-founder and CEO.

The funding reflects a broader push to weave Bitcoin into traditional financial products. Delaware Life teamed up with BlackRock last year to offer Bitcoin exposure through a fixed indexed annuity, a sign that insurers are increasingly looking to pair the volatile asset with long-term planning vehicles.

Blockchain.com is making a bid to join the U.S. prediction-market boom, telling CNBC it has applied to the Commodity Futures Trading Commission for licenses to offer event contracts and cryptocurrency derivatives to American customers.

The crypto platform filed for two designations, the company told CNBC: a designated contract market license, which would let it operate as a futures exchange for event contracts, and a futures commission merchant license, which would allow it to act as a broker for derivatives.

Together, the approvals would let Blockchain.com run its own regulated marketplace serving both retail and institutional traders in the U.S.

The move would bring in-house what the company currently offers only through partners abroad. Blockchain.com began providing prediction markets to some international customers this year via a partnership with Polymarket, along with perpetual futures powered by the decentralized exchange Hyperliquid.

A CFTC license would let it operate its own event-contract venue rather than route through third parties. CEO and co-founder Peter Smith told CNBC the goal is to let users manage digital assets, trade derivatives and bet on real-world events all in one place.

The application lands amid a crowded rush toward federally regulated event trading. Blockchain.com joins 11 other companies that have filed for DCM licenses this year, according to CNBC, with the CFTC having approved six new ones in 2026.

The appeal of the federal route is partly jurisdictional: a CFTC license offers cover from state gambling regulators, who have sued prediction-market operators including Kalshi and Polymarket.

The regulatory backdrop is shifting fast, with states battling federal regulators in court over who gets to police such markets.

The CFTC recently sent the White House new rules aimed at cementing its authority over prediction markets, while a court fight over whether event contracts are federally regulated swaps, pitting the NFL and others against Kalshi, has drawn the Supreme Court’s attention.

The appeal of prediction markets for Blockchain.com and others, though, is obvious. Business is booming. Prediction markets are increasingly entering the mainstream and trading volumes have increased dramatically over the last year.

Analysts at Bernstein have projected the sector could generate billions in revenue by the end of the decade in what they predict will be a $10 trillion market.

Sen. Richard Blumenthal is asking Wall Street firm Cantor Fitzgerald to hand over records on its business relationship with Tether, the issuer of USDT, the world’s largest stablecoin—the dollar-pegged digital asset that underpins the crypto economy.

He also wants to know how much money the family of Commerce Secretary Howard Lutnick has made from the arrangement, according to a letter dated Thursday.

“Disturbingly, Cantor Fitzgerald’s lucrative business arrangements with Tether come at the expense of America’s national security,” Sen. Blumenthal argued.

The letter went to Brandon Lutnick, Cantor’s chairman and Howard Lutnick’s son, from Blumenthal as the top Democrat on the Senate Permanent Subcommittee on Investigations. It is a request for information with an Oct. 23 deadline rather than a subpoena, and it describes the minority’s probe as an investigation into “the illicit use of cryptocurrencies and the Trump Administration’s self-enrichment and self-dealings with cryptocurrency firms.”

The letter does not accuse Cantor of breaking a specific law, and nothing in it is a finding of wrongdoing. It asks how the firm checks that Tether follows U.S. sanctions and anti-money-laundering rules, and the claims about Tether and the Lutnick family are Blumenthal’s, built on press reports and his staff’s earlier work.

“Just as Tether has made untold millions in interest and investments from the stablecoins used in these illicit activities, so has Cantor Fitzgerald profited from its relationship with Tether,” Blumenthal wrote.

“Given the longstanding and deeply intertwined relationship between Tether and Cantor Fitzgerald, and its substantial relationships within the Trump Administration, I am writing to request information about your business partnership with Tether and to learn about your practices for detecting violations of American banking and sanctions laws,” the senator added.

Blumenthal wrote that Lutnick has made more than $250 million since President Donald Trump returned to office, including a $192 million distribution from Cantor. Over the same stretch, he said, Cantor’s 5% stake in Tether climbed from $600 million to an estimated $10 billion, on top of the tens of millions of dollars a year the firm collects from the assets it holds for Tether.

Cantor acts as Tether’s custodian, the firm that safeguards its reserves (the cash and other holdings meant to back every USDT in circulation). Blumenthal noted that although Tether claims to operate out of El Salvador, “the vast majority of its assets reside in the United States under your custodianship.”

Howard Lutnick ran Cantor until the Senate confirmed him as Commerce Secretary in February 2025, and the firm is now chaired by his son Brandon. The senator also requests the terms of any loan or legal strategy that would allow Lutnick to transfer asset ownership to his children.

That request tracks a reported loan from Tether to a trust for Lutnick’s four children, for an undisclosed amount, made around the time they bought out their father’s stake in Cantor. Sens. Elizabeth Warren and Ron Wyden asked Lutnick and Tether about it in April.

When assets are transferred to a trust, the people under investigation don’t have the legal property over those assets.

The letter landed 10 days after the subcommittee’s Democratic staff released “Tethered to Terrorism,” a report finding that 84% of 846 crypto wallets (addresses that hold tokens) sanctioned or targeted for seizure over links to Iran and its proxies used USDT exclusively or nearly exclusively.

The headline number rests unevenly on two lists: 87% of the 757 wallets designated by Israel’s counter-terror finance bureau transacted mostly in USDT, against 57% of the 101 designated by OFAC, the Treasury office that enforces U.S. sanctions.

Tether, which can lock a wallet so the tokens inside cannot move, published a statement the same day citing roughly $550 million in Iran-linked freezes this year, including $344 million in April and more than $130 million in July. It did not address the report directly.

Blumenthal also wants Howard Lutnick’s own records on Tether from his time at Cantor, including “any discussion of Tether with the White House or state and federal regulators, lobbying on Tether’s behalf.”

Other requests cover Cantor’s know-your-customer policies (the checks a firm runs on who it does business with), whether it has ever weighed ending the relationship, and its communications with the President’s Council of Advisors on Digital Assets.

Cantor has until Oct. 23 to respond, and the letter asks the firm to preserve every document, record and communication tied to its Tether relationship in the meantime.