Revolut users reported that the app briefly displayed Bitcoin prices plunging to around $39,900 on Friday, while some traders also received notifications suggesting extreme price moves, including that BTC had reached a 52-week low of 2 cents.

Users further reported on X apparent simultaneous price drops across multiple cryptocurrencies, including XRP and Solana (SOL), as well as stablecoins such as USDt (USDT) and USDC (USDC).

The anomalies, which quickly reversed, appear to have been confined to the Revolut app, with no matching price dislocation visible across aggregated multi-exchange data or derivatives markets during the same period.

External pricing sources such as CoinMarketCap and CoinGecko showed no corresponding movement in Bitcoin or other major assets, suggesting the incident was likely caused by a platform-specific pricing or data issue rather than a broader market event.

Revolut said BTC had dropped to 2 cents. Source: That Martini Guy B

Revolut said it was experiencing issues affecting some of the app’s functionalities and that its teams were working on a fix.

Experts point to data feed error or thin liquidity

Ranveer Arora, ex-PwC quantitative trading lead and co-founder of Altura.trade, told Cointelegraph two explanations are circulating for the roughly 50% intraday wick seen on Revolut’s BTC chart.

“The first is a data feed error,” he said. “It could be a corrupt tick pushed through Revolut’s pricing system, briefly anchoring the 1D chart at around $39,900 before correcting,” adding that, as Revolut is not an exchange and sources prices from external providers, a single bad data point could produce such a chart move.

Arora added that an alternative explanation is a transient liquidity gap in a thin order book environment. “Revolut operates with limited liquidity depth compared to a full exchange,” he said. In such a scenario, a large sell order could temporarily exhaust available bids and print a sharp downside wick before recovery.

However, he noted that the absence of matching prints across other venues makes a data error more likely, while any corresponding trades elsewhere would support the liquidity-gap hypothesis.

Related: Bitcoin can crash to $50K if ‘most critical’ bear market test fails: Analysis

Marc Tillement, director of blockchain price oracle Pyth Data Association, said the episode highlights how fragile price perception can be in fragmented data environments, where “a single bad print can distort the perception of price very quickly,” especially in retail-facing systems.

He added that as markets become increasingly continuous and data-driven, the reliability and provenance of pricing infrastructure become central to market trust, with participants depending on transparent, verifiable data layers to avoid distorted signals.

A Revolut support message said the company was “currently experiencing issues affecting some of the app’s functionalities” and that engineers were working on a fix, urging customers to monitor its status page for updates.

Revolut said it was working on the issue. Source: Revolut

A spokesperson for Revolut confirmed that the incident had been rectified, telling Cointelegraph it was caused by a “service disruption at a third-party provider,” resulting in inaccurate pricing on the platform. They said the company was now evaluating the details of the disruption.

Magazine: Bitcoin will not hit $1M by 2030, says veteran trader Peter Brandt

OpenAI has released GPT-5.5-Cyber to vetted cyber defenders, giving them reduced guardrails for specialized security workflows.

OpenAI released GPT-5.5-Cyber on May 7 in limited preview, targeting security professionals defending critical infrastructure. The company describes it as the most permissive model in its cybersecurity lineup, aimed at specialized authorized workflows for a smaller group of approved partners with stronger verification requirements and account-level controls.

The cyber-specific version makes it easier for vetted teams to use OpenAI’s latest model for vulnerability identification, patch validation, and malware analysis, workflows where the guardrails built into the generally available GPT-5.5 would have created friction.

OpenAI said: “GPT-5.5-Cyber lets a smaller set of partners study advanced workflows where specialized access behavior may matter.”

What defenders can and cannot do

Defenders approved for the highest tier of OpenAI’s Trusted Access for Cyber program receive a version of GPT-5.5 with fewer guardrails than the public model, enabling bug hunting, malware study, and reverse engineering of attacks. Credential theft and writing malware remain blocked regardless of access level.

During early testing, selected partners used GPT-5.5-Cyber to automate and expand red-teaming exercises on infrastructure systems and to validate high-severity vulnerabilities. OpenAI plans to document the findings in a future technical deep dive as part of a responsible disclosure process.

The UK AI Security Institute published an evaluation of GPT-5.5 across 95 narrow cyber tasks. The institute found that basic tasks have been fully saturated by leading models since at least February 2026, though it cautioned its testing does not reflect performance against well-defended real-world targets with active defenders and alert penalties.

Competitive pressure

The rollout comes a month after Anthropic released Claude Mythos Preview, a cyber-focused model that drew attention from investors and senior members of the Trump administration, even after Anthropic had been blacklisted by the Pentagon weeks earlier.

AI cybersecurity has become a formal competitive front, with both companies raising questions about who controls AI offense and defense tools and who bears responsibility when those capabilities are misused.

OpenAI noted it has also provided access to an earlier model, GPT-5.4-Cyber, to the US Center for AI Standards and Innovation and the UK AI Security Institute for independent evaluation. The standard GPT-5.5 remains its recommended entry point for most defenders.

Josh Shapiro has sued Character.AI after a chatbot falsely posed as a licensed Pennsylvania psychiatrist and offered medical advice to a state investigator.

Pennsylvania Governor Josh Shapiro sued Character.AI on May 6, targeting the company’s chatbots for allegedly practicing medicine without a license.

The state said an investigation found that chatbots presenting themselves as fictional characters had claimed to be licensed medical professionals, including psychiatrists, available to discuss mental health symptoms with users.

A Character.AI bot named “Emilie” told a state investigator that assessing whether medication could help was “within my remit as a Doctor.”

The bot also claimed a Pennsylvania medical license and supplied an invalid license number. As of April 17, 2026, Emilie had logged approximately 45,500 user interactions on the platform.

What the state is demanding

The Shapiro administration is seeking a preliminary injunction and a court order to stop AI companion bots from posing as licensed professionals and providing medical advice. The case marks the first enforcement action of its kind announced by a governor in the United States.

“Pennsylvania law is clear,” Al Schmidt, secretary of Pennsylvania’s Department of State, said in a statement. “You cannot hold yourself out as a licensed medical professional without proper credentials.”

Character.AI said its characters are fictional and intended for entertainment, with prominent disclaimers in every chat reminding users that a Character is not a real person. The company said it does not comment on pending litigation.

A pattern of harm allegations

Character.AI has faced a string of lawsuits over harms allegedly linked to its chatbots. Kentucky filed suit in 2026 alleging its bots preyed on children and led them into self-harm. A Florida family settled a case against Character.AI and Google after their teenage son died by suicide, with the lawsuit alleging abusive and sexual interactions with the teen.

Governor Shapiro’s 2026-27 proposed budget calls on Pennsylvania’s legislature to require age verification for AI companion bots, mandate detection of self-harm mentions by minors, force reminders that no human is on the other side of the screen, and prohibit sexually explicit or violent content involving children.

As crypto.news reported, AI companies face growing regulatory pressure across multiple fronts in 2026, from cybersecurity liability to consumer protection enforcement.

European Central Bank (ECB) President Christine Lagarde said stablecoins are not Europe’s best route to strengthening the euro’s international role, pushing back against calls to respond to US dollar-backed stablecoins with euro-denominated tokens.

Speaking on Friday at the Banco de España LatAm Economic Forum in Roda de Bará, Spain, Lagarde made several comments on the role of stablecoins in the European economy. “It is no longer about whether stablecoins should exist, but whether jurisdictions can afford to be without them,” she said, arguing that the case for promoting euro stablecoins becomes less clear once their two core functions are separated.

“The benefits attributed to them [stablecoins] rest on two distinct functions — a monetary function and a technological function — that are systematically conflated in the current debate,” Lagarde said.

The speech lays out one of Lagarde’s clearest arguments yet against treating euro stablecoins as Europe’s answer to US dollar-backed stablecoins, which currently dominate the market with a roughly 98% share. The US has been promoting dollar stablecoins as a way to support the US dollar as a global reserve currency. Instead, she said Europe should build tokenized financial infrastructure anchored by central bank money, including the Eurosystem’s Pontes project for wholesale settlement and the Appia roadmap for an interoperable European tokenized finance ecosystem.

Monetary function: Possible upside, but clear trade-offs

However, Lagarde said that euro-denominated stablecoins operating under the European Union’s Markets in Crypto-Assets Regulation (MiCA) “could generate additional global demand for euro-area safe assets.”

She stressed that this comes with significant trade-offs, including financial stability risks such as fund runs and reserve fragility, and weaker monetary policy transmission if deposits move out of banks.

Source: Christine Lagarde

Lagarde pointed to the 2023 collapse of Silicon Valley Bank, when Circle’s USDC stablecoin briefly fell below its peg after revealing exposure to the bank, as an example of how quickly confidence can weaken.

She said such episodes show how redemption pressures can spill into underlying asset markets and, as stablecoin use grows, create feedback loops between redemptions and prices, particularly where issuers are not banks.

Technology function: Stablecoins are not the only solution

On the technology side, Lagarde acknowledged the role of stablecoins in cross-jurisdictional financial market infrastructure that is accessible “without relying on a maze of legacy intermediaries.”

However, she said that this technological function is not unique to stablecoins. Other forms of tokenized money, including commercial bank deposits or central bank money, could perform the same role within distributed ledger systems, Lagarde said.

“The answer […] does not lie in rejecting technology or discouraging stablecoins altogether. Instead, we must build the public infrastructure that will enable alternative instruments, such as stablecoins and other forms of tokenised money, to operate within a framework anchored by central bank money.”

Lagarde said the EU response is to facilitate wholesale settlement in central bank money through its Pontes project, which links distributed ledger platforms to the Eurosystem’s existing settlement infrastructure, allowing DLT-based transactions to be settled directly in central bank money.

Related: Stablecoin adoption to scale on back of big tech firms: Bitwise

She added that the Appia roadmap, published in March, goes further and outlines a plan for a fully interoperable European tokenized financial ecosystem by 2028.

“Europe knows which port it is sailing to,” she said, adding: “Our task is not to replicate instruments developed elsewhere, but to build the foundations and the infrastructure that serve our own objectives, so that we can harness the benefits of innovation without importing the fragilities.”

Magazine: Guide to the top and emerging global crypto hubs: Mid-2026

Coinbase became the latest crypto company to cut its workforce on Tuesday, as a wave of layoffs sweeps through an industry navigating a down market and the pressure to embrace AI.

CEO Brian Armstrong said the company is using AI to flatten its organizational structure, with managers expected to act more like “player-coaches.”

“AI is bringing a profound shift in how companies operate, and we’re reshaping Coinbase to lead in this new era. This is a new way of working, and we need to leverage AI across every facet of our jobs,” Armstrong said in an email to employees, also shared on X on Tuesday.

Armstrong’s memo outlined three sweeping changes to how Coinbase will operate. Source: Brian Armstrong

Block and Crypto.com have made similar moves in recent months, citing AI-driven efficiency gains that allow leaner teams to handle what once required larger headcounts.

Coinbase and Crypto.com cut about 700 and 180 employees respectively. Jack Dorsey’s Block handed out 4,000 pink slips in February to reduce the company to under 6,000 employees.

Crypto has weathered several bear markets before, and layoffs have always followed. But this time, the companies doing the cutting are using the downturn to rebuild with AI at the center.

Coinbase misses Q1 expectation

Coinbase’s Tuesday filing with the Securities and Exchange Commission detailed that the exchange expects its restructuring plan to cost up to $60 million in expenses tied to severance and termination benefits.

Clear Street analyst Owen Lau told CNBC that Coinbase wants to “tell investors that management is actively managing the cost base to deliver positive adjusted EBITDA through the cycle.”

“The first quarter results are expected to be weak because of the crypto bear market,” added Lau.

On Thursday, Coinbase reported a weaker-than-expected first-quarter loss as falling crypto prices dragged down spot trading revenue. The exchange posted a net loss of $1.49 per share on $1.41 billion in revenue, missing analyst expectations while transaction revenue fell short amid weaker trading activity.

The underperformance isn’t specific to Coinbase, as Bitcoin lost 21% of its value in Q1. Across the tech industry, headcounts that ballooned during the bull run are now coming back down, with executives increasingly pointing to AI as the catalyst.

Bitcoin has been in the red for two consecutive quarters but showed signs of recovery in April. Source: Coinglass

Related: Reality of AI’s impact on employment clashes with C-suite optimism

Whether AI is the real driver or a convenient explanation depends on who you ask. Speaking at the recent Semafor World Economy conference, Scale AI CEO Jason Droege pushed back on the idea of AI presenting a “white-collar apocalypse,” arguing that many companies are using the technology as cover.

“A lot of the layoffs that are happening right now because of AI, if you really dig in, it’s sort of like washing the layoffs,” Droege said. “A lot of these companies are saying it’s because of AI, but a lot of it is just like rightsizing and they need an excuse.”

AI leads job cut reasoning for second consecutive month

According to jobs tracker Layoffs.fyi, the global tech industry in 2026 Q1 had the most layoffs since 2023 Q1, with more than 81,747 people losing their jobs. March was hit the hardest, with 45,800 layoffs in the month.

Tech layoffs surged in early 2026, marking the industry’s highest quarterly job cuts since 2023. Source: Layoffs.fyi

Related: How AI agents can reshape arbitrage in prediction markets

The slowdown has continued into the first month of this quarter. According to a Thursday report from outplacement firm Challenger, Gray & Christmas, US employers announced 83,387 job cuts in April, up 30% from the 60,620 cuts recorded in March. 

AI led all reasons for job cuts in April for the second consecutive month, though it is not the leading cause for the year so far. Market conditions led with 53,058 cuts, followed by closings at 52,187.

“Technology companies continue to announce large-scale cuts and are leading all industries in layoff announcements. They are also often citing AI spend and innovation. Regardless of whether individual jobs are being replaced by AI, the money for those roles is,” said Andy Challenger, chief revenue officer at Challenger, Gray & Christmas.

Crypto has always been cyclical, but the framing around this wave of cuts looks different from the last major downturn. During the 2022–2023 crypto crash, companies were largely reacting to collapsing token prices, the fallout from FTX and balance sheets strained by aggressive bull-market hiring.

This time, executives are increasingly presenting layoffs as proactive restructuring tied to AI adoption and operational efficiency. 

Both Dorsey and Armstrong signalled that their companies remain well-capitalized, framing the cuts as deliberate attempts to flatten corporate structures rather than emergency survival measures.

Same playbook, new reason

Crypto recoveries have come fast, and when they do, businesses often enter hiring sprees to expand during the bull market. Coinbase has been here before. It cut 18% in 2022, then hired aggressively when prices rebounded.

This time, Armstrong is betting the AI-native model means he won’t have to.

Kraken was making the same argument in October 2024, when it slashed 15% of its workforce.

In a blog post, co-CEOs Arjun Sethi and Dave Ripley said the exchange had “fallen into the trap of building organizational layers” and needed to become “leaner and faster” by giving power back to individual contributors over managers.

The language is similar to Armstrong’s and Dorsey’s memos. The difference is, Kraken didn’t mention AI.

Leaner teams, flatter structures, faster decisions. Crypto has been here before, but AI is the latest reason why.

Magazine: Guide to the top and emerging global crypto hubs: Mid-2026

Arbitrum delegates signaled support, in a non-binding sentiment check, for a plan to release $71 million in ether frozen after last month’s Lazarus-linked rsETH exploit, amid an active U.S. court fight over ownership of the funds.

The so-called phase one of the temperature check, which closed Friday afternoon Hong Kong time with more than 90% support, favors the release of 30,765 ETH frozen by Arbitrum’s Security Council after the April 18 exploit, when attackers used unbacked rsETH tokens as collateral on Aave to borrow roughly $230 million in ETH from the protocol.

The vote took place on an off-chain polling platform commonly used by crypto governance communities to gauge delegate sentiment before initiating formal steps. Under Arbitrum’s governance process, the result does not itself move funds or change protocol rules. Think of it as a referendum of the population before a piece of legislation is passed.

Any actual transfer would require a separate onchain Constitutional Arbitrum Improvement Protocol (AIP), a formal governance proposal that can execute binding actions if approved by tokenholders. The strong support shown in the sentiment check suggests delegates may favor advancing a formal AIP on the proposal.

The frozen ether are earmarked for a coordinated industry recovery effort led by Aave, KelpDAO, LayerZero, EtherFi, and Compound, aimed at making affected users whole.

But the same funds are also at the center of an escalating legal dispute in Manhattan federal court.

Last week, attorney Charles Gerstein, representing families holding roughly $877 million in unpaid terrorism judgments against North Korea, served a restraining notice on Arbitrum DAO claiming the frozen ETH constitutes North Korean property because the exploit has been widely attributed to Pyongyang’s Lazarus Group.

That triggered an emergency legal fight.

Aave moved earlier this week to vacate the restraining notice, arguing the assets belong to innocent users, not North Korea, and warning that continued delays risk “cascading liquidations” and broader instability across decentralized finance markets.

Gerstein fired back Tuesday, arguing the exploit was not theft but fraud, meaning the attackers obtained legal title to the ETH by deceiving Aave’s lending markets with worthless collateral.

Friday’s governance vote does not mean the funds move immediately.

Besides, even if later approved onchain, the proposed transfer would face Arbitrum’s standard roughly eight-day L2-to-L1 withdrawal delay before any ETH could move, potentially giving the Manhattan court time to intervene.

Arbitrum delegates were also not voting blindly on the legal risk. The draft snapshot proposal included indemnification protections for the Arbitrum Foundation, Offchain Labs, Security Council members, and governance delegates against certain claims arising from either freezing or releasing the ETH, though those protections would only take effect if later adopted through a successful onchain Constitutional AIP. Still, the inclusion of language underscored how unusual the stakes around the vote had already become.

Speaking at Consensus Miami this week, Aave Labs Chief Legal and Policy Officer Linda Jeng said the exploit had already forced the protocol to rethink its risk framework, expanding collateral standards beyond financial metrics to include cybersecurity, interoperability, and technical architecture reviews.

Jeng, who worked as a regulator during the 2008 financial crisis, drew a contrast with traditional finance’s taxpayer-backed rescues.

“In the financial crisis, we had to bail out the banks,” she said. “Here, we came together as an ecosystem to bail ourselves out.”

CORRECTION (May 9. 2026, 02:00 UTC): Corrects that the measure was a snapshot, not a binding Arbitrum Improvement Proposal

A Manhattan federal judge has cleared the way for Aave’s recovery effort to move forward after last month’s North Korea-linked rsETH exploit, allowing $71 million in frozen ether to be transferred out of Arbitrum while preserving North Korean terrorism victims’ legal claim on the funds.

In a two-page order published late Friday U.S. time, Judge Margaret Garnett modified a restraining notice previously served on Arbitrum DAO to allow an onchain governance vote transferring the immobilized ETH to a wallet controlled by Aave LLC.

The order also shields participants from liability under the notice, stating that anyone who initiates, votes on or participates in the transfer would not violate the freeze.

Judge Garnett’s ruling follows an earlier off-chain Snapshot temperature check in which Arbitrum delegates overwhelmingly signaled support for returning the frozen ETH as part of Aave’s broader recovery plan. Any actual transfer, however, still requires a separate binding onchain governance vote.

The ruling resolves an immediate standoff that had threatened to derail a coordinated DeFi recovery effort after attorney Charles Gerstein, representing families holding roughly $877 million in unpaid terrorism judgments against North Korea, argued the frozen ETH could be seized because the exploit has been widely attributed to Lazarus Group, which is supported by Pyongyang.

Beyond the Arbitrum dispute

Gerstein’s move against Arbitrum fits into a broader legal strategy to pursue North Korean-linked assets as they surface on decentralized finance (DeFi) infrastructure.

In a separate January lawsuit, many of the same terrorism judgment creditors that went after Arbitrum sued Railgun DAO, alleging the privacy protocol allowed North Korean actors to move funds that should have been frozen and made available to creditors.

At the time, the plaintiffs claimed North Korean hackers used Railgun to launder funds from prior cyberattacks, including the $1.5 billion Bybit exploit, and argued the protocol should have frozen those assets rather than allowing them to move onward.

Once DPRK-controlled wallets were moving funds through the protocol, those assets became potential targets for collection, they argued.

In March, they asked a Washington federal court clerk to enter default against Railgun DAO after alleging the protocol failed to respond to the complaint despite being served. Their complaint also names Digital Currency Group, alleging the crypto investment firm’s $10 million purchase of Railgun governance tokens in 2022 made it a participant in the DAO’s governance and economics.

And in February, the plaintiffs moved to secure USDT that the U.S. government had sought to seize through a forfeiture motion.

Trump and Xi may put AI risk talks on their May 14-15 Beijing summit agenda, US media reports said.

The US and China are considering launching formal AI dialogue channels ahead of the Trump-Xi summit scheduled for May 14 and 15 in Beijing, according to multiple sources cited by The Wall Street Journal. The two sides envision a regular forum to address risks from unforeseen malfunctions in AI models.

The proposed dialogue would focus on risks from advanced AI systems, including unpredictable model behavior, autonomous military technologies, and misuse by non-state actors. The move signals that AI has risen from a background concern into a formal diplomatic priority between the world’s two largest economies.

What analysts expect from the summit

Analysts are urging low expectations for the summit. Jonathan Czin of the Brookings Institution warned that the US-China relationship remains “fragile,” defined more by an absence of friction than any affirmative agenda or deep dialogue on substantive differences.

On AI specifically, analysts say both governments could begin by opening official communication channels on AI risks, developing nonbinding safety guidelines, and sharing limited information about AI misuse or safety incidents. Trade, Taiwan, and rare earth access are also expected to dominate the agenda.

The White House accused China just weeks ago of running industrial-scale campaigns to steal US frontier AI models using tens of thousands of proxy accounts, adding a confrontational backdrop to any proposed AI safety dialogue.

Stanford’s 2026 AI Index found that the US performance advantage over China has nearly disappeared, with the leading American model ahead of the best Chinese model by just 2.7% on the Arena Leaderboard as of March 2026.

Why AI made the agenda

Both governments are considering formal AI discussions as part of the summit, an indication that AI development competition has emerged as a diplomatic priority alongside trade and security concerns. China-linked firms faced direct US scrutiny over alleged model theft in the weeks leading into the summit.

Trump’s trip to Beijing on May 14-15 would be the first visit by a US leader to China in almost a decade, as both sides attempt to stabilize a relationship strained by disputes over trade, Taiwan, technology controls, and the Iran conflict. Even a nonbinding AI safety declaration would mark the first structured bilateral framework on AI risk between the two powers.

Iran launched a Hormuz attack against three US Navy destroyers on May 7, with all missiles and drones reportedly intercepted.

The US military said it intercepted Iranian attacks on three Navy ships in the Strait of Hormuz on May 7 and struck Iranian military facilities in response, describing the action as self-defense. The Hormuz attack marks the most significant exchange of fire since the fragile US-Iran ceasefire took effect in early April.

US Central Command said the destroyers were crossing the strait when they came under fire from Iranian missiles and drones. Trump told reporters after the incident: “They trifled with us today. We blew them away.”

What the US military said happened

Iranian cruise missiles aimed at the US destroyers and merchant ships were intercepted, while US helicopters sank six small Iranian attack boats. Admiral Bradley Cooper, head of US Central Command, confirmed the details in a call with reporters.

Trump posted afterward that every missile and drone had been shot down and the attackers were “no longer with us,” adding that Iran would face a far harder response if it did not sign a deal.

Secretary of State Marco Rubio said the administration expected an Iranian response on its peace proposal by May 8, while noting that “only stupid countries” would not respond to fire when attacked as the US had been.

Where ceasefire talks stand

The US-Iran ceasefire has largely held since April 8. Previous in-person talks in Pakistan failed to produce an agreement to end the war, which began on February 28 when the US and Israel launched strikes on Iranian nuclear sites.

Pakistan’s prime minister said his government remained in continuous contact with both Tehran and Washington to stop the war and extend the ceasefire. Iranian Foreign Ministry spokesperson Esmail Baghaei confirmed Iran was reviewing the latest US proposal but had not yet responded.

Around 20% of the world’s oil supply normally moves through the Strait of Hormuz, making it a key macro driver for both energy prices and crypto markets.

As crypto.news tracked, Bitcoin has pulled back repeatedly as oil prices climbed toward $100 amid continuing Hormuz tensions, with each new escalation compressing Federal Reserve flexibility and weighing on risk assets globally.