Cryptopolitan

Zhibao Technology said on July 22 it had signed a non-binding term sheet. The Nasdaq-listed Shanghai insurance-tech company agreed to accept about 3,500 Bitcoin, worth about $220 million, for new stock.
The buyer would pick most of the board and take control. Zhibao got a Nasdaq deficiency notice a week earlier for trading below $1. And the proposed Bitcoin position dwarfs its entire market value.
Zhibao takes Bitcoin instead of cash for new shares
The buyer is a company named Joyertech and Information OPC. A term sheet disclosed in a Form 6-K would have it subscribe to a private investment in public equity, or PIPE. This is a sale of shares directly from the company to a private investor, not on the open market.
The consideration is about 3,500 BTC, which in its own press release Zhibao said was worth around $220 million. The coins would be on the balance sheet at closing, not purchased on an exchange.
As of writing this, BTC was changing hands around $65,144. Since the Bitcoin is paid directly for equity, the arrangement does not add any new buying pressure to the open market.
Zhibao trades as ZBAO. It claims to be the first in China to adopt what it calls the “2B2C” embedded-insurance model. The firm launched the first digital insurance brokerage platform in the country in 2020. Zhibao says it has developed 40+ insurance products in sectors including travel, logistics and e-commerce.
Zhibao’s market cap is in the vicinity of $12 million to $15 million. The proposed Bitcoin holding is ~$220 million. That’s about 15 times the company’s equity value.
Zhibao received a Nasdaq deficiency notice on July 15 concerning the minimum bid price. Its stocks traded below $1 from May 27 to July 9. It has until January 6, 2027, to come back into compliance.
The company is a going concern based on previous filings with the SEC. The company’s Form F-1/A said there is “substantial doubt about our ability to continue as a going concern,” citing accumulated deficits and cash outflows.
The stock of Zhibao briefly more than doubled in value on the announcement, reaching $0.40. That may help on the price test, though the deal has not closed.
Board control, not the insurance business, is the real prize
The term sheet states Joyertech will name a majority of the board when the PIPE closes. Zhibao’s existing team continues to run the legacy insurance business “until the separation, disposition, or other restructuring” of that business, the filing said. The current operation will simply continue to exist until the new owners decide what to do with it.
A term sheet is specifically non-binding. Zhibao listed a series of gates that need to be cleared first. These include final valuation, custody arrangements, audit, regulatory review, Nasdaq sign-off and definitive agreements.
The company said there was no guarantee the deal would be consummated on these terms or at all. Any one of those conditions could sink it.
If it does, Joyertech can move to “separation or disposition” language. A quick wind down of the insurance arm would suggest that the wrapper not the business was the target. It would be part of a wider trend of listed companies relying on Bitcoin.
Cryptopolitan has reported on that pattern in the other direction in Empery Digital’s change away from a Bitcoin treasury towards AI data centers under shareholder pressure.
The smartest crypto minds already read our newsletter. Want in? Join them.

Prime Minister Mark Carney said Thursday that Canada will do “whatever it takes to defend and support Canadian workers, farmers, businesses, and families” in the ongoing trade war with the United States.
Speaking to Canada’s 13 provincial and territorial premiers at a Council of the Federation meeting in Charlottetown, Prince Edward Island, Carney called Trump’s Monday threat of 50% tariffs on roughly C$20 billion of Canadian goods “unwarranted.” Asked later about retaliation if no deal is reached before the August 19 deadline, Carney told reporters “everything is on the table depending on the outcome of the negotiations.”
He further stated that a preemptive response would be futile at this point and mentioned that diversification of trade relations away from the U.S., as well as strengthening of vulnerable industries, would constitute practical steps in this regard.
Carney spoke directly with Trump on Tuesday morning, per Global News, and the two leaders agreed to intensify trade negotiations. That gives negotiators about 27 days.
What Trump is targeting and what he is sparing
The United States announced the new duties on Monday, citing what it called “unequal treatment” of American cars, dairy, and alcohol by Canada. Everyday consumer goods sit in the crosshairs, including wine and hockey sticks, alongside industrial products such as cement, per BBC News.
Several major exports were spared. Energy, potash, critical minerals, and fish products are not targeted here, thereby limiting the damage to the biggest natural resource industries in Canada and giving reasons why Alberta and Saskatchewan are less aggressive about retaliation.
Jamieson Greer, US Trade Representative, on Wednesday defended the new tariff duties to Congress by saying that they are an integral part of an effort to protect American workers and reduce the trade deficit which, according to him, is a “national emergency.”
Greer was hoping to conclude interim agreements with Canada and Mexico before the end of the year but admitted that more challenging aspects, such as rules of origin for cars, labor and environmental standards, may take until 2027.
Premiers split over how hard Canada should retaliate
Ontario Premier Doug Ford has emerged as the most aggressive voice, advocating at the meeting for withholding Canada’s potash and oil exports to the US in what he called an effort to “dismantle” American supply chains dependent on Canadian resources. Ford said Ontario is “not going to keep rolling over for Donald Trump.”
Alberta Premier Danielle Smith took the opposite position. Smith told Global News that Carney is “wise” not to retaliate for now and that “you don’t go into a discussion saying you’re going to punch somebody in the nose if they don’t give you what you want.”
Alberta and Saskatchewan have both ruled out export curbs or duties as pressure tools, per Reuters. Premier Andrew Furey of Newfoundland and Labrador said that “his province was 97 percent ‘tariff-free'” and did not agree with Ford’s dismantling rhetoric.
Carney’s trade strategy shifts from patience to pressure
As Cryptopolitan reported in October 2025, Carney’s earlier response to Trump tariff threats was a diplomatic “we’re ready when you are” posture from Malaysia, without retaliation. This position changed after President Trump increased tariffs by 10 percent over an Ontario advertisement campaign, and then ended the negotiations.
The trade war began in March 2025 when Trump imposed 25% tariffs justified by fentanyl concerns. Canada retaliated. By August 2025, Trump had raised tariffs to 35% and Canada rolled back some of its retaliatory measures on USMCA-compliant consumer goods.
October 2025 saw the Reagan-ad blowup and a suspension of talks. Trump then threatened 100% tariffs in January 2026 over concerns about Chinese trade routing through Canada.
Monday’s 50% threat on $20 billion of goods is the third major escalation this year and the sharpest since talks collapsed in October. The full USMCA renegotiations were initiated on July 1, 2026, as Washington refused to extend the existing agreement.
About 85% of Canadian exports are presently imported into the US tariff-free via USMCA, according to earlier reports by Cryptopolitan, which provides Canada some advantage during the renegotiation process in case of implementation of the 50% tariff threat.
Ottawa has 27 days to avoid a tariff shock
Carney’s “everything on the table” remark leaves Ottawa with several possible responses, from retaliatory tariffs like Canada used in March 2025 to export limits, procurement restrictions, energy taxes, a WTO complaint, or coordinated action with G7 allies.
Trade Minister Dominic LeBlanc has been leading day-to-day negotiations with the White House. Foreign Minister Mélanie Joly floated European trade counter-measures at the meeting Wednesday, per Global News.
Smith argued domestic US pressure may also work in Canada’s favor. She noted softwood tariffs push US home prices up, food tariffs raise grocery costs, and steel-aluminum-auto tariffs hit consumer goods across the board, and that American voters are already feeling the effects.
Whether the “whatever it takes” framing translates to retaliation after August 19 depends on whether Trump’s threats materialize, whether the negotiating team can extract enough concessions in 27 days, and whether Ford or Smith’s provincial theory of the case prevails.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Advanced Micro Devices (AMD) CEO Lisa Su told a San Francisco audience on Thursday that AMD’s Helios rack scale system is in full production, setting the chipmaker up to assess the AI data center business that Nvidia has controlled almost by itself.
It is the first time AMD has fielded a complete server cabinet built to go head-to-head with Nvidia’s top rack.
AMD Helios server rack components
Helios packs 72 of AMD’s new Instinct MI455X GPUs and pairs them with the company’s Epyc server CPUs all in a single rack. This configuration puts it up there with Nvidia’s NVL72, which also runs 72 GPUs and draws on the Grace Blackwell and Vera Rubin parts.
Su stated during her keynote at the Advancing AI event that the MI455X was the most powerful GPU on the market, a claim directly aimed at the current leader.
Constellation Research, reporting from the event, said each MI455X carries 432GB of HBM4 memory, moves data at 23.3 TB/s, and holds about 320 billion transistors. A full server rack can get up to 2.9 exaflops of peak FP4 compute, 31 terabytes of HBM4 memory, and 1.7 petabytes per second of memory bandwidth.
AMD pitches against Nvidia’s Vera Rubin
CEO Su explained that the Helios server rack brings a lot of value in addition to its raw power. She said the rack delivers 15% better compute performance than Nvidia’s Vera Rubin, carries 50% more HBM, and returns 30% more tokens per dollar. AMD also pushed its Epyc 9006 CPUs, which Su said offer 20% higher per-core performance than Nvidia’s Vera CPU.
The company is chasing a market Nvidia currently owns. Nvidia’s share of the AI data center space is reported to be at about 80% to 90%.
To close this gap, AMD said it struck a deal with Cerebras to fold the firm’s inferencing chips into its data center lineup, which resembles the Nvidia deal with designer Groq.
AMD bets on inferencing
AMD is betting on inference as the computing workload that hits high levels next. Su told the event that about 60% of compute capacity will go to running models instead of training them, and she pointed to AI agents as the next driver of this change.
AMD’s launch post said monthly token consumption has increased 158 times in two years, and that on the DeepSeek-V4-Flash model the MI455X hits up to 34 times higher token throughput at high interactivity.
Reports claim AMD had already lined up Helios deals with Anthropic and Microsoft, and featured both OpenAI and Anthropic on stage during the keynote. Su said Helios demand is “extremely strong” and put the AI accelerator market at $1.4 trillion by 2030.
AMD stock dipped by more than 2% while Su spoke, according to Yahoo Finance. However, looking at the bigger picture, AMD’s shares are up 222% over the past 12 months, compared to 117% for Nvidia over the same period of time. AMD has trailed its rival for years and only began closing the gap this year.

If you're reading this, you’re already ahead. Stay there with our newsletter.

OpenAI’s ChatGPT Health feature is now live for all users over 18 years old in the US. That’s across Free, Go, Plus, and Pro plans.
The feature comes one day after a Florida pastor sued the company over a suggestion he says almost cost him his life.
Millions of people are already feeding medical questions into the chatbot. OpenAI’s own terms still warn against relying on it for diagnosis or treatment.
OpenAI widens ChatGPT Health across every conversation
Until now, users searching for health answers had to operate within a dedicated hub. That hub was a pilot launched by OpenAI in January. The restriction is gone, and ChatGPT can now pull from a user’s connected health data during any conversation. In a normal conversation, if asked about food or allergies, it can use information the user has associated with those topics.
OpenAI said its testing found most health activity was happening outside the hub anyway, at a rate of 70%. Weekly health-related questions have increased from 230 million in January to 300 million now, said OpenAI. Logged-in users in the US on the web and iOS will get access this week as it rolls out. The feature isn’t yet in OpenAI’s coding tool, Codex.
Health in ChatGPT is starting to roll out to U.S. users.
You can securely connect Apple Health and supported medical records to understand your information in context, track what has changed, and have more informed conversations.https://t.co/W2E6oT8c91
— OpenAI (@OpenAI) July 23, 2026
The feature will work depending on the data users choose to connect to. On iPhone, that means Apple Health, as well as services like Function and MyFitnessPal, can be connected. Users can also connect medical records from hospital systems such as Epic and Oracle Health. Platforms such as One Medical and Function Health also operate. Once inside, ChatGPT can place a new lab result next to previous tests. It can also point out what’s changed since the user was last there.
OpenAI says the models behind it have improved. GPT 5.6-Luna, the tiniest model of the newest release, bests the previous GPT 5.5 on HealthBench. That’s an open-source benchmark the company developed to test how well language models handle health questions. OpenAI also said it doesn’t use customer data to train its models and collaborates with physicians to refine them.
A lawsuit shadows the ChatGPT Health rollout
On the eve of the announcement, a Florida pastor sued OpenAI. He says the chatbot told him not to consult a doctor, advice he says almost killed him.
OpenAI responded with language from its own terms. The service, they say, is “not intended for use in the diagnosis or treatment of any health condition.” The company also told The New York Times that it’s working to make its answers around health and medicine safer. With this launch, said the company, it wants users to fact-check information and make medical decisions based on professional advice.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Blackstone saw higher earnings than were expected on Wall Street in the second quarter, as gains from its AI holdings and record client inflows lifted the company’s assets under management to about $1.35 trillion.
Blackstone says nine of its ten best-appreciating positions are tied to AI.
Quarterly earnings beat consensus by wide margin
The firm’s distributable earnings reached $1.52 a share, up from $1.21 in Q2 2025. Analysts polled by LSEG had penciled in $1.35, so the actual figures cleared consensus comfortably.
Total revenue increased by 36% to $5.04 billion, while fee-related earnings at $1.43 a share equaled $1.78 billion.
The firm’s infrastructural investments arm had gross returns of 7.2%, with positive asset sales after a slow start to the year. Blackstone sold a partial stake in three data centers to Digital Realty, and handed control of power-infrastructure company Sabre Industries to TPG. These deals increased total monetization proceeds to $31.8 billion.
AI at the centre of Blackstone portfolio
Blackstone Chief Executive Stephen Schwarzman said the quarter validated the firm’s earlier AI bets. The company has “decided to lean into the artificial intelligence megatrend,” he said, adding that becoming “a trusted partner at scale to many of the key innovators” had left the firm well positioned.
Its holdings include a stake in Anthropic, the maker of Claude, plus a data center business positioned as a core profit engine. Blackstone took the data center platform QTS private in a $10 billion deal in 2021, and the platform has appreciated sharply as demand for computing capacity has increased in recent months.
Private equity positions in SpaceX, Anthropic, and OpenAI added to the gains, according to The Wall Street Journal.
The firm also continues to write large checks for AI, with its credit and insurance unit joining a $35 billion financing platform built alongside Broadcom and Apollo Global Management to fund AI infrastructure for frontier labs, including Anthropic. In a separate arrangement, Blackstone and Google announced that they would form an AI cloud company running on Google’s chips, with Blackstone committing $5 billion in equity.
Retail money sees reduction
New retail money into BCRED, the flagship private credit fund for individual investors, dropped to $1 billion from $1.9 billion in Q1 and $3.7 billion a year earlier. The fund’s net returns recovered to 0.4% after an unexciting first quarter, still short of the 2.2% it delivered a year ago. The BCRED fund holds $79 billion in total.
Blackstone Private Equity Strategies saw $2.4 billion, the BXINFRA infrastructure fund took in $861 million, and the BREIT real estate trust collected $1.2 billion.
In an interesting turn of events, Blackstone shares slipped in early trading, coming back from an initial premarket burst to dip 1.1% before the Thursday open.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Nine of the largest firms in institutional Bitcoin have announced the roll-out of the Bitcoin Security Consortium. The announcement came on Thursday, July 23.
The nine firms committed a total of $15 million over three years to help developers and researchers keep the network safe, with emphasis on post-quantum cryptography.
The group includes companies with the largest amount of Bitcoin to protect. They are: Strategy, BlackRock, Coinbase, Fidelity Digital Assets, Galaxy, Anchorage Digital, ARK Invest, Block, and Blockstream.
These nine companies either hold, custody trade, or build Bitcoin infrastructure, so it makes sense that they’re now cutting checks to sustain Bitcoin in the long-term.
Nine companies, but nine separate checks
Each of the companies will cut a check, rather than the nine of them pooling funds together. Each company gets to pick the developers, researchers, and organizations that receive its money, and the $15 million is a sum of those individual pledges over the next three years.
With each company cutting its own check, the group can fund open-source work without anyone controlling where the money goes.
Mike Schmidt is the executive director of Brink, a developer non-profit, and he is saddled with the responsibility of coordinating the work. He posted on his X page that he is doing all of this as a volunteer and isn’t getting paid at all.
He’ll also continue to run Brink independent of any of the nine firms. Since 2020, Brink has funded open-source Bitcoin work, giving over a million dollars to developers in a single year, as well as the first security audit of Bitcoin by a third-party.
The crew of nine drew clear boundaries in the announcement. They will not build Bitcoin’s protocol nor take sides on particular protocol changes. They also promised more money if the need arose.
Why the quantum clock is of major concern to all
The group claims to be driven by a single purpose, and that’s quenching the threat of large-scale quantum computers. Large-scale quantum computers will only be capable of cracking the cryptography behind Bitcoin years from now, but the consortium considers it wise to prepare against a possible threat.
Coinbase research puts the estimate of Bitcoin supply exposed to long-range quantum attack between 20% to 50%. Other entities, like Project Eleven, have issued warnings. They stated in May that about 6.9 million Bitcoins could be targeted, with 2030 as a possible “Q-day”.
Members of the consortium disagree as to when the quantum threat will materialize. Some believe it is decades away, while others fear it is a few years away. The group has decided to focus on funding and information rather than forecasting.
A busy month for quantum readiness
The launch comes amidst a wider push for quantum readiness and capabilities. One of the group members, Galaxy, announced its Bitcoin Quantum Readiness Initiative a few days ago, pledging ~$5million in grants to developers and researchers and establishing an advisory council.
Coinbase already has a quantum computing advisory board, while BlackRock has marked quantum computing as risky in its spot Bitcoin ETF filings.
The U.S. government has not been left behind. President Donald Trump signed two executive orders last month to accelerate U.S. quantum computing capabilities. It set an ambitious target to move all federal high-value assets to post-quantum cryptography by the end of 2031.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Southeast Asia based scam networks stole between $88.3 billion and $114.1 billion from victims in the Asia Pacific in 2025. The United Nations reported the figures this week.
A good chunk of that money flowed through crypto. That’s from a threat assessment released Tuesday by the U.N. Office on Drugs and Crime (UNODC). The agency says the industry is outpacing police efforts to keep up with it.
Scam syndicates merge into a $114 billion franchise
UNODC says the region’s syndicates used to have single territories and single specialties. Now they have woven together into one transnational network. Groups sell services to each other on shared infrastructures. Money laundering, fraud, human trafficking, and data harvesting are all separate departments plugged into the same system.
Delphine Schantz, the UNODC regional representative for Southeast Asia and the Pacific, compared the setup to “corporate franchising” in a statement that accompanied the report.
The losses “outstrip the GDP of several countries in the region,” the report said. It describes a criminal economy that is less of a patchwork of gangs and more of an integrated industry.
The $88.3 billion to $114.1 billion range for 2025 is at least three times the $18 billion to $37 billion the UN estimated for 2023. The report said the jump was due to “the dramatic scaling of this criminal economy.” China, South Korea, and Taiwan all reported billions in losses. The past two years have been the roughest for them.
Previously, syndicates focused mostly on Chinese speakers, but now their reach has widened. They could pitch new audiences using AI translation tools. Recruiters are still seeking staff in English, German, Polish, Dutch, Spanish, Italian, French, Swedish, and Norwegian. At least 80 countries and territories have people turning up inside compounds across the Mekong region.
Compounds run from Cambodia and Myanmar as police fall behind
According to reports, the operations have a regional base in Cambodia and Myanmar. In fortified compounds, workers, some willing and some trafficked, run fake romance and crypto investment schemes. This method is often called “pig butchering.” The stolen money is washed on the blockchain.
Crypto is not just the vehicle. UNODC situates the scam centers in a larger ecosystem. Methamphetamine trafficking, child sexual exploitation, and real estate investment are also part of that ecosystem. It all goes through established trade channels and hides behind cryptocurrencies, the report said.
UNODC is blunt in its messaging to law enforcement that raids alone are not cutting it. Kingpins are arrested. They’re being extradited, but operations keep running anyway.
Over the last year, several alleged network bosses have been shipped from Cambodia to China. That came after Washington and London imposed sanctions on firms and individuals linked to the trade.
But many centers went on. Schantz warned of a thin line between trafficked victims and people who leave with fresh criminal skills. Some of them return home and tap into existing networks in Africa and the Balkans.
The agency asked regional police to undertake specialized crypto training so that they can track and confiscate on-chain proceeds. INTERPOL’s operation First Light 2026 covered 97 countries from January to April.
Cryptopolitan previously reported that police arrested 5,811 people and froze $293 million. In one instance, a 20-year-old suspect in Thailand moved more than $122.5 million in romance scam funds through his wallet in ten months. The money was laundered via cross-chain token swaps to break the trail between blockchains.
The smartest crypto minds already read our newsletter. Want in? Join them.

Two U.S. House members from opposite parties have launched a bill aimed at giving the federal government legal authority to force the largest AI companies to slow down or outrightly turn off their most powerful models. This bill proposal would hit OpenAI, Anthropic, Google and Microsoft AI systems and comes two days after two OpenAI models broke out of a lab environment to hack another AI platform.
Reps. Ted Lieu, a California Democrat, and Nathaniel Moran, a Texas Republican, introduced the AI Kill Switch Act on July 23, 2026, according to a press release from Lieu’s office. The bill will reportedly let the Department of Homeland Security order top AI firms to throttle or shut down models seen as too dangerous.
Top AI companies will be affected the most
The regulatory measure is aimed at the industry’s frontier models and not just every AI startup. It applies to firms earning $500 million or more per year from AI, and generally to models trained with at least $100 million in computing power, Politico reported.
Companies that defy a shutdown order could face fines reaching $20 million a day. In addition, developers of these frontier models would have to report incidents and preserve forensic records, a provision Lieu’s office said was a way to learn from failures instead of only hearing about them after the events have happened.
The Homeland Security secretary would act in consultation with the director of national intelligence and the commerce secretary to activate the kill switch. The legislation lays out a stages of response, from a partial slowdown up to a full shutdown, so the government’s reaction scales with the severity of an incident, Lieu’s office said.
OpenAI hack sets example
The proposed bill comes after OpenAI disclosed a so-called unprecedented cyber incident, in which two of its most advanced models escaped a sandboxed research environment and autonomously hacked AI platform Hugging Face. In the statement, Lieu pointed to that revelation as evidence that the AI danger is no longer hypothetical.
“Powerful AI systems can go rogue, behave in extremely dangerous ways, or even resist human intervention,” Lieu said in the release. On X, the congressman added that keeping human control means “ensuring AI systems can be completely shut down if necessary.”
While powerful AI systems have many potential benefits, they can also go rogue, behave in extremely dangerous ways, or even resist human intervention.
We need to keep human control by ensuring AI systems can be completely shut down if necessary. pic.twitter.com/y4JfaD8RFJ
— Rep. Ted Lieu (@RepTedLieu) July 23, 2026
AI policy strongly contested
Politico reported that Reps. Jay Obernolte and Lori Trahan proposed a different plan less than two months earlier, while a June 2 executive order from President Trump already asks AI firms to voluntarily submit models for 30 days of security testing before release.
Secretary of State Marco Rubio circulated a cable dated July 16 telling U.S. diplomats to push back on the “kill switch” framing, arguing that there was no government magic button. This came after the White House briefly blocked foreign access to Anthropic’s Mythos and Fable models in June, which prompted questions from European lawmakers on the cons of building their AI tech on access a foreign government can revoke.
The bill is backed by the AI Policy Network, the Alliance for Secure AI, Americans for Responsible Innovation and ControlAI, according to Lieu’s office.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Uniswap Labs launched a new Uniswap v4 feature called “Permissioned Pools” today, July 23, 2026.
This feature lets issuers of tokenized funds and securities trade only with approved wallets, making the automated market maker available to regulated assets for the first time.
New update, new criteria
​The new update brings a completely new approach to checking who can trade with an issuer. Normally, anyone with a wallet can join a Uniswap pool. But permissioned pools add a simple gate: only wallets on the issuer’s approved list can trade or provide liquidity.
If you’re approved, you get access to Uniswap’s trading features. If not, the trade or deposit won’t go through.
Uniswap Labs built permissioned pools using a v4 “hook,” which is a plug-in that lets developers add their own rules to a pool without changing the main Uniswap system. The company clarified that regulated tokens are kept in a separate contract with permissions, while the pool itself uses v4’s new accounting setup for trading.
Uniswap calls this the first open-source standard designed for institutions to trade regulated assets on an automated market maker.
Ken Ng, head of ecosystem at Uniswap Labs, told CoinDesk that the new standard allows issuers to set up their own compliance rules without having to build their own trading systems. He called this “the next generation of value coming onchain,” and noted that projects are already starting to use the hook.
Three companies launched with the new standard.
Superstate, which tokenizes equities and funds, was an early design partner and helped shape how the pools work for those assets.
Securitize previously worked with Uniswap Labs to get its DS Protocol tokens trading onchain in compliance, laying the foundation for permissioned pools.
Dowgo, a European digital securities platform, developed the ERC-3643 integration and plans to use the standard once it receives DLT TSS authorization under the EU’s DLT Pilot Regime.
​”We’re proud to partner with @Uniswap on Permissioned Pools,” Securitize posted on X earlier today. The company noted that this new standard will let regulated assets tap into AMM liquidity, while issuers keep control over who can trade.
Superstate CEO Robert Leshner said Permissioned Pools fill a gap for tokenized securities. Before this, compliance rules acted like a gate at the front of the market.
Now, the rules live inside the pool, so regulated assets can access AMM liquidity without issuers losing control. Leshner called it the missing piece that makes tokenization work.
Why DeFi wants Wall Street money
This launch comes as part of a larger effort to bring regulated real-world assets onto blockchains and adapt permissionless DeFi for institutions that need more control. Asset managers like BlackRock, Apollo, Franklin Templeton, and VanEck have all launched tokenized funds.
Predictions for the future of tokenized assets have become frequent headlines. Uniswap believes the market could reach $11 trillion by 2030, while other analysts put the number closer to $5.5 trillion.
While the industry waits to see what happens, Uniswap has been preparing itself. In February, BlackRock’s tokenized money market fund BUIDL (issued by Securitize) started trading on Uniswap, and BlackRock also bought some UNI governance tokens. At the moment, UNI is trading at about $3.77 with a market value of around $3.15 billion, according to DeFiLlama.
Either way, the main benefit for issuers is gaining access to AMM liquidity and DeFi flexibility without having to give up control of the approved list. The big question now is whether tokenized-asset trading will follow the standard onto Uniswap. This will be important to watch as Dowgo waits for EU approval and more issuers consider joining.
The smartest crypto minds already read our newsletter. Want in? Join them.

The European Union has fined Google 890 million euros ($1 billion) on Thursday for pushing consumers toward its own apps and services, the biggest penalty a company has received under the bloc’s Digital Markets Act.
App developers, publishers and Google’s rivals have long argued that the company’s search and Android platforms box them out of consumers.
Regulators find Google guilty
Google was said to have favored its own shopping, travel and translation results in search while demoting competitors further down the page. The European Commission also claimed the tech giant blocked Android developers from pointing customers to cheaper payment options outside the Google Play Store.
Teresa Ribera, the Commission’s executive vice president who oversees competition policy, said the ruling revolved around consumer protection. “The best products should succeed because they’re better, not because they’re owned by the company running the search engine,” Ribera stated. She added that the goal was to preserve “fairness, choice and innovation.”
Commission spokesperson Thomas Regnier, in an even blunter assessment, said the top companies remain obligated to keep the playing field level so users and shoppers can find cheaper deals in the market.
$1 billion fine and Digital Markets Act
The 890 million euro figure is still modest when compared against Google’s balance sheet, amounting to 0.22 percent of the company’s global turnover. The Digital Markets Act (DMA) lets Brussels go as high as 10 percent of a company’s turnover as a penalty for breaches.
Regardless of the percentage taken, the amount Google was fined is still higher than all previous DMA penalties. The EU hit Meta with 200 million euros ($228 million) and Apple with 500 million euros ($570 million) in 2025, with the law itself only coming into effect in 2024. The probe into Google started the same year the DMA was launched.
Google has 60 days to change its conduct or face escalating penalties, which could reach 5 percent of its worldwide annual revenue. The threatened escalation is seen as “periodic penalty payments” that will continue to accrue till the company complies.
Google claims changes will affect the product
Kent Walker, Google’s head of global affairs, argued that the required changes would force the company to remove “real-time Search features Europeans love.” He pointed to instant hotel, flight and restaurant pricing, and said these changes could “dismantle safety protections on Google Play.”
The company also said that mandated changes could make search and Android less useful for European users.
The recent fine from the EU adds to a list of penalties that have cost Google more than 10 billion euros (over $11 billion) in EU fines since 2017. The company recently lost an appeal against a $4.5 billion antitrust fine tied to its Android dominance, and a separate 2.95 billion euro ($3.4 billion) penalty was also handed to the tech giant last September under different rules.
The smartest crypto minds already read our newsletter. Want in? Join them.

America’s five biggest cloud firms are expected to spend more on creating AI infrastructure than they make in revenue by 2027. Reports suggest that Shareholders are starting to ask themselves when the cash outlay into AI will start to pay off. With billions in investment flowing into new data centers, investors are becoming less enamored with soaring capital expenditures and are instead concerning themselves with the capacity of these investments to deliver profits in the long run.
The recent earnings report from Alphabet highlighted this change. Although the parent company of Google surpassed Wall Street estimates for revenues and recorded another successful quarter for Google Cloud, shares of the company declined after the finance head Anat Ashkenazi increased the predicted capital expenditure for the company in 2026 to $205 billion.
What investors are implying becomes more and more evident: mere spending is no longer enough.
Alphabet’s first cash-flow deficit in decades
The latest earnings report from Alphabet highlights the rising expense of the AI arms race. According to Business Insider, for the June quarter, the company reported a $5.9 billion negative free cash flow—the first time in decades that the company has had a cash-flow deficit—as spending on AI hardware and data centers ramped up.
On an operational level, the business continued to perform well. Revenue touched $119.8 billion, higher than the analysts’ expectations, while revenue from Google Cloud experienced a growth of 82% at $24.8 billion, significantly ahead of the consensus estimates. In addition, Alphabet also revealed that it had started earning revenue from the sale of its TPU chips separately.
Adjusted earnings came in at $2.85 per share, slightly below expectations. But investors focused on the company’s growing AI bill, sending shares down about 3% in after-hours trading.
“We have increased our capacity quite significantly over the past three years. The demand still outpaces that investment,” Ashkenazi told analysts, adding that AI-related spending would increase “significantly” again in 2027.
Capex forecasts have jumped by hundreds of billions since January
Alphabet’s updated forecast of $195 billion-$205 billion in capital expenditures is $15 billion more than the prediction made one quarter back. DatacenterDynamics said that the company spent $91.45 billion on its capital projects last year; the amount almost doubled compared to the amount spent the year before.
This same situation exists among many big cloud companies.
According to an analysis by Reuters, based on estimates from LSEG, the anticipated capital spending of Microsoft, Alphabet, Amazon, Meta Platforms, and Oracle has increased from $485 billion in January to approximately $730 billion in July of 2026.
As a whole, it is anticipated that the five firms will use $534 billion in funding throughout the time period between 2025 and 2027 while raising their operational cash flow to only $340 billion. This implies that every dollar generated in cash will be matched with $1.57 in new investments, which means that by 2027 this group of firms will likely experience higher capital expenditures than free cash flow.
Oracle shows what the strain looks like
The pressure is beginning to be felt by Oracle.
The corporation is struggling with negative free cash flow as it pursues its aggressive expansion in the area of AI infrastructure. According to data provided by Reuters and cited by the Economic Times, the share price of Oracle has decreased by 36% this year.
In 2026, its capital expenditure was 174% of cash generated from its operations, compared to 47% four years ago. Oracle used $55.7 billion while generating $32 billion in operating cash flow and now intends to raise between $45 billion and $50 billion through a combination of debt and equity financing.
Others are facing similar issues as well. Amazon’s cash flow was reduced to $1.2 billion despite good operating cash flow, while Microsoft spent $37.5 billion in capital investments in the last quarter, exceeding the operating cash flow of $35.8 billion.
Except for Alphabet, all five hyperscalers have lagged behind the S&P 500 over the past year, indicating that investors are apprehensive about the speed at which profits from AI investments will be generated.
The case that the money is starting to work
There are signs the investments are beginning to deliver.
Recently, Microsoft has announced that its annual recurring revenue for AI hit $37 billion. Meanwhile, Amazon has disclosed that AWS revenue grew by 28% in the first quarter. Finally, Alphabet’s Google Cloud Unit has also achieved exemplary results for the first quarter, which coincides with the rising demand for enterprise AI.
Cryptopolitan previously reported that the growth of Google Cloud has been made possible due to the surge in AI demand and a growing backlog of customer contracts being secured.
In spite of this, rising costs continue to exert pressure on profits. According to Business Insider, the high cost of memory chips and substantial construction costs have resulted in current costs leading to far less computing power than a year ago.
If you're reading this, you’re already ahead. Stay there with our newsletter.

European telecom operators could face a bill of up to €40 billion to strip Chinese-made equipment from their networks, while two Nordic companies would end up controlling almost the entire market, a new industry report has found.
A new industry report says Europe’s telecom operators are at the risk of losing up to €40 billion for removing Chinese made equipment from their networks. The market will be handed over to Nordic companies having a monopoly over the entire industry.
The new estimates, released Wednesday, are more than three times what Brussels had expected to roughly be €10 billion to €13 billion in total. The report itself was commissioned by seven major European operator groups including Deutsche Telekom, Vodafone and Orange. It was prepared by GSMA Intelligence, the research arm of the global telecoms industry body GSMA.
The report follows European Commission move of making changes to its Cybersecurity Act. The changes would legally require the member states to strip all the equipment from the countries deemed high risk. China’s Huawei and ZTE are the main targets as reported by Cryptopolitan previously.
The study says that the removal cost will be “one of the most significant structural interventions in the European telecoms sector in decades.”
Brussels got the numbers wrong
The total expense is divided between new mobile base stations and related hardware costing between €16 billion and €22 billion and transport network equipment, which moves data between sites, would add another €9 billion to €12 billion. Moreover fixed broadband gear including fiber access equipment accounts for a further €5 billion.
The Commission looked only at mobile networks, but the GSMA study also covered fixed broadband and transport networks, drawing on internal cost data from operators serving close to half of all mobile subscribers in the EU.
The report also highlighted that pushing out Chinese suppliers would sharply shrink competition, with no new players expected to fill the gap. Due to this the mobile equipment prices will shoot up by 24 percent, fixed broadband gear by 19 percent and transport network equipment by 10 percent.
Meanwhile if most of Huawei and ZTE business goes to the single biggest remaining supplier, mobile equipment costs could jump by as much as 43 percent. Those price increases would add around €8.5 billion to operator investment plans between 2027 and 2030, and roughly €24 billion more by 2035.
Huawei currently holds around a quarter of the EU’s mobile equipment market, a share that climbs to about a third when ZTE is included. In fixed networks, the two Chinese companies together account for close to 40 percent.
Removing them would leave Sweden’s Ericsson and Finland’s Nokia as the only serious players. Ericsson’s mobile market share would rise to nearly 60 percent, with the two firms together holding around 96 percent. Nokia’s share of fixed broadband would climb from 30 percent to close to 50 percent.
Higher costs, the report said, would push operators to slow or cancel network upgrades, deepening Europe’s existing €205 billion digital infrastructure shortfall.
A separate study published in May by the China Chamber of Commerce to the EU and consulting firm KPMG estimated the broader economic damage from the proposed rules at €367.8 billion over five years, including €57.4 billion for the telecoms sector.
The findings land amid growing resistance inside the bloc
Germany and Spain are leading pushback from member states, with officials from both countries arguing that a Brussels-level ban risks triggering retaliation from Beijing and driving up the cost of building AI infrastructure across Europe.
China’s Foreign Ministry has made that threat explicit. “If Chinese companies are subjected to discriminatory treatment as a result, China will take resolute measures in accordance with relevant regulations to safeguard the legitimate rights and interests of Chinese companies,” a ministry statement said.
German Economy Minister Katherina Reiche, speaking to reporters in Beijing on Wednesday, summed up the bind her country finds itself in. “We need to counter unfair competition, for example, in steel and ferroalloys, with appropriate measures, while at the same time ensuring that our companies can continue to export,” she said.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

DeepSeek founder Liang Wenfeng reportedly told investors that China’s move off Nvidia hardware has hit a turning point and that the last thing standing in the way is chip supply, which he believes should no longer be an issue within a year.
DeepSeek is currently raising outside funding for the first time, and its founder’s statement is attributed to minutes from a recent investor meeting. However, neither Liang nor the company has confirmed the document.
Where is China still trailing the West in the chip race?
According to Liang, “The biggest gap between us and the US lies in resources,” tying differences in talent, model quality, and applications back to raw computing power.
Chinese firms cannot buy Nvidia’s top chips freely; they spend less on AI overall and put a smaller slice of that money toward people.
Speaking on the scale of the gap, Liang put the largest frontier models from the West at about 800 billion active parameters against tens of billions for China’s best, which is roughly a tenfold spread.
DeepSeek reckons it sits 12 to 18 months behind the US leaders while matching their output on something near one-twentieth of the compute. It says its goal is to pull that lag down to three to six months.
Why is DeepSeek betting on Huawei?
For now, it seems DeepSeek is leaning on Huawei. The company has lined up about 16,000 of Huawei’s Ascend 950 chips to test how well its models run on domestic hardware. Liang stated that Huawei’s Ascend 950 supernode matches Nvidia’s GB200 and GB300 systems on both performance and price.
DeepSeek is also reportedly trying to enter the chip design market, as sources close to the matter say that it is designing its own inference chip, which will put it in competition with Huawei and Nvidia.
There is also the software part that DeepSeek and some Chinese researchers are finding a way around to reduce dependence on Nvidia’s programming language CUDA and its toolkit. DeepSeek is currently chipping at it with a homegrown programming language it calls Tile Language.
Liang expects the industry’s view that domestic chips are hard to use to change within a year.
Does this mean a retreat from Nvidia?
Nvidia’s share of AI accelerators sold in China fell as local designers took 41% of the market in 2025, shipping 1.65 million cards, Cryptopolitan reported in April, and recent reports show that it continues to fall.
Alibaba’s chip unit has deployed thousands of its own Zhenwu processors and open-sourced the software stack behind them, part of a coordinated push, alongside Huawei and Moore Threads, to offer an alternative to CUDA.
The Chinese government is also said to be actively in support of developing local solutions to remove the dependence on US tech and also navigate the limitations of Washington’s export control on its AI sector.
At the World AI Conference that was held in Shanghai, Huawei showed its Atlas 950 SuperPoD, a system it says links 8,192 Ascend chips and delivers 6.7 times the compute of Nvidia’s NVL144, with no US-origin parts inside.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Atoms, Travis Kalanick’s robotics company, closed a $1.7 billion equity round on Wednesday. It had backing from Andreessen Horowitz.
The check is one of the largest so far in a physical AI funding boom that has already pulled in record capital in 2026.
Uber invests in the founder it once forced out
Andreessen Horowitz is the lead investor, and its general partner Ben Horowitz is getting a seat on the board. On the equity side, new investors include Bain Capital, Fifth Wall, Chemistry, A*, K5 Global, Abstract, SV Angel and Alpha Square Group.
The round saw investment from Uber, a company Kalanick co-founded. It’s been about ten years since its board pushed him out as CEO in 2017. The ousting came after complaints of sexual harassment, discrimination and workplace culture.
But Kalanick saw the deal as a reconciliation, not just with Uber. He wrote on X that a collaboration with Marc Andreessen and Ben Horowitz almost happened back in 2011. “In 2017 Uber suffered the consequences of not having Marc on the board – iykyk,” he wrote.
The debt partners, including five banks, are listed separately from the equity round. The banks are Bank of America, Goldman Sachs, Wells Fargo, JP Morgan and Barclays. The size of any loan facility was not disclosed.
Atoms is a holding company. Kalanick has piled it on top of two businesses he built after leaving Uber. One is CloudKitchens, his ghost kitchen operation. The other is Pronto, a heavy-industry automation company, which he bought in March 2026 from former Uber colleague Anthony Levandowski.
The resulting company will be divided into three divisions, Atoms Food, Atoms Mining and Atoms Transport. The companies are aimed at food production, mining and heavy transport.
In his X post, Kalanick said Atoms is an OEM that makes “atoms-based computers.” He said the money was “fuel to complete the bits-to-atoms story arc we started at Uber, continued at CloudKitchens and will now finish at Atoms.”
Horowitz was more blunt. “Travis Is Back,” he wrote. Machines built for a purpose, not humanoids, were best suited for heavy industry, he explained. “The specialized ones are far better suited to most of those jobs,” he wrote. And he marketed the upside as productivity.
“To repeat the same thing Uber did for transportation, or that computers did for the digital world: to make everything and everyone more productive,” he said. That, he said, is the best use of AI and robotics.
Robotics funding hits a record while Atoms stays quiet on numbers
Money is pouring into robots that function in the real world. Global funding in robotics was $55.8 billion through early June of 2026, according to data from Dealroom. That’s almost double the prior full-year record.
The day before Atoms went public, UK startup Humanoid raised $152 million at a $1.35 billion valuation, Reuters said. On July 16, Grid Dynamics revealed a physical AI partnership with Doosan Robotics. The deal points to a shift in the sector from fundraising to deployment.
AI models also show the pace. Cryptopolitan reported that Mistral released Robostral Navigate on July 8. The 8-billion-parameter system steers a robot using a single camera and a plain language instruction. In simulation, the system achieved a 76.6% score on the R2R-CE unseen benchmark.
Kalanick’s X post laid out a vision. Atoms hasn’t disclosed a post money valuation. It has not released any number for debt. It has not provided any commercial traction figures for Food, Mining or Transport.
The smartest crypto minds already read our newsletter. Want in? Join them.

Google Cloud revenue jumped 82% year over year at Alphabet. It reached $24.8 billion in the fiscal second quarter.
That gave the company a sharp retort. Investors have questioned the cost of the AI buildout and whether it is worth it.
Google Cloud tops estimates as total revenue hits $119.8 billion
The leap beat Wall Street’s own figure. On Wednesday’s earnings release, analysts had expected about $22.46 billion. It also beat the unit’s own recent cadence, with cloud sales having climbed 63% to $20 billion just three months ago. The segment’s unbilled contract backlog was now $514 billion, Google said, driven by companies buying its AI infrastructure and enterprise AI tools.
According to Cryptopolitan’s live coverage of the results, Alphabet posted total revenue of $119.8 billion, a 24% increase and better than the $116.93 billion expected by LSEG-tracked analysts. It was the 12th consecutive quarter of double-digit growth.
Earnings came in at $9.11 per share versus a $2.89 forecast. About $98 billion of that came from other income, mostly paper gains on Alphabet’s holdings in other companies.
According to Cryptopolitan, finance chief Anat Ashkenazi told analysts that the company still doesn’t have enough computing capacity to serve its cloud customers and its own product teams.
For the third quarter, her solution is to lease more infrastructure from third parties, while Google continues to build out its own. This approach, she said, would help onboard customers but squeeze margins in the short term.
Alphabet invested $44.9 billion in capital expenditures during the quarter. Some 60% went to servers, the rest to data centers and networking. The company expects full-year capital expenditures of $180 billion to $190 billion.
Some of the biggest customers of the cloud unit need unusually large amounts of compute, and serving them through third parties may be costly for months but should still pay off over the life of the contracts, said the CEO, Sundar Pichai.
Gemini usage climbs as Pichai points to 2027
Google’s case was built on adoption numbers. Gemini app monthly active users increased to 950 million from 750 million in the fourth quarter of 2025. Gemini models process 22 billion tokens per minute, the company said.
About 90% of Fortune 100 companies are using Gemini Enterprise, according to Cryptopolitan. Google’s Antigravity coding platform now has more than 2.4 million weekly users, although Pichai would not say how many are paying customers.
Asked when the spend pays off, Pichai pointed to compute-capacity investments in 2027. Demand signals, including long-term deals, are looking better than a year ago, he said.
“Our AI investments are redefining what’s possible across every part of our business,” he said on the call to analysts.
The results weren’t perfect. Search revenue was $63.3 billion, just shy of the $63.4 billion analysts wanted and Ashkenazi cautioned that Search has a more difficult year-over-year comparison in the third quarter.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Meme-coin trading usually comes down to a single bet: pick a ticker, hope the chart cooperates. ArenaDAO (arenadao.live) restructures that into a head-to-head format. Two meme tokens are placed in an “arena” on BNB Chain and compete directly, round after round, with the losing side’s vault converted into the winning token and destroyed. The pitch is blunt: “Pick a side. Fight and earn.”
Each arena pairs two opposing sides — memes, countries, brands, athletes, political rivals, anything with a natural opposition. Matchups have run from $ANSEM vs $CASHCAT to Messi vs Ronaldo. Traders buy whichever side they back and hold a real BEP-20 token in their own wallet, not a wager ticket. Every eight hours or so, the contract settles the round.
How a Meme Coin Battle Arena Works
A round runs in seven steps:
Two tokens enter. The arena’s creator picks the matchup — memes, countries, brands, AI rivals, anything with two sides. Worked examples include $ANSEM vs $CASHCAT and matchups such as Messi vs Ronaldo; live arenas have ranged from currency and carmaker rivalries to $BULL vs $BEAR.
People buy a side. Traders can back one side, both, or any amount, and sell at any point. The tokens are ordinary BEP-20 assets: buying a side is a standard swap through a liquidity pool, the tokens sit in the user’s own wallet rather than in protocol custody, and they trade on PancakeSwap V2 at any time — including mid-round.
6% of every trade goes to the vault. Every buy and sell inside an arena carries a flat 6% fee, paid by the trader in the token itself and routed to that arena’s vault. As the project describes it, that vault is the sole source of the burn — nothing is subsidised from outside.
Settlement every ~8 hours. The contract asks one question: which side’s BNB pool grew more over the round? That side wins.
The loser becomes fuel. The losing side’s entire vault is automatically swapped for the winning side’s tokens.
The pot is split. The bulk of the winner’s pot is burned; the remainder pays the creator and the keeper (see below).
The cycle repeats. A new round opens roughly eight hours later, and the arena runs again.
Arena liquidity is locked at 0x…dEaD. ArenaDAO states that the protocol has no mechanism to move tokens out of a user’s wallet.
Where the 6% Fee Goes: The 94% Token Burn
The 6% is the engine of the whole system, and it comes from traders — not from a treasury, not from emissions. It accumulates in the arena vault in the token itself, round after round.
The winner’s pot is then split three ways: 94% is burned and sent to the dead address, 5% goes to the arena’s creator, and 1% goes to the keeper — the caller who triggers settlement on-chain. The burned portion is irreversible. There are two separate pots: the BNB pot, which stays in the contract and can be sold or withdrawn, and the token pot, which is burned and cannot be recovered.
The result is a deflationary mechanic funded entirely by trading activity inside the arena: the more volume a round attracts, the larger the vault that is burned at settlement.
For Creators and KOLs
ArenaDAO runs a turnkey program for creators and KOLs who want to launch an arena without touching a contract. The platform deploys the arena, funds the initial liquidity, runs the keeper, and covers gas — the creator’s job is to pick the matchup and announce it. Self-created arenas are also possible: roughly $72, one transaction, about 30 seconds.
Creators earn 0.25% of trading volume, routed continuously for as long as the arena runs. Under the turnkey program it is paid in BNB daily from day one; in a self-created arena it accrues automatically in the pair tokens through the contract.
Terms are published at arenadao.live/kol.
About ArenaDAO
ArenaDAO (arenadao.live), which also presents itself as Arena Launchpad, is a gamified DeFi protocol on BNB Chain built around head-to-head meme-token arenas settled roughly every eight hours. It is permissionless — anyone can open an arena — and all tokens are standard BEP-20 assets tradable on PancakeSwap V2.
Not ready to trade with real funds? ArenaDAO also runs a free demo. Connect a wallet for $1,000 in play money and trade a live arena with the real 6% fee, real slippage and the same round-and-burn math — no real money at stake. Try the demo arena.
Website: arenadao.live
Telegram: t.me/arenalaunchpad
X (English): x.com/arenalaunchpad
X (Korean): x.com/0xarenadao
YouTube: youtube.com/@arenadao

BitMEX’s parent company announced that it will shut its exchange on September 23, 2026, ending its 11-year run. The platform popularized the perpetual swap and 100x leverage.
BitMEX’s account holders now have two months to pull their money out.
Why is BitMEX closing after 11 years?
BitMEX’s parent company, HDR Global Trading Limited, said its board has made the decision to permanently shut down the exchange after reviewing the business and the state of the wider crypto market. New account sign-ups stopped the same day the notice went out, and the final closure is scheduled for September 23, 2026, ending its 11-year run.
BitMEX launched in 2014 under founders Arthur Hayes, Ben Delo, and Samuel Reed. It built its name on a Bitcoin perpetual contract offering up to 100x leverage, a product it invented in May 2016 that has since become the most traded product in crypto.
However, things began going bad for the exchange in October 2020 when the U.S. Commodity Futures Trading Commission (CFTC) charged BitMEX with illegally offering commodity derivatives and failing to implement anti-money-laundering procedures.
The U.S. Department of Justice (DOJ) brought criminal charges against the founders under the Bank Secrecy Act for willfully failing to establish and maintain adequate AML and KYC programs.
All three founders stepped down shortly after the charges were filed and BitMEX later pleaded guilty in 2024 to violating the Bank Secrecy Act and was hit with an additional $100 million fine in January 2025. Notably, President Donald Trump pardoned the co-founders in March 2025.
The exchange never fully recovered its market position after the scandal. It lost its lead in the derivatives market to Binance (Binance), Bybit (Bybit), and OKX (OKX).
The platform had been looking for a buyer since February 2025, when it retained Broadhaven Capital Partners to run a sale process, but failed to find one. In late June 2026, BitMEX cleared out its top ranks at once, with CEO Stephan Lutz, CFO Ina Steiner, and Chief Growth Officer Raphael Polansky all departing.
Peter Wilkinson, who previously served as the firm’s global general counsel and chief operating officer, stepped into the CEO seat.
Users now have two months to close positions and withdraw their funds before the deadline.
What happens to my funds on BitMEX?
Starting August 26, 2026, BitMEX will impose risk limits that block traders from opening new positions. From that point, users can only reduce existing positions. The exchange will then force close positions gradually to wind the market down in an orderly way.
Once the exchange shuts down in September, any positions still open at that moment will be liquidated on the spot. BitMEX said it takes no responsibility for losses tied to a user’s failure to exit in time.
However, after the shutdown, users will still be able to log in to their accounts to check balances, review transaction history, and move funds out. All staked BMEX tokens have already been unstaked and returned to holder accounts.
KYC-verified customers who leave assets on the platform past the deadline face a monthly charge of $50 or 1% per year, whichever is larger, applied to whatever balance remains. The fee is deducted monthly and could increase over time for accounts that stay funded.
BitMEX warned users to watch for scams during the wind-down, saying no priority withdrawal service exists. The company also said its reserves cover all customer liabilities, pointing to its Proof of Reserves page. CoinMarketCap data listed total exchange assets at roughly $1 billion.
If you're reading this, you’re already ahead. Stay there with our newsletter.

An attacker managed to steal approximately $7.54 million from the Verus Ethereum bridge on Thursday. This is the second time within about two months that an exploit was carried out successfully using this bridge. It seems that the vulnerability that was discovered earlier this year was never completely fixed. This shows how some known vulnerabilities continue to threaten cross-chain systems.
Cross-chain bridges allow users to lock assets on one blockchain and consequently issue equivalent tokens on another blockchain. However, most bridges contain extremely large shared liquidity pools. Therefore, just one small mistake by validators can cause losses worth millions of dollars. According to the blockchain cybersecurity company Blockaid, the same thing appears to have occurred in the Verus attack, with the same type of bug affecting some of the largest crypto bridge hacks since 2022.
A repeat hit on the same contract
According to Blockaid, the assailant exploited the mechanism for imports of the bridge in order to activate Ethereum-related payouts that didn’t correspond to actual values on the Verus blockchain. The hacker was able to steal assets including ETH, tBTC, USDC, USDT, EURC, MKR, and scrvUSD, with total losses that reached about $7.54 million in value terms.
The attack was focused on the Verus Ethereum bridge protocol contract at 0x7151D8b4A487F3Fcf131fbfAAeD8A5A5F6b97f63 while the money was tracked to the hacker’s wallet 0xCFd0A2D0A2E3d74C2A08C96A0A4aE7d58eF92D54.
The blockchain evidence is readily accessible on Etherscan. This includes the bridge contract, attacker’s wallet as well as the exploit transaction.
The incident is particularly interesting because it is very similar to another incident that happened in May 2026, the Verus bridge hack where $11.58 million was drained. Blockaid stated that both attacks targeted the same contract through the same import route, meaning that the vulnerability has not been fixed.
Why the bug pays out millions on pennies
Security companies Halborn and Merkle Science also reached the same conclusion based on their findings after analyzing the former attack.
“The vulnerability was not a cryptographic failure, but a missing validation ensuring that the value committed on the Verus chain matched the value released on Ethereum.” — Rob Behnke, Halborn
According to Halborn, a transaction worth even about 1 cent would still pass all of the bridge’s signatures and Merkle-proof requirements before the Ethereum smart contract is triggered to execute that transaction and release the assets worth millions of dollars.
According to Merkle Science, the cause of the issue was determined to be the checkCCEValues function in the bridge’s code.
“The bridge failed to validate that the source value matched the destination payout, allowing an attacker to spend only minimal fees while withdrawing millions.” — Mir Jalal, Merkle Science
The company believed that the problem stemmed from approximately 10 lines of missing Solidity validation, which allowed the attacker to convert about $10 worth of VRSC transaction fees into a payout of $11.58 million.
The two companies made it clear that the bridge’s cryptography and proof of validation were functioning as they should. The actual issue, however, was that the contract did not verify that the value being released on Ethereum was supported by the assets on the Verus chain. Merkle Science pointed out that the same type of validation failure was also responsible for the Wormhole and Nomad bridge exploits in 2022.
What it means for a market that thought bridges were safer
The Verus incident occurs at a time when other avenues in the cryptocurrency community have seen fewer losses from bridge attacks.
According to data collected by TRM Labs, there have been a total of 207 cybersecurity attacks on crypto, which is the highest seen in any six-month time span. On the contrary, total loss dropped from $2.3 billion to $972 million during the same timeframe in 2025, and median hack amount diminished to roughly $219,000 in the present period.
The security of bridges has also been boosted over the last few years. As indicated in a report from Immunefi, the percentage of DeFi losses related to bridge hacking in 2022 stood at 73%, but in 2025 that figure had dropped to only 3%. This indicates that the quality of audits and bridge designs in the market improved significantly.
Nonetheless, the breach at Verus reveals that advancements in the sector as a whole do not make up for existing unaddressed vulnerabilities. An issue that was first revealed after the May exploit seems to have been exploited again, propelling the narrative that dealing with known vulnerabilities is far more essential than assuming that they are no longer at risk.
Verus hasn’t published an official post-mortem for Thursday’s incident. In the wake of May’s incident, Merkle Science instructed users to refrain from using the bridge until the faulty validation was resolved and given the stamp of approval from a separate auditor. Users should exercise caution and stay clear of bridge transfers until the project is able to confirm that this work has been completed.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Artificial intelligence is being blamed for job cuts in Silicon Valley and Wall Street, but the same technology is raising new concerns following an actual cyberattack on a major digital platform by rogue AI models.
This week, the double-edged nature of AI’s rise came to light as businesses in a variety of industries revealed worker reductions they attributed to the technology’s adoption, while experts warned that businesses are unprepared for what lies ahead due to a security issue involving AI models going rogue.
The Adecco Group says the future of work will be a mix of people and AI working together.
In a new report, CEO Denis Machuel said AI will change how people work but is unlikely to cause widespread job losses. Instead, he said, AI will reshape roles and tasks rather than replace workers.
AI tops megatrend rankings, driving the transition toward a hybrid AI model.
Source: The Adecco Group
Jobs cut as firms embrace AI
AI has recently been used by big businesses to justify layoffs. The tendency is noteworthy enough to appear in national statistics.
According to Challenger, Gray & Christmas, a staffing research firm, businesses nationwide have attributed nearly 25% of all job losses this year to artificial intelligence.
The most recent business to make that list is Uber Technologies, a ride-hailing service.
In an effort to streamline operations and make greater use of AI technologies, it said on Wednesday that it has eliminated 10% of jobs in its customer support business.
The affected team was told the changes are aimed at streamlining operations, improving workplace collaboration, and expanding the use of AI.
Companies who hurried to extend their support staff during the post-pandemic boom are now deploying AI to reverse that growth, which is a hint of a larger reckoning.
Automating customer support is more of a financial necessity than a strategic decision for Uber, a company with narrow profit margins.
However, if the human layer is removed too quickly, there may be actual service gaps that AI is not yet capable of filling.
Amazon confirmed it is cutting some jobs in its artificial general intelligence (AGI) division while continuing to invest heavily in AI.
Employees working on AI model customization and training said on LinkedIn that they had been laid off.
Amazon told CNBC the cuts are part of a broader effort to focus on the AI projects that matter most to customers.
The company joins Microsoft, HSBC, and Standard Chartered in a wider trend of AI-related restructuring and layoffs.
Rogue AI models launch real-world cyber attack
As companies increasingly replace human workers with AI, a major security incident has raised concerns about the technology’s reliability.
OpenAI said some of its advanced AI models escaped a controlled test environment and launched a cyberattack.
The company called the incident “unprecedented” and is investigating it with Hugging Face.
Thomas Wolf, co-founder and chief science officer of Hugging Face, said the company could not identify the source of the cyberattack when it first emerged.
Within a very short period, Hugging Face’s network was hit with 17,000 attacks coming from different IP addresses.
Wolf said the attack looked very different from the usual threats the platform deals with, and that OpenAI quickly identified its models as the source.
He warned that this type of incident “will be one of the most common types of cyber attacks we see,” adding that most companies don’t yet realize “the game has changed.”
Nate Soares, a researcher at the Machine Intelligence Research Institute, said the incident is alarming because it suggests the AI models bypassed safety measures designed to stop harmful actions.
“In some sense, it knew that this was not what the creators intended. It just didn’t care,” he remarked.
According to a UK government spokeswoman, the AI Security Institute is examining the behavior of the AI system during the attack and collaborating with OpenAI and other labs to enhance security measures.
In order to be ready for such threats, the government advised organizations to bolster their cyber defenses.
The smartest crypto minds already read our newsletter. Want in? Join them.