San Antonio orders warning signs on all 193 Bitcoin ATMs after $39 million in scam losses
San Antonio has passed an ordinance requiring fraud warning signs at every crypto kiosk in the city. The move targets 660 scam reports and about $39 million in losses police recorded from January 2024 to April 2026. The San Antonio Police Department said the scams follow a similar pattern. A caller claims to be a law enforcement officer, court official, government agency or utility company and invents an emergency. They then tell the victim to deposit cash into a Bitcoin ATM. The caller says the payment will clear up an arrest warrant, unpaid fine or overdue bill, according to KENS 5 reporting. Police say the callers keep victims on the phone during the entire transaction. This means they can’t ask a relative, ask a store employee for help or call 911 before the money is converted to crypto and moved out. Bilingual signs hit 193 crypto kiosks on July 1 During the ordinance approval process, SAPD data showed nearly 38% of the victims identified were 66 or older, according to the San Antonio Report. Victims were as young as teenagers and as old as 90 years. About 88% of the cases resulted in losses of less than $50,000. But four individual cases were over $1 million each. The city identified San Antonio’s 193 crypto kiosk locations. That number beats Dallas, Fort Worth or Austin, according to the San Antonio Report. Operators are required to post warning signs in English and Spanish at each crypto machine. The signs must be in color-coded 18-point font and must be placed so that users standing at the kiosk can read them directly. The signs will list common crypto scam tactics and tell anyone feeling pressured to send money to call 911 instead. The SAPD must generate and distribute the signs, and enforce compliance. Failure to post them subjects the company to a fine of $100 to $500 for each violation, each day that a violation continues being a separate offense. The ordinance goes into effect on July 1. Texas pushes for a statewide ban San Antonio isn’t the only Texas jurisdiction to go after crypto kiosk fraud. Smith County Sheriff Larry Smith met with state legislators earlier this week to lobby for a statewide ban on the machines. Sheriff Smith first publicly called for a ban in May after an elderly woman lost $13,000 to an inmate running a scam from a Georgia prison. Among those who attended the meeting were the office of State Sen. Bryan Hughes, State Rep. Cole Hefner, State Rep. Daniel Alders and officials with the Texas Financial Crimes Intelligence Center. The report noted that Indiana, Tennessee and Minnesota already have a ban on crypto ATMs at the state level. Cryptopolitan has previously reported that Bitcoin Depot, which has more than 9,000 crypto ATMs across North America, filed for Chapter 11 bankruptcy in May after its Q1 2026 revenue fell by ~50% year over year. In February, the Massachusetts Attorney General sued the company for more than half of its ATM revenue in the state being linked to transactions associated with scams. U.S. Secret Service Investigative Analyst Laura Bravo told KSAT that crypto moves faster than traditional money and gets harder to recover once it reaches a foreign exchange. Crypto ATMs also isolate victims by stripping out the human interaction a bank teller would provide, Bravo noted. This gives the scammers more control, and the victim follows phone instructions at a crypto machine. SAPD’s advice is blunt. No government agency or utility company will ever ask you to pay via a Bitcoin ATM. That’s the case even if the caller knows personal details, uses the name of a real officer or spoofs a real phone number. When cash is converted to crypto and sent to a scammer’s wallet, the transaction cannot be reversed. The smartest crypto minds already read our newsletter. Want in? Join them.
Retired Florence man loses $222K after falling for a pig butchering crypto scam
A retired man in Florence, Alabama, lost $222,000+ to a crypto romance scam. Federal court documents filed this week say he was manipulated by a person posing as a young woman to transfer his savings into fraudulent crypto wallets. The U.S. government has filed a civil forfeiture complaint to permanently seize the funds. WAFF reported June 26 that investigators tracked the money through several wallets and exchanges before seizing it under a federal warrant. How the scam worked Court papers described a classic “pig butchering” scheme. In this scam, fraudsters gradually build victims’ trust over weeks or months and then direct them to fake investment platforms. The federal complaint states the victim connected with a person calling herself “Bella” who claimed to be a 23-year-old woman and offered to help him invest in crypto. The conversations moved to the encrypted messaging app Telegram, where they became romantic. “Bella” then provided step-by-step instructions on how to transfer money out of the victim’s Alabama bank account to a Coinbase account. The money was then moved to cryptocurrency wallets controlled by the scammers. Federal prosecutors say more than $222,000 of the stablecoin USDT was laundered through a series of wallets and exchanges before being seized by law enforcement, according to WTVA’s reporting of the court documents. Now the government wants a federal judge to order the forfeiture of the seized money as proceeds of wire fraud. Pig butchering drained $5.5 billion in 2024 In February 2025, on-chain security firm Cyvers reported that crypto investors lost $5.5 billion to pig butchering scams in 2024. That grew to 200,000 identified cases on the Ethereum network alone, per Cryptopolitan’s earlier coverage. Data from Chainalysis during the same period found that pig butchering accounted for 33.2% of all crypto scam revenue by subclass. Deposits in these schemes were up by about 210% year-on-year. Michael Pearl, vice president of GMT strategy at Cyvers, called pig butchering “by far the biggest threat” to crypto investors. He ranked it above outright hacks, which stole $2.3 billion in assets across 165 incidents in 2024. Across the data, 75% of pig butchering victims lost more than half their net worth, Cyvers found. The grooming period ran one to two weeks in 35% of cases. But some scams stretched to three months. Men aged 30 to 49 were the most targeted group, according to the firm’s data. Cyvers found 75% of pig butchering victims across the data lost more than half their net worth. The grooming period was 1–2 weeks in 35 % of the cases. But some scams take up to three months. The firm’s data showed that men aged 30 to 49 were the most targeted group. “The rate at which bad actors are using elaborate pig-butchering scams to defraud innocent people is despicable,” Special Agent Stacey Moy of the FBI’s San Diego Field Office said in a statement tied to a separate 2024 forfeiture case in the District of Columbia. Federal prosecutors have not identified any suspects in the Florence case. It is not known whether any of the recovered funds will be returned to the victim. The forfeiture action is against the crypto itself, a normal step in cases of fraud when the perpetrators are located abroad. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Chinese startup's AI mosquito laser blows past its Indiegogo goal by 130 times
A Chinese startup has raised $2.7 million on Indiegogo for a portable device that detects and kills mosquitoes in mid-flight using AI and lasers. 4,000 backers from more than 50 countries donated. Photon Matrix Lab is located in Changzhou, Jiangsu, China. The company has built what it calls a consumer-grade laser defense system against mosquitoes. The device is about the size of a smartphone. The company’s Indiegogo campaign describes the device as a combination of a computer vision module with lidar to find and kill flying insects. Each backer shelled out about $630 to reserve a unit, according to data from Indiegogo cited by South China Morning Post. The campaign was first targeted to raise $20,000. It surpassed that by 130+ times. The device hunts mosquitoes with lidar and AI The system distinguishes between dust particle movement, mosquito movement, and sensor artifacts. Creator Jim Wong says the device can hit insects flying at up to one meter per second, with body sizes ranging from 2 to 20 millimeters. The company says the device can handle up to 30 mosquitoes a second. That figure has not been independently confirmed by anyone. Sand flies and fruit flies are also in the device’s targeting range. In April, Photon Matrix CTO Li Ran told China Daily that Chinese manufacturing gave the company an edge over Western rivals. Li said it takes two weeks to prototype a high-precision fiber laser module in Changzhou, because suppliers are nearby. “In Silicon Valley, it’s hard to find a supplier who can prototype a high-precision fibre laser module in two weeks,” he told the outlet. “But in Changzhou, the supply chain is right downstairs.” Chinese firms repurpose military or industrial tech for consumer products with the help of mature local supply chains, cheap lidar, and edge computing, according to SCMP. Mass production moves to August Photon Matrix initially promised deliveries in early summer 2026. But the company is now hoping to begin mass production in August. This sort of delay is common in hardware crowdfunding. Indiegogo and Kickstarter campaigns often do not ship on the original date, and some never ship. Photon Matrix still has to comply with Western safety standards for a consumer laser product. That may mean further delays. Laser mosquito defense technology has existed since 2007. Astrophysicist Lowell Wood proposed the idea in a Bill and Melinda Gates Foundation brainstorming session on malaria elimination, according to Wikipedia. Wood was an aide to the architect of the Strategic Defense Initiative, conceived during the Reagan years. He envisioned using missile defense principles against disease-carrying insects. Since then, a number of teams have attempted to create commercial versions. But none of them could be produced cheaply on a mass scale. Photon Matrix is a pricey $630 device, but it could be the first viable consumer product in the category. Climate change is expanding the range of mosquito species that carry dengue and Zika into previously temperate regions. In 2024, the European Union has seen over 300 cases of locally acquired dengue. According to a ClimaHealth report, that’s more than the combined total of dengue cases of all the previous 15 years. If you're reading this, you’re already ahead. Stay there with our newsletter.
Tether plans Gold-backed loans with Ledn using XAUT
USDT issuer Tether and crypto lender Ledn have laid out plans to let holders of Tether Gold (XAUT) borrow against the it later in the year, which would open a lending channel taking advantage of the stablecoin issuer’s $23 billion physical gold reserve. Tether partners with Ledn Lending platform Ledn announced it will add XAUT to its platform alongside Bitcoin (BTC) and Tether’s dollar-pegged stablecoin USDT. This is expected to go live before the end of 2026, and would let XAUT holders use their holdings as collateral for loans instead of selling off the gold they own. Each XAUT token represents one troy ounce of physical gold stored in Swiss vaults, according to Tether. The structure of this product mirrors Ledn’s handling of bitcoin-backed lending for the past few years. Client collateral is held on a 1:1 basis and is not lent out or used to generate yield, the company said. Gold-backed lending has been controlled by central banks, large financial institutions, and bullion dealers since forever. Tether and Ledn’s partnership is betting on bringing the same concept on-chain, giving holders access to liquidity without forcing the sale of their gold assets. “As digital assets become an increasingly important part of the global economy, demand is growing for solutions that combine long-term ownership with financial flexibility,” Tether CEO Paolo Ardoino said in a statement. Tether expands past the dollar Profits from USDT, the world’s largest stablecoin, have funded expansion into areas well beyond dollar-pegged tokens. The company has invested in precious metals marketplace Gold.com, partnered with crypto financing firm Antalpha on XAUT lending and physical redemption, and has also backed AI infrastructure provider Northern Data. Tether has also put massive amounts of capital into bitcoin mining and multiple renewable energy projects. The lending product backed by Tether’s physical gold reserves is expected to expand Ledn’s collateral options beyond its current supported assets. As at the time of writing, XAUT traded at $4,070.34 per token. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Arthur Hayes faces fresh exit liquidity accusations over $CARDS token sale
BitMEX co-founder Arthur Hayes is facing another round of exit liquidity allegations after on-chain observers flagged that his fund, Maelstrom, appeared to offload $1.92 million worth of $CARDS tokens within days of Hayes publicly promoting the project. The move, which is recognized as using others as “exit liquidity” to get out of a trade, is coming just roughly three weeks after blockchain investigator ZachXBT called out Hayes for similar actions that involved four different tokens. Why is Arthur Hayes getting criticized? On June 23, Hayes posted on X that “$CARDS degens” had a “solid” thesis and predicted the token’s price would be “pamping,” according to his post on X. Maelstrom’s official account shared a link to the project around the same time, according to a post from the fund’s X account. Four days later, crypto analytics account SolanaFloor reported on X that Hayes had set a $4 price target for $CARDS when the token was trading around $0.30 and that Maelstrom sent $1.92 million worth of $CARDS to market maker Flowdesk the following day. SolanaFloor added that it was “likely for selling.” That was all Crypto Twitter needed to fire a barrage of posts and criticism at the socially active Arthur Hayes. The token was trading near $0.23 at the time of SolanaFloor’s post, a decline of roughly 23% from where it sat when Hayes endorsed it. Currently, it trades around $0.24 Another on-chain analyst, Ericonomic, also flagged the sequence on X, noting that Hayes “shilled $CARDS 4 days ago” and that three days later an address sold “his entire stack through Fireblocks.” Ericonomic added that the wallet address was never publicly disclosed by Hayes, and the link was based on timing and token patterns. What did ZachXBT call out Hayes? On June 6, Cryptopolitan reported that ZachXBT confronted Hayes over a similar cycle of endorsing tokens and then liquidating his holdings of those tokens. That time it involved four tokens, HYPE, NEAR, ZEC, and WLD. ZachXBT documented how Hayes exited all four positions within a two-week window after publicly endorsing each one. Hayes had called HYPE, ZEC, and NEAR the “Holy Trinity” on May 22, then went on to sell his HYPE and NEAR holdings by June 4 and dumped ZEC on June 5 after citing an exploit in its Orchard Pool. He also closed his WLD position the next day, less than 24 hours after framing Worldcoin as a SpaceX IPO play. ZachXBT asked Hayes directly how much exit liquidity his followers had absorbed. Hayes responded that he “sold to a willing seller at a price” and that he “happened to call it right this time” regarding his trading goals. ZachXBT’s history of flagging suspicious actions ZachXBT has built a track record of flagging this kind of promote-then-sell dynamic across crypto. His investigations into RAVE, SIREN, and LAB tokens over the past two months have all centered on the role insiders or prominent figures play in generating retail buying interest and then selling into the demand they created. In a May 14 investigation into LAB, ZachXBT documented how insiders allegedly controlled over 95% of the token’s supply while the project reached a fully diluted valuation above $6 billion. He characterized that case as “everything wrong with the current meta of retail extraction on major centralized exchanges.” So far, Hayes has not publicly responded to the latest $CARDS allegations, and the connection between the Maelstrom fund wallet and the Flowdesk transfers has not been independently confirmed beyond what was cited by SolanaFloor and similar sources. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Coinbase CEO floats Chinese open-weight AI models as antidote to runaway bills
Coinbase’s CEO has proposed experimenting with cheaper open-weight AI models to keep AI spending in check as token consumption climbs. This proposal has led to concerns over the security and geopolitical risks of directing enterprise workloads through Chinese-origin systems. Why are companies using Chinese AI models? U.S. export controls have made it harder for Chinese companies to access American AI chips, but that hasn’t stopped them from building competitive models and selling them at much lower prices. For instance, Zhipu’s GLM 5.2 costs $1.40 per million input tokens and $4.40 per million output tokens compared to Anthropic’s Opus 4.8 at $5 and $25 for the same volume. GLM 5.2 scored 62.1 on SWE-bench Pro, a key coding benchmark, beating OpenAI’s GPT-5.5 at 58.6. One AI researcher said GLM 5.2 “is at least as good as Opus 4.8 and GPT 5.5.” Another called it “the first open model that can really compete with closed-source systems.” Is Coinbase using Chinese AI models Coinbase’s CEO Brian Armstrong says the best way to control rising AI costs is to use cheaper open-weight models, including systems from China like GLM 5.2. Armstrong said instead of spending more and more on AI, companies need “better defaults, routing, and caching.” His suggestion to use Chinese models, even if they are cheaper, has drawn concerns about security and political risks. Beyond its convenient pricing, GLM 5.2 uses an MIT license, meaning companies can download it, modify it, and run it on their own servers, removing any risk of sending sensitive company data to an outside API. AI spending has become a genuine issue, causing companies to roll back the use of the technology in operations. Cryptopolitan recently reported that Uber used up its entire 2026 AI coding budget by April and now caps engineers at $1,500 per tool each month. Meta sent a memo warning of an “exponential increase” in AI usage and started building spending controls. Amazon scrapped an internal leaderboard that ranked employees by AI consumption because people were gaming it and driving costs up. A KPMG survey found only 26% of companies have full visibility into their AI costs, while 22% discover spending only after receiving the bill. Goldman Sachs projects that AI token consumption could increase 24-fold by 2030, reaching 120 quadrillion tokens per month. The International Data Corporation predicts 70% of leading AI-driven enterprises will use multiple models by 2028 rather than relying on a single provider. What makes Chinese AI models risky? Z.ai’s cloud API, which allows developers and companies to use its AI models (including GLM 5.2), falls under China’s National Intelligence Law. That raises real concerns for any company handling sensitive information. U.S. lawmakers opened a formal inquiry in May into cybersecurity risks from Chinese-origin AI models in critical infrastructure. There are also concerns that models trained under different legal systems could carry undisclosed behaviors. Adding to that, an AI builder tested GLM 5.2 against GPT-5.5 on a debugging task and found it “not even close” to the OpenAI model’s ability to spot problems, despite reports that Chinese models outperform their more expensive counterparts. Anthropic disclosed in an open letter to the Senate Banking Committee that Alibaba Qwen operators ran 28.8 million Claude exchanges through about 25,000 fake accounts between April and June. They called it the largest known campaign to steal a model’s capabilities. Self-hosting the open weights eliminates the API data-routing risk, as companies that run the model on their own servers don’t send data to China. But the concern about the models themselves remains. If you're reading this, you’re already ahead. Stay there with our newsletter.
Apple cuts five older iPad models from iPadOS 27 support
Apple will not roll out iPadOS 27 to any iPad with an A12 or A12X chip. That leaves five models out in the cold for the company’s next tablet OS. Devices sold from 2018 to 2020 are affected by the change, as Apple is raising iPad prices across the board. Apple kills all five A12 and A12X iPads According to 9to5Mac, the cutoff removes the 2019 iPad Air (3rd generation), the 2019 iPad mini (5th generation), the 2018 11-inch and 12.9-inch iPad Pro models, as well as the 2020 iPad (8th generation). All five iPads are powered by either the A12 or A12X processor. Apple gave these devices six to eight years of software updates, longer than most tablet makers provide. But the timing is suspicious. All iPhones that ran iOS 26 will support iOS 27. The iPad side is taking a harder line. There has been a lot of criticism for how well iPadOS 26 performs. Michael Burkhardt of 9to5Mac also argued that Apple should either allow owners to downgrade to iPadOS 18 or find a way to bring iPadOS 27 to the dropped hardware. A downgrade path is technically doable. Apple still ships security patches for iOS 18. iPadOS 18.7.9 landed about a month ago, so the older operating system stays actively maintained. Apple’s current policy blocks downgrades by “unsigning” old firmware versions when a new release ships. Apple first restricted Stage Manager to M1 iPads on iPadOS 16 in 2022, then relented and released it for the 2018 and 2020 iPad Pro models following criticism. Those older iPad Pros ran Stage Manager without support for an external display. Full windowing in iPadOS 27 may require a bit more GPU headroom and RAM than 4 GB can provide on the A12. Apple never made a public explanation. The iPadOS 27 beta is about three weeks old. If the public backlash increases, Apple may broaden device compatibility ahead of the fall release. Apple’s iPad prices jump by $100 to $200 On June 25, Apple hiked the iPad prices. The base iPad is now $449, up from $349. iPad Air was increased to $749 from $599. The iPad Pro increased from $999 to $1,199. Apple attributed the increases to the cost of memory components driven by the build out of AI infrastructure. “The consumer electronics industry is facing an unprecedented challenge,” Apple told CNBC. “The rapid expansion of AI data centers has created an extraordinary surge in demand for memory and storage. We have never seen a component price increase this much, this quickly.” Memory prices jumped more than fourfold since late 2025, and higher bills of materials are a lasting challenge said Tarun Pathak, research director at Counterpoint Research, to TechCrunch. Counterpoint said smartphone DRAM prices increased 50% and NAND flash storage prices rose by more than 90% Q/Q in Q1 2026. For a user owning the 2019 iPad Air, which is out of software support, it’s now $100+ more expensive to trade in for the lowest-end iPad than last quarter. If you're reading this, you’re already ahead. Stay there with our newsletter.
Singapore's MAS puts Hyperliquid on investor alert list
The Monetary Authority of Singapore has placed fast developing decentralized trading platform Hyperliquid (HYPE) on its Investor Alert List on Friday, flagging both the Hyper Foundation website and the Hyperliquid trading application as unlicensed entities operating without regulatory authorization in the city-state. The listing makes Hyperliquid one of the first major decentralized finance protocols to appear on the register. HYPE joins centralized exchanges Binance (BNB), KuCoin (KCS), Bitget (BGB), and Bybit on the alert list. What even is Singapore’s alert list? MAS created the Investor Alert List in 2004 as a consumer-protection tool, identifying financial service providers without proper licenses to operate in Singapore. However, it is important to note that appearing on the list does not mean there is any fraud at play or any enforcement action underway. Entities on the register have not gone through Singapore’s regulatory process. This process covers capital requirements, compliance with anti-money laundering actions, and consumer safeguards. The consequence for Singaporeans who use platforms on the list is that MAS protections do not apply if anything goes wrong in their use of these platforms. Hyperliquid responds In a post on X, Hyperliquid said the listing “does not constitute a ban, an enforcement action, or a finding of wrongdoing.” The protocol also stated that it has never represented itself as holding MAS authorization and described itself as permissionless infrastructure where users maintain self-custody and transactions are settled purely on the blockchain. Hyperliquid has been added to the MAS’s Investor Alert List (IAL). IAL listing does not constitute a ban, an enforcement action, or a finding of wrongdoing. The IAL provides a list of entities that, based on information available to MAS, may be wrongly perceived as being licensed… — Hyperliquid (@HyperliquidX) June 26, 2026 Hyperliquid added that its operations remain unchanged and that it would “continue to engage constructively with regulators and institutions.” The platform currently ranks as the ninth-largest decentralized exchange by trading volume on CoinGecko, with roughly $5.7 billion in total value locked per DefiLlama estimates. HYPE was trading around $64 to $65 at the time of the announcement, down by about 1% over 24 hours, according to Yahoo Finance. Singapore tightens crypto regulation Over the last few years, MAS has increased its efforts to shrink operational freedom available to unlicensed crypto operators in Singapore. In 2024, rules were put into place to bar crypto and digital payments service providers from offering credit, leverage, or trading incentives to retail customers while also prohibiting them from lending or staking retail assets, according to Yahoo Finance. Crypto firms have also not been allowed to market themselves to the general public since January 2022. In May 2025, MAS told crypto companies serving overseas customers from Singapore to either obtain licenses or stop operating, putting an end to a loophole used by some firms to avoid licensing by restricting services to non-Singaporean users. The smartest crypto minds already read our newsletter. Want in? Join them.
Anthropic nears deal to restore public access to Claude Fable 5
Anthropic’s Fable 5 AI model could be available to the public by next week, according to a report by Axios. Talks between Anthropic and the US administration are expected to run through the weekend, and a second source told Axios that Anthropic anticipates restoring access soon. Anthropic’s Fable 5 shutdown unfolded fast The Commerce Department told Anthropic on June 12 to block all users from outside the US. Anthropic couldn’t set up regional blocks quickly enough, so it took everyone off of both Mythos 5 and Fable 5. The two AI models run on the same underlying architecture. Some safety features have been taken out of Mythos 5 so that it can be used in certain situations. Fable 5 is the consumer product and is priced at $10 per million input tokens and $50 per million output tokens, according to Anthropic’s June 9 launch announcement. The payments company Stripe used Fable 5 to work on a 50-million-line script in just one day. This is faster than what its coders could have done by hand in more than two months of tests, according to Anthropic. When Fable’s access was revoked on June 12, coders found that automated work had stopped in the middle of a job, and companies rushed to replace it with cheaper Chinese models or competitors’ models. The American government cited cybersecurity concerns. During a congressional hearing, Senator Mark Warner said NSA Director General Joshua Rudd told him that Mythos could identify weaknesses across almost all US classified systems within hours, according to Cryptopolitan’s earlier reporting. A US official made it clear that just because Mythos found weaknesses, that didn’t mean it could exploit them right away. Cryptopolitan reported that Anthropic said the government’s findings were too narrow and that OpenAI’s GPT-5.5 could find software flaws just as well without hacking methods. On June 26, the Commerce Department gave Anthropic a partial win. It allowed Mythos 5 to be redeployed to a vetted list of American organizations, with more than 100 companies and institutions cleared, including several Fortune 500 firms, according to a recent report by Cryptopolitan. The White House asked OpenAI to gate GPT-5.6 access on a customer-by-customer basis with government sign-off, and only ~20 companies got limited preview access, Axios reported. Anthropic and OpenAI, two of the three largest American frontier AI labs, now run under some form of government access controls. Anthropic filed a confidential S-1 with the SEC on June 1, raising $65 billion at an implied valuation of $965 billion, Cryptopolitan reported. And the Fable 5 suspension piled on top of a separate controversy. On June 11, Anthropic apologized and reversed a policy that had secretly degraded Fable 5’s output quality on frontier AI research queries without notifying users. Researchers, including former Anthropic staff, called the approach anti-competitive. The company said future safeguards in that category would be visible to users. If you're reading this, you’re already ahead. Stay there with our newsletter.
GameStop will pursue eBay takeover despite initial rejection
GameStop will continue to pursue its unsolicited eBay takeover, weeks after the e-commerce company’s board dismissed the initial $56 billion cash and stock proposal from the retail gaming company as not attractive and borderline insulting. The video game retailer confirmed that it remained dedicated to completing the deal and said “additional materials regarding the proposed transaction are forthcoming,” according to Reuters. The company had promised to release a detailed presentation covering the strategic rationale and operational plan for combining the two companies earlier in the week, but the document had not been published by Friday’s close. GameStop’s first eBay bid GameStop CEO Ryan Cohen’s first pitch in May involved an offering of $125 per share and an interest in running the new entity post-merger. He described the merger as a path to building a solid competitor to Amazon. The eBay board, led by Chairman Paul Pressler, rejected the proposal on May 12 calling it “unappealing” and questioning GameStop’s financial capacity to support the bid, as reported by Crypropolitan. At the time of the bid, GameStop’s market cap was approximately $10.3 billion. GameStop posts stronger earnings In addition to confirming an unending interest in the eBay pursuit, GameStop has predicted an adjusted EBITDA hitting more than $600 million for fiscal 2026, up from $345.4 million in 2025. This led to a 2% increase in share prices in after-hours trading on Friday. The earnings projection may be aimed at bolstering the game retailer’s credibility as a buyer. A standalone GameStop generating nearly double last year’s EBITDA strengthens the argument that the company can service acquisition debt. GameStop has still not filed a formal tender offer after the rejection, and the promised presentation, when or if it does arrive, will need to address the financing shortfall and integration plans in concrete terms to further move the needle on investor sentiment. Debt concerns amid GameStop’s financial solidity Moody’s Ratings also dropped a breakdown of the proposed acquisition shortly after the initial offer was announced, warning that the merger could alter eBay’s finances. The credit ratings agency explained that eBay ended 2025 with approximately $7.2 billion in adjusted debt, along with a gross leverage ratio of about 2.3x. Moody’s explained that a huge-money merger financed through additional borrowing could possibly increase the company’s leverage levels and affect its credit metrics and financial flexibility negatively. CEO Cohen has previously argued that the deal would generate about $2 billion in annual synergies within a year of closing, with 60% from sales and marketing cuts, 25% from administrative savings, and 15% from product development. The credit agency agreed that these figures would translate to about 3.25x in deleveraging potential but expressed caution since there was no assumption of revenue losses or additional offsetting costs. If you're reading this, you’re already ahead. Stay there with our newsletter.
ARK Invest's Wood backs Bitcoin over AI as 'insurance policy' for global wealth
ARK Invest CEO Cathie Wood has stated that Bitcoin stands apart from the artificial intelligence trade because it offers protection against sovereign currency risk, something no tech stock can offer. “Capital outflows from less stable countries around the world will light another fire under bitcoin and other digital assets,” Wood wrote on X on June 27. She acknowledged that AI “has launched a technology revolution, deservedly sucking a lot of oxygen out of the investment world,” but added that it “cannot serve as the insurance policy” that Bitcoin provides. How much hold does AI have in capital markets? Wood is not the only voice drawing a line between the two asset classes. BlackRock’s head of digital assets, Robbie Mitchnick, said in a June 22 interview that Bitcoin’s weak performance since October 2025 is not in isolation, as everything outside AI trade has also suffered. According to Mitchnick, it is not a crypto-specific problem. “It’s been a tough stretch for Bitcoin since last October for all of crypto, and that’s consistent in many ways with just about everything that is not AI-centric,” Mitchnick said in an interview. He described the AI boom as “certainly sucking a lot of the oxygen out of the room.” Bitwise chief investment officer Matt Hougan recently called crypto a “contrarian bet” now that institutional capital has shifted toward AI stocks, robotics firms, and SpaceX. “Who needs crypto when the Nasdaq-100 is up 43% year-over-year?” he wrote. Bitcoin trades around $60,000 as of June 27, which is a decline of over 50% from the all-time high of above $125,000 reached last October. U.S.-listed spot Bitcoin ETFs recorded more than 45 consecutive days of outflows totaling $7.8 billion, according to Cryptopolitan’s earlier reporting. While these bleak numbers are coming out, the largest cryptocurrency by market capitalization, AI-linked semiconductor stocks from Nvidia, AMD, Broadcom, and Marvell have all outperformed Bitcoin on a year-to-date basis. Are crypto miners ditching crypto for AI? The capital flight extends beyond trading desks, as miners have also been pivoting to AI infrastructure services in droves. Bitcoin’s network hashrate fell from a peak of 1.151 zetahashes per second in October 2025 to roughly 0.888 zetahashes per second. A report from late March found that listed miners could derive as much as 70% of their revenue from AI infrastructure by the end of 2026, up from about 30% at the start of the year. In a mid-year update, Fidelity Digital Assets described the dynamic as “structural retooling,” noting that miners appear to be redirecting power and infrastructure toward higher-margin AI data center workloads. The 30-day average hashrate and mining difficulty both dropped roughly 8% to 9% from earlier highs before partially recovering. The Bitcoin difficulty adjustment that occurred on June 14 was the 11th largest downward adjustment in the network’s history, according to Galaxy Research. Is a Bitcoin bull run ahead or a pipe dream? Despite the pressure, several institutional voices say that the pendulum of investments will swing back to crypto. Mitchnick identified rising U.S. government debt and deficits as “ultimately the most important fundamental driver ahead” for Bitcoin, predicting the issue could resurface around the 2026 midterm elections. BlackRock itself recommended a 1% to 2% Bitcoin portfolio allocation on June 23, calling the asset a “complementary diversifier.” For Hougan at Bitwise, the bear market may be closer to its end than its beginning. Hougan wrote, “When crypto stops being a momentum trade, fundamentals start to matter, and this rotation is proof it’s already underway.” In her post, Wood stated that AI creates wealth, but it cannot serve as an insurance policy to protect it. If capital flight from weaker currencies goes up, or if U.S. fiscal policy triggers inflation fears, Bitcoin’s value proposition as a non-sovereign store of value becomes harder for allocators to ignore, regardless of how many AI chips Nvidia sells. How this plays out in the future remains to be seen. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Zuckerberg asks Meta to seek partnerships with Kalshi and Polymarket
Mark Zuckerberg has told Meta executives to pursue potential partnerships with prediction market platforms Polymarket and Kalshi with the company planning the development of its own competing app called Arena. Meta has seen an aggressive push into the prediction markets sector that has exploded since the 2024 U.S. presidential election. Combined trading volume on Polymarket and Kalshi hit almost $50 billion in 2025 and has already seen more than $130 billion so far in 2026, according to Binance Square figures. Arena’s operations, a different approach Meta’s prediction market app, currently in internal testing, is expected to take a totally different approach compared to its potential partners cum competitors. Individuals with knowledge of the matter told the New York Times that Arena will use a video-game-style points system rather than accepting real-money wagers, as previously reported by Cryptopolitan. Polymarket operates its markets via USDC settlements on the Polygon blockchain, while Kalshi operates as a CFTC-regulated exchange using cash. Arena’s points-based model will offer a significantly different style of operation to those employed by these top platforms. Meta has, however, not ruled out adding financial mechanisms that use real cash down the line after launch. Meta targets 100 million users Zuckerberg’s team is aiming for at least 100 million monthly active “predictors” on Arena, with the app targeting users between 18 and 34 years old. Meta plans to eventually integrate parts of Arena into Facebook and Messenger. Zuckerberg’s interest in exploring a partnership with Polymarket and Kalshi could suggest Meta intends to move faster in the development of the Arena platform instead of totally building from scratch. The tech giant is also said to be looking at multiple ways to integrate existing prediction markets into its ecosystem. Meta’s track record, market reactions Meta has had other attempts at financial infrastructure that have not gone so well or hit the ground running. The company’s Libra stablecoin project, later renamed Diem, collapsed due to regulatory opposition and was eventually sold off. Meta is also reported to be separately pursuing stablecoin payments through third-party integrations, with a USDC creator-payout pilot already running in Colombia and the Philippines. The news has slightly rattled the shares of companies in similar markets. DraftKings saw a drop of more than 2%, with Flutter Entertainment (FanDuel’s parent company) dipping nearly 2% before seeing a partial recovery. Robinhood, a company already offering Kalshi markets, also saw a decline. Bernstein analysts have estimated prediction markets could reach $1 trillion in annual volume by 2030, per Binance Square. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Ripple CEO Garlinghouse calls Saylor's bitcoin funding model a 'damning indictment'
Ripple’s chief executive, Brad Garlinghouse shared doubts about whether Strategy’s CEO Michael Saylor has been a positive influence on the broader crypto market during a Friday interview with CNBC. Garlinghouse specifically described the untethering of Strategy’s Stretch (STRC) preferred stock as a “damning indictment. However, the Ripple chief executive maintained that he remains bullish on Bitcoin, regardless of Saylor or anybody. What did Ripple’s CEO say about Strategy’s preferred stock slide? Ripple’s CEO Brad Garlinghouse singled out Strategy’s STRC perpetual preferred stock, which has been on a sustained slide below its $100 target, as evidence that Michael Saylor and Strategy had started to lose the plot during a CNBC interview on Friday. The way it was designed, STRC is supposed to hold near its $100 par value while paying an 11.5% annual dividend. However, it’s well below those levels, sitting at $74.57 when markets closed for the weekend. That 26% discount, according to Garlinghouse, is “a pretty damning indictment.” “Financial engineering does not drive long-term value,” Garlinghouse said during the interview. “Team Michael Saylor wasn’t focused on the right stuff and that has hurt the overall market.” Garlinghouse’s comments did not do Strategy any favors, piling on a company that has seen its common shares fall to their lowest level since February 2024, closing near $82 on Friday, while its primary reserve asset also dipped to $59,000. Garlinghouse also flagged Strategy’s leverage damage Garlinghouse also talked about how the leverage game that Strategy played to amplify gains on BTC’s rallies is now leaving deep wounds on the way down. “You start to see that in a place that can actually compound negatively,” the executive told CNBC. Annualized dividend obligations across Strategy’s preferred share classes have grown to roughly $1.2 billion. Strategy’s own disclosures show it is not ignorant of the financial pressure. The late May sale of 32 Bitcoins for $2.5 million to fund a dividend payment became bigger news than usual for a firm that still had 843,706 BTC that it paid nearly $64 billion for in reserve. That sale was also the first time the company liquidated any of its holdings since at least December 2022, per Cryptopolitan at the time. The financial pressure is not lost on other market stakeholders either. CryptoQuant’s head of research Julio Moreno estimated that the firm’s runway on dividends has gone from more than seven years to approximately just 14 months, per Cryptopolitan reporting. CryptoQuant recommended in a separate report that Strategy hit the brakes on buying Bitcoin for a bit to beef up its cash reserves. What did Garlinghouse recommend for Bitcoin? Garlinghouse made sure to separate his perspective on Bitcoin from Saylor’s Strategy. The Ripple CEO maintained his long-term bullish view of “digital gold.” His criticism was strictly directed at Strategy’s approach and funding structure. Garlinghouse was basically reading out of the Ripple handbook when he said utility will tether a digital asset’s long-term value. He pointed to Ripple’s XRP-powered cross-border payment infrastructure, which processed nearly $16 trillion in payment and prime brokerage volume last year, as an example. How has Saylor responded to skeptics? Saylor has not shown any signs of letting up on the Bitcoin-first playbook as of yet. Cryptopolitan reported on the CEO’s defiant tone during the week after CryptoQuant and Peter Schiff hinted at impending problems for the firm. Earlier in the week, he also promoted the STRC instrument when he posted that “Digital Credit is income for investors who believe in Bitcoin.” Strategy is down about $13 billion on the cash it paid for its over 847,000 BTC stash, according to BitcoinTreasuries data cited by Cryptopolitan. Ironically, Saylor is now down about 25% on the dollar cheque he wrote for every BTC that Strategy has bought, the same rate the STRC preferred stock currently trades below par value. The smartest crypto minds already read our newsletter. Want in? Join them.
On-chain investigators link KelpDAO and Humanity Protocol exploits to same attackers
The $292 million KelpDAO bridge exploit in April and the Humanity Protocol private key theft in June were already suspected as connected, as both incidents carried hallmarks of DPRK-linked operations, with fingers pointing to the notorious Lazarus group. Now, on-chain evidence shows the proceeds of those attacks are now flowing into shared wallets, which is a pattern consistent with a single laundering pipeline, according to blockchain analyst Specter. How did the attackers move the Kelp DAO and Humanity protocol funds? According to Specter, the Humanity Protocol attacker moved 15,403 ETH, which is around $23.6 million, to a relatively new Ethereum address. The funds were then crossed onto the Bitcoin network, where they mixed with proceeds that have been traced to the KelpDAO exploit. The funds stolen in the Humanity Protocol and KelpDAO attacks have landed in the same wallets, per ZachXBT and Specter. Source: TRM Labs This action is a well-documented Lazarus Group technique, where they consolidate proceeds from separate operations into unified Bitcoin wallets before routing them through mixers and over-the-counter desks. What connects the two exploits? According to Chainalysis’s investigation, the attackers behind the KelpDAO exploit on April 18 compromised internal RPC nodes operated by LayerZero Labs and launched a DDoS attack against external nodes simultaneously. The attackers tricked the Ethereum bridge contract into releasing 116,500 rsETH without a corresponding token burn on the source chain. The attack was attributed to the Lazarus Group. The Arbitrum Security Council froze over 30,000 ETH of the attacker’s downstream funds, and KelpDAO’s emergency pause also prevented another $95 million from being drained. Although the Humanity Protocol breach did not follow the same pattern as the Kelp DAO attack, post-mortem reports now show that North Korea-linked bad actors were involved. A Quantstamp incident report, prepared for Humanity Protocol on June 11, found that the attacker phished a company director, Chong Yee Wai, with a malicious email impersonating the Korean exchange Bithumb. Quantstamp stated that the attack was “characteristic of DPRK intrusions.” The malware gave the attacker remote desktop access to Chong’s Windows machine. From there, the attacker copied MetaMask wallet keys and used them to mint and sell unauthorized $H tokens on both Ethereum and BNB Smart Chain. This caused the token to crash by roughly 89%. Proceeds at known attacker addresses are worth over $21 million in ETH, according to Quantstamp’s findings. Legal complications add a twist to recovery efforts Currently, plaintiffs hold over $877 million in unpaid U.S. court judgments against North Korea. In May, they served the Arbitrum DAO with a restraining notice on April 30, seeking to seize approximately 30,766 ETH (about $71 million) of frozen funds. The plaintiff claimed that since the funds were linked to North Korea, they had the right to seize any funds from groups linked to the country as part of the money owed in unpaid judgments. Arbitrum already had a governance proposal in motion to transfer the frozen funds to a recovery initiative backed by Aave Labs, KelpDAO, LayerZero, EtherFi, and Compound, which would compensate affected users. A court later approved the Arbitrum vote to move the Kelp funds back to Aave. How the plaintiff reacts to this newfound confirmation of North Korea’s involvement is yet to be seen, but going by past incidents, chances are high that the Humanity Protocol loss and possible recovery could also come under litigation. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Hong Kong's crypto-asset reporting bill enters legislative review as stablecoin launches near
Crypto exchanges and service providers might soon be required to collect and share tax-residency data on their users in Hong Kong. Hong Kong lawmakers are considering passing a bill that sets up a way for tax authorities to see crypto activity for the first time ahead of the launch of the country’s first regulated stablecoins. What does the new crypto reporting bill require? The Crypto-Asset Reporting Framework (CARF) bill is currently being reviewed at the Legislative Council, following the structure of a related tax-information-sharing law that passed on June 17. Lawmaker Priscilla Leung wrote in Ming Pao on June 26 that the CARF legislation has a similar compliance setup to the recently amended Tax Ordinance. Under this bill, licensed crypto platforms must identify which users must be reported to tax authorities, gather and check documents that show where users pay taxes and they must be registered with the government. All reporting platforms must create an account with the tax department by January 31 each year. These platforms will also be required to keep detailed records, even if the business closes down. The government plans to add about 8,000 more financial institutions to the reporting system. However, most of them will likely file empty returns. The rules are set to start on January 1, 2027, with the first information exchange happening in 2028. When will Hong Kong’s first regulated stablecoins launch? Hong Kong’s first regulated stablecoins are expected to reach the market between mid-2026 and the end of the year. The Hong Kong Monetary Authority (HKMA) granted stablecoin licenses to the Hongkong and Shanghai Banking Corporation Limited (HSBC) and Anchorpoint Financial Limited, a joint venture backed by Standard Chartered, Hong Kong Telecom, and Animoca Brands. These two licensees were chosen from 36 applicants. Both plan to issue stablecoins tied to the Hong Kong dollar and HSBC has said it plans to connect its stablecoin to its PayMe mobile app. The HKMA’s Chief Executive, Eddie Yue, said the licensed issuers will focus on areas like cross-border payments, local payments, and tokenized asset trading because they have strong banking backgrounds. In a separate development, the Financial Services and the Treasury Bureau and the Securities and Futures Commission recently completed a one-month consultation period for licensing virtual asset advisory and management services. The emerging proposal will create new separate licenses for firms advising on virtual assets (VA advisory) and those managing VA portfolios (VA management). Firms that provide trading advice or market analysis without holding client assets must maintain at least HK$100,000 in liquid capital under the proposed rules while those that do hold client assets face a HK$5 million paid-up share capital threshold and HK$3 million in liquid capital. If you're reading this, you’re already ahead. Stay there with our newsletter.
GraniteShares brings first U.S. leveraged SK Hynix ETFs closer
GraniteShares plans to create new leveraged exchange-traded funds (ETF) based on SK Hynix stock for trading on U.S. exchanges. These funds will allow U.S. investors access to a rapidly growing Asian memory chip manufacturer whose share price has increased more than 300% in 2023 because of increasing demand for artificial intelligence products. The proposed funds—a 2x long daily ETF (SKUU) and a 2x short daily ETF (SKDD)—would deliver twice the daily performance of SK Hynix shares in either direction, according to ETF Tracker on X. A prospectus filed with the U.S. Securities and Exchange Commission in March described the proposed ETFs, while subsequent EDGAR filings show GraniteShares filed three delaying amendments, the latest setting an effective date of July 2, 2026. AI chip rally turns SK Hynix into ETF favorite SK Hynix, a dominant supplier of high-bandwidth memory (HBM) chips used in AI data centers, closed trading on June 22, 2026, with a market capitalization of roughly $1.35 trillion, briefly surpassing Samsung Electronics’ common-share market value to become South Korea’s most valuable listed company by that measure. 🚨 COMING SOON GraniteShares is launching the first U.S.-listed leveraged SK Hynix ETFs: -GraniteShares 2X Long SK Hynix Daily ETF $SKUU -GraniteShares 2X Short SK Hynix Daily ETF $SKDD$SKUU aims for 2X the daily move of SK Hynix, the South Korean memory-chip maker, while… pic.twitter.com/xePindWFLV — ETF Tracker (@TheETFTracker) June 26, 2026 The company’s explosive stock performance has already spawned a wave of leveraged products in Asia and Europe. In Hong Kong, CSOP Asset Management’s 2x leveraged SK Hynix ETF has ballooned to more than $16.8 billion in assets under management, overtaking the Tracker Fund of Hong Kong to become the city’s largest ETF. Analyst Rebecca Sin said the CSOP product had returned close to 900% year-to-date. On the other hand, Leverage Shares started offering a 3x long SK Hynix exchange-traded product (ETP) on the London Stock Exchange on June 12. GraniteShares’ U.S. filing adds another venue, and another leverage tier, to a rapidly expanding global ecosystem of single-stock derivatives built around one chipmaker. GraniteShares bets on SK Hynix demand GraniteShares is not standing alone as six more U.S. asset managers (Tuttle Capital (T-REX), Themes ETFs, ProShares, Direxion, Defiance ETFs, and Amplify ETFs) filed for SEC registration of leveraged ETFs based on the planned Nasdaq-listed American depositary receipts (ADRs) of SK Hynix. This effort is part of one of the fastest group-wide product launches for a single foreign semiconductor stock in history, and all seven providers are requesting ten different leveraged and inverse products that are likely to be issued before the ADR is released. Based on SEC filings, the manner in which these providers are competing is evident. In most cases, they are competing on the basis of brand, timing of the launch, and distribution instead of leverage ratios. Whereas GraniteShares is attempting to capture the first mover advantage through issuance of its 2x long (SKUU) and 2x short (SKDD) funds, other providers wish to wait for the ADR to be listed before introducing their products. This rapid cycle of competitive product introduction is consistent with previous high-profile companies like Nvidia and Tesla that also saw larger amounts of capital and higher volumes due to the first and early issuers compared to later entrants. The rush to file reflects expectations that SK Hynix could become the next major single-stock ETF trading vehicle for U.S. semiconductor investors. Kim Sun-woo, a researcher at Meritz Securities, told Global News Top that the company’s valuation discount to Micron Technology could narrow once U.S. investors gain direct ADR access. SK Hynix currently trades at roughly seven to eight times forward earnings, versus around eleven times for Micron. If that valuation gap compresses as international ownership broadens, leveraged ETF issuers stand to benefit from higher trading activity regardless of whether investors are positioned long or short. South Korean regulators voice concern The proliferation of leveraged products tied to Korean chip stocks has drawn scrutiny at home. Lee Chan-jin, governor of South Korea’s Financial Supervisory Service, said at a June 22 press conference that approvals for leveraged ETFs linked to Samsung Electronics and SK Hynix had been “prepared hastily,” according to Reuters. “Maybe I should have lain down on the floor to block it,” Lee said. “I personally regret (I didn’t).” The regulator said the products had contributed to heightened volatility and helped push margin-financed retail equity positions to a record 60 trillion won ($39 billion) by the end of May. Lee said the watchdog was considering stabilizing measures but declined to provide specifics. South Korea’s KOSPI index has risen more than 110% in 2026, with Samsung Electronics and SK Hynix accounting for over half its weighting, Reuters reported. Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
Apple’s CXMT request tests US restrictions on Chinese chipmakers
Apple is reportedly requesting approval from the White House to buy memory chips produced by ChangXin Memory Technologies (CXMT). The Pentagon lists this firm among the military-linked companies from foreign nations, which means that Apple will have quite a hard time closing this deal. Apple needs to make this request because the increasing cost of memory powered by AI has started putting quite a lot of strain on the profitability margins of the company’s hardware offerings. Not only does Apple’s request show us that the AI revolution has changed the way semiconductors around the globe operate, but it also proves that hyperscale cloud service providers (which, of course, are data centers) compete with each other for getting their hands on scarce amounts of DRAM and NAND flash memory made either in North America or Asia. This implies that large corporations like Apple now have to take into account both the increases in the price of obtaining parts and Washington’s restrictions regarding national security. According to reports, Apple had reached out to the Commerce Department over a month ago to ask for permission to purchase memory chips from CXMT, which is China’s largest producer of DRAM chips. CXMT is a “military Chinese company” according to the U.S. Department of Defense, 1260H of the National Defense Authorization Act of June 2026. It is said that an inter-agency committee agreed to add CXMT to the list of companies in the Department of Commerce Entity List, with strict export license requirements for any goods, software, and technology manufactured in the U.S. However, there has been a delay in issuing the new Entity List because of extensive negotiations with China, Cryptopolitan reported previously. AI demand is reshaping the memory market Apple’s request is being made in an environment where AI infrastructure investment is generating one of the most vigorous upturns the memory sector has experienced in years. According to TrendForce, DRAM contract prices are expected to increase by 58% to 63% in Q2 2026, and enterprise SSD contract prices will also be on the rise by 48% to 53%, due to the rapid growth of demand from hyperscale AI data centres. Apple indicated this has affected them too, since they have just increased the prices of iPad and MacBook products, stating they will no longer be able to absorb the increased costs of memory and storage, as reported by the BBC. The iPhone maker has hiked the prices of some laptops and tablets by almost 20%. The price of its basic console will go up by $100 (£75) to $499, while the price of a console with more memory will go up by $150 to $749. New prices will take effect from August. Unlike logic chips, which are manufactured primarily at Taiwan Semiconductor Manufacturing Company (TSMC), global DRAM is dominated primarily by Samsung Electronics, SK Hynix, and Micron. Hence, there are not many choices available to customers. Because of the significantly greater DRAM and enterprise storage demands of AI servers, DRAM manufacturers shift to produce higher margins on server products and reduce supplies to makers of smartphones or PCs. Micron states that AI servers have seen a considerable increase in memory capacity to enable the operation of larger language models and data-intensive workloads. For Apple, gaining access to CXMT would not simply add another supplier—it would diversify procurement in one of the semiconductor industry’s most concentrated supply chains while potentially reducing exposure to rapidly rising contract prices. A frozen blacklist complicates the picture The Commerce Department’s Entity List has not been updated since October 2025—the longest gap in more than a decade, according to Philip Luck of the Center for Strategic and International Studies. Reuters separately reported that CXMT, DeepSeek and more than 100 other Chinese companies had completed the interagency review process but have not yet appeared on the published list. The United States has not put any companies in the “Entity List” since October 2025. According to former Commerce Department official Kevin Kurland, this indicates that trade policy takes precedence over a very important national security tool. According to Reuters, Undersecretary of Commerce Jeffrey Kessler has sought since late 2025 to avoid adding Chinese firms to the published Entity List while trade negotiations remain ongoing. CXMT is not currently included in the published Entity List as of now and, unless the Department of Commerce issues an explicit authorization to Apple, it is possible for them to create new supply chains independent of Samsung, SK Hynix and Kioxia without violating any existing export controls. Apple’s history with Chinese chipmakers In 2022, Apple examined the option of purchasing NAND flash memory from Yangtze Memory Technologies Co. (YMTC) for use in its iPhones sold in China. However, the company dropped its plans after the Biden administration introduced stricter export controls, and U.S. lawmakers warned that the deal could have security implications for the international technology supply chain. Later on, the Commerce Department added YMTC to the Entity List to show how quickly geopolitical factors can change semiconductor sourcing strategies. Washington’s actions will be a key indicator of how the administration will manage AI’s competitive advantages against national security restrictions. Should Apple gain approval, this would set a precedent for multiple U.S. technology firms struggling with AI-related memory shortages. More importantly, it will show the difference in outcomes from having the Pentagon classify a company as a military contractor versus the Commerce Department implementing export restrictions while Entity List changes are historically “frozen”. If Apple’s request is denied, Apple and other manufacturers of consumer electronics will remain under pressure to raise memory prices as investments into AI infrastructures become the largest consumer of global DRAM and NAND supply. The smartest crypto minds already read our newsletter. Want in? Join them.
Anthropic regains limited Mythos 5 access after two-week US government standoff
On June 26, the U.S. Commerce Department approved Anthropic’s request to resume selling Claude Mythos 5 to a select group of U.S. customers, allowing Anthropic to have its previous export-control order partially reversed. Two weeks prior to this order, the company was compelled to remove its two most advanced AI products, Mythos 5 and Fable 5, from the market. Meanwhile, Fable 5, the consumer version based on the same underlying architecture as Mythos 5, continues to be offline with no definite date for its redeployment. The partial reversal of the original export-control order has changed Anthropic’s competitive landscape at a crucial juncture in the company’s growth. Anthropic filed a confidential S-1 on June 1 with the SEC in preparation for a proposed IPO. Each additional day that these flagship models are offline allows competitors to fill that void with their own offerings. For example, only one day after the original export ban was imposed, Chinese company Zhipu launched its open-weight GLM 5.2 model. According to a CNBC report referenced by Cryptopolitan, Zhipu’s token traffic rose faster than any other Chinese model tracked in 2026. Why the US restricted Anthropic’s Mythos and Fable models On June 12, the U.S. Commerce Department issued a directive to Anthropic to block all users outside the U.S. from accessing Anthropic’s models, including Anthropic’s own non-American employees. Anthropic was unable to block foreign users quickly enough; therefore, both models were disabled for all users. The reasons given by the U.S. government for the block include cybersecurity concerns regarding methods of breaking into or bypassing Fable 5’s security features to find vulnerabilities in the software. Anthropic publicly disagreed with the government’s position, calling the findings narrow. The company argued that other models available for purchase, such as GPT-5.5, developed by OpenAI, could provide similar results in finding vulnerabilities without having to use jailbreak techniques. In addition to the jailbreaking dispute between Anthropic and the U.S. government, there are broader cybersecurity implications. During a congressional hearing, Senator Mark Warner of Virginia said NSA Director General Joshua Rudd told him Mythos could identify weaknesses across nearly all U.S. classified computer systems within hours. The NSA also conducted Project Glasswing, a covert effort with Anthropic and U.S. intelligence agencies to find and remediate critical software vulnerabilities, according to AP reporting and Cryptopolitan’s earlier coverage. A U.S. official later clarified that identifying vulnerabilities within hours did not mean Mythos could exploit them in the same timeframe. Still, the demonstration gave the administration political cover to act. What the Commerce Department now allows Mythos 5 to do According to a letter from Commerce Secretary Howard Lutnick to Anthropic co-founder Tom Brown regarding Anthropic’s models, the company appears to have made “significant progress” toward mitigating the risks associated with both Mythos 5 and Fable 5. The letter is dated June 26 and was seen by both Reuters and The Verge. In addition to granting access to the Mythos model without the need for an export license to some identified U.S. partners, Anthropic will also be able to provide access to non-U.S. employees of these registered U.S. partners and its own non-U.S. employees. A source familiar with the directive told Reuters that there are currently more than 100 companies and institutions approved to use Mythos 5, including a number of Fortune 500 companies. Anthropic spokesperson Danielle Ghiglieri confirmed the company had “received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers,” according to The Verge. Lutnick’s letter, however, did not reference the Fable 5 model. A source told Reuters that the U.S. government is progressing toward a resolution on the release of Fable 5 to the public, but no anticipated timeline has been established. How the shutdown helped Chinese AI rivals gain ground The ban on Anthropic’s Fable 5 and Mythos 5 was extremely inconvenient. The models were released on June 9 as Anthropic’s newest and best models, costing $10 per million input tokens and $50 per million output tokens. Three days later, both were forced offline. Zhipu’s GLM 5.2 launched to subscribers on June 13, one day after the shutdown. This model, which has 744 billion parameters, has scored within four points of Anthropic’s Opus 4.8 on Terminal-Bench 2.1 and costs approximately one-fifth of the Anthropic model, Cryptopolitan reported previously. Full open weights of GLM 5.2 were shipped on June 16 under an MIT license, meaning that a government cannot retroactively deny access to anyone after downloading it. OpenAI has a similar, but softer, limitation. The White House requested that OpenAI provide customers access to its latest model in an incremental format based on government approval from customer to customer. Axios reported that GPT-5.6 launched under restricted access, with only around 20 government-approved companies getting limited preview access. This similar gating model will now also apply to providing access to Mythos 5. Two of the three largest U.S. frontier labs now operate with government gate requirements. Why Anthropic’s government tensions matter before its IPO The Mythos suspension was not an isolated event. Throughout much of 2026, Anthropic’s relationship with the U.S. government was strained. The company would not permit the U.S. military to use its models to conduct domestic surveillance or work with fully autonomous weapons; the government reacted by placing the company on its national security blacklist. President Trump later changed his mind, telling Axios on June 19 that he no longer considered Anthropic to be a national security risk, an event covered by Cryptopolitan at the time. In addition, on June 10, Anthropic informed the Senate Banking Committee that operators of Alibaba carried out 28.8 million Claude transactions via about 25,000 fraudulent accounts between April and June 2026. Anthropic described this as the biggest attempt at extracting capabilities from Claude that it had ever seen, according to Cryptopolitan. Previous such distillation attempts have been made by DeepSeek, Moonshot, and MiniMax. What happens next for Fable 5 and Anthropic’s rollout Lutnick’s letter retained all other restrictions set forth in the June 12 letter but reserved the right to modify the terms should circumstances require. For Anthropic, the next step is ensuring that Fable 5 becomes available again before its proposed IPO. With each day the model remains unavailable to the public, enterprise customers continue to run cost-benefit analysis on open-weight alternatives from China that no government restriction can easily take back.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
China’s tungsten export controls and AI chip demand collide, reshaping the global specialty gas s...
A recent research report by China International Capital Corporation (CICC) published June 26 shows that increased spending on infrastructure for AI and the prohibition by the Chinese government on exporting high purity tungsten are putting pressure on the world’s supply of electronic specialty gases (ESG), which are a critical component needed to manufacture all advanced AI chips. The importance of the current disruption stems from the current combinations of two separate forces: The demand for these gases is growing rapidly due to AI’s growing influence and the supply chain through Japan has been broken. Chip manufacturers who are creating processors needed for data centers that support large language models, cloud-based machine learning inference systems and AI training facilities will face a bottleneck in their production, as will manufacturers of these gases. The bottleneck is not the capacity of fabs or packing/assembly but rather the semiconductor etching gases used to create electrical pathways that will limit production levels. AI chip demand increases pressure on specialty gases According to an in-depth report on ESG in semiconductor materials by Nanda Optoelectronics, ESG account for approximately 13% of the materials used in wafer fabrication. This makes ESG the second largest group of materials used in wafer fabrication, next to silicon wafers. What really differentiates the impact of AI from other things is not only the higher volume of chips being produced, but the fact that the total gas consumption per wafer has increased exponentially due to the decrease in transistor geometries. In 65nm process technology, there are typically about 20 etch processes for each wafer and in comparison, at the 7nm node, which is used to make many current AI accelerator devices, this figure has increased to approximately 140. Adding pressure to gas consumption is high bandwidth memory (HBM), which is the backbone of AI training hardware. Deep etch processes for creating through-silicon vias (TSVs) in HBM use sulfur hexafluoride and octafluorocyclobutane. Because of the increasing number of 3D NAND layers in AI-related non-volatile storage, the demand for these gases continues to grow. That being said, the dollar amount spent on increased semiconductor production as a result of AI continues to grow rapidly, as evidenced by forecasts for both foundries and cloud service providers. TrendForce has forecasted that global foundry revenue will be $218.8 billion in 2026, representing a 24.8% year over year growth. Capital expenditures made by the 8 largest cloud service providers would increase roughly 61% from the previous year, while global AI server shipments are expected to increase approximately 28%, according to the CICC report. China’s tungsten controls hit Japan’s WF6 supply In February 2025, Beijing placed export restrictions on tungsten and four other metals, requiring licenses for all twenty business-related products to export these materials, as reported by Reuters. These restrictions were framed as a response to U.S. tariffs and are designed to protect China’s national security interests. The downstream effects from these restrictions impacted Japanese manufacturers of tungsten hexafluoride (WF6) much more than expected. WF6 is an essential compound used in depositing tungsten interconnects into logic chips, DRAM, HBM, and 3D NAND. Japanese manufacturers Kanto Denka and Central Glass notified their customers in South Korea that they had run out of inventory and would likely be unable to meet sustainable supply through the second half of 2026. Japanese manufacturers account for approximately 24% of global WF6 production. There are few alternatives for filling the shortfall left by Japanese producers from suppliers located in Western countries. For instance, according to Reuters, the U.S. ceased tungsten mining operations in 2015 and has not produced refined bismuth since 1997. Chinese gas producers move to fill the shortage Rapid expansion of domestic Chinese gas production is being undertaken to capitalize on this opportunity. CSIC Special Gases, for instance, currently operates WF6 production capacity of 2,000 tons per year with plans to add an additional 1,000 tons to reach 3,000 tons by 2027, making it the world’s largest producer according to the CICC report. Meanwhile, HaoHua and Zhongju Core have capacity of 600 tons each and Heyuan Gas intends to commence trial production from a 600-ton facility this year. Price data is indicative of this transformation in production levels. Based on customs data from China’s General Administration of Customs, the average export price for WF6 between January and May 2026 exceeded RMB 950,000 per ton. Furthermore, by the end of June the market price quote for 6N (99.9999% pure) WF6 was between RMB 2 million and $2.5 million per ton. Electronic specialty gases become a strategic chip bottleneck The global ESG market is estimated to reach $6.9 billion by 2032, which represents a compound annual growth rate (CAGR) of 4.4%; it was approximately $5.1 billion in 2025, as per Persistence Market Research. The Asia Pacific accounts for 69% of ESG consumption, driven by semiconductor manufacturing concentration in China, Taiwan, and South Korea. According to the CICC report, the four major worldwide corporations, Linde, Air Liquide, Air Products, and Nippon Sanso, still control over 70% of the total electronic gases market. At present, approximately 25% of electronic gases for integrated circuit manufacturing in China are locally sourced. That number is climbing. The CICC report states that there is an opportunity for the production of ESG from China to increase simultaneously and aggressively gain both volume and price advantages, which is seldom the case for similar commodities along the supply chain. Chipmakers face new supply risks in AI hardware production Hyperscaler firms and the chip manufacturers that they fund both face immediate questions regarding potential delays to fab ramp schedules in South Korea and other countries due to limitations on the supply of WF6. Major HBM producers Samsung and SK Hynix have disruptions in their supply chains, which have largely relied on Japan until this point. In addition, the prevailing trend towards reliance on Chinese sources for specialized gases also presents a new dependency for an industry that is already experiencing challenges as result of U.S.-China technology restrictions. Acceptability of this trade-off will likely be determined by quality certification for purity of special gases, lead times for qualification of leading edge fabs, and the speed at which alternative sources of tungsten that do not originate from China can be found.
If you're reading this, you’re already ahead. Stay there with our newsletter.