Cryptopolitan

On Tuesday, Palo Alto Networks reported outstanding quarterly results and a new acquisition, which marks a big move towards a segment of enterprise technology that is growing very fast: security for AI agents.
The cybersecurity company (NASDAQ: PANW) has reported better-than-expected results for the fourth quarter and acquired Console, an AI-native platform designed to help organizations automate complex operational tasks.
This happened at the right time when the market is growing rapidly. According to Gartner, the worldwide spending on AI will reach $2.59 trillion in 2026, which is almost 47% higher than in 2025. The spending on AI cybersecurity will also grow, almost doubling from $25.9 billion to $51.3 billion in 2025, while in 2027 it may approach $86 billion.
For Palo Alto, Console is a bet that securing AI agents will become an increasingly important part of that spending.
Why Palo Alto is buying an agent company
According to Palo Alto Networks, Console lets organizations use AI to analyze problems and take action across their operations. Its technology will be integrated into Cortex.
CEO Nikesh Arora said customers will be able to “build agentic workflows in natural language” that identify and remediate problems automatically.
“This is the shift to software-as-an-agent, giving our platform the arms and legs to deliver autonomous security outcomes across the entire enterprise.”— Nikesh Arora, chairman and CEO, Palo Alto Networks
According to Console’s co-founder and CEO Andrei Serban, the platform’s key idea is very simple: users should only have to set an operational goal while smart software does everything else necessary to achieve that goal.
Palo Alto did not disclose how much it paid to acquire Console.
The quarter that beat the Street
Palo Alto has announced its revenue for the fiscal fourth quarter to be $3.41 billion, which reflects a 34% growth as compared to the previous year. Revenue surpassed analysts’ projections as they estimated revenue of only $3.35 billion for the company. They have also earned adjusted earnings per share of $1.02 but the analysts’ prediction was $0.98.
The projections announced by the company seems equally strong. They have stated that the company expects revenue of $14.10 million to $14.20 million in fiscal 2027, which exceeds the projection made by Wall Street of only $13.79 billion. The company is also expecting adjusted earnings to be $4.16 to $4.19 rather than $4.11.
The annual recurring revenue of its Next-Generation Security experienced an increase of 63% and currently stands at $9.10 billion. The company also projected approximately $11.13 billion for the mid-point of its fiscal guidance of 2027 and hopes this figure will increase to $20 billion by 2030.
This trajectory explains the Console acquisition. Palo Alto does not need to sustain today’s 63% NGS ARR growth; however, it must achieve over 20% growth per year after FY2027 in order to reach $20 billion smoothly.
However, this quarter was not without its hiccups. The company reported a GAAP net loss of $282 million, as opposed to a net income of $254 million during the same period last year.
Agentic AI is spreading faster than the guardrails
Demand for AI agents is rising, but governance is struggling to keep pace. McKinsey found that 40% of organizations with more than $1 billion in annual revenue are scaling AI agents, up from 27% a year earlier.
Yet Stanford’s 2026 AI Index found agent deployment remained in the single digits across nearly all business functions.
Security concerns are also becoming more concrete. Britain’s AI Security Institute recorded unauthorized autonomous actions in 10 of 122 cyber-testing runs, although it found no real-world harm. The International AI Safety Report 2026 also found strong evidence that AI can assist with several stages of cyberattacks.
The OECD has warned that cybersecurity and operational risks could slow adoption while also increasing demand for bundled AI and security products.
As Cryptopolitan reported in July, autonomous agents are increasingly viewed as “trusted insiders” because they can operate within legitimate permissions and still create security or compliance risks.
That is the opening Palo Alto is targeting. If AI agents become a routine layer of enterprise software, securing what they can see, decide and execute could become core infrastructure.
What to watch next
Execution now matters more than the announcement. Palo Alto still has to integrate Console while absorbing other recent acquisitions and prove the deal can strengthen Cortex.
For fiscal first-quarter 2027, the company has guided to revenue of $3.300 billion to $3.310 billion and NGS ARR of $9.54 billion to $9.56 billion.
Beyond whether Palo Alto beats those numbers, the bigger test is whether Console helps turn the AI-security boom into the sustained recurring-revenue growth needed to reach its $20 billion FY2030 targets.


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

According to a September 1 post from BitcoinTreasuries.NET, Strive’s preferred stock has raised enough money to fund Bitcoin (BTC) purchases for nine days in a row, most recently buying 104 BTC on Tuesday. It is noteworthy that while corporate balance sheets have become essential for Bitcoin demand, the number of buyers has sharply decreased. Thus, Strive’s purchases are among the few regular bids for the cryptocurrency at approximately $77,000.
The amount of $77,000 refers to the spot market price, rather than Strive’s acquisition cost. According to BitcoinTreasuries.NET, the live dashboard put BTC at $77,100, whereas Strive paid $79,431, in fees and costs included, between August 24 and August 28.
A lone bid in a market that lost its buyers
The timing of this trend is important. According to the Glassnode Strategy Watch Report on August 27, it was reported that Bitcoin treasury vehicles were net purchasers in July, even if the corporate buying had dropped as the spot ETF demand returned. The amounts held in these treasuries ranged from 2,300 Bitcoins to 7,600 Bitcoins, meanwhile, the flows for ETFs changed from -70,400 Bitcoins at the beginning of July to +5,400 Bitcoins by the end of July. Glassnode explained that this was a sign of growth of regulated demand and the decrease in balance sheet activity.
Galaxy Research had warned about recent strains on the digital asset treasury model and advised that a significant drop in equity premiums could make the new issuances of shares become more dilutive.
However, the situation with Strive is different. BitcoinTreasuries.NET noted the ongoing issuance of SATA that has not been accompanied by any buyback activities. This makes the nine-day BTC buying streak a good illustration of the real demand for the preferred stock.
What SATA actually is
SATA is not a subsidiary; instead, it is a ticker symbol for Strive’s Variable Rate Series A Perpetual Preferred Stock that is traded at Nasdaq together with Strive’s Class A common stock, ASST. In May, it was announced that SATA would pay dividends every business day beginning June 16 at an annual rate of 13.00%.
Strive CEO Matthew Cole called SATA “a true zero-to-one innovation” and said it is “the first listed security in the history of U.S. capital markets to pay cash dividends every single Business Day.”
That structure is the driving force behind the strategy. The March article by NYDIG’s Greg Cipolaro described SATA and the STRC of Strategy as:
“actively managed, capital markets–dependent liability structures backed by a reserve asset, bitcoin.” — Greg Cipolaro, NYDIG
The Digital Credit report for July published by BitcoinTreasuries.NET explains why investors find this approach appealing: preferred equity enables them to raise permanent capital without incurring any dilution of common shares or running debt refinancing risks, and converts Bitcoin collateral into a yield-generating investment. In the last round of funding in this sector, preferred equity was leveraged 3x-4.5x, or $3-$4.5 of Bitcoin for every $1 of preferred stock issued.
The purchases behind the streak
Strive has been ramping up its acquisitions for weeks. Its Form 8-K dated August 31 showed that it bought 1,800 BTC in the period from August 24 to 28, at an average price of $79,431, thus raising its total holdings from 21,356 BTC to 23,156 BTC.
According to Cryptopolitan, the approximately $143 million acquisition raised Strive’s position to the fifth place among the largest public holders of Bitcoin, whereas its weekly pace of acquisitions sped up from 20 BTC at the end of July, through 1,110 BTC to 1,800 BTC.
According to the research note published on August 31 by TD Cowen, the company raised its price target for ASST from $28 to $32 and expects Strive to reach nearly 4,300 BTC in acquisitions for the third quarter.
The 8-K document also sheds light on the manner of financing achieved for this acquisition, showing the growth of outstanding shares in SATA from 8.27 million to 9.07 million and in Class A common shares from 79.89 million to 83.47 million. BitcoinTreasuries.NET calculated that these two ATM programs raised around $154.6 million combined — approximately $80.3 million from SATA and $74.3 million from common shares — and $143 million went to Bitcoin.
Why the flywheel cuts both ways
The same reflexivity that powers the strategy can work in reverse. Cipolaro warned the loop:
“can abruptly halt if market confidence falters.” — Greg Cipolaro, NYDIG
Galaxy documented drawdowns exceeding 98% at Nakamoto and steep losses across DAT equities as premiums turned into discounts. Cryptopolitan noted that Strive itself booked a $257.6 million GAAP net loss in the second quarter, including $234 million tied to fair-value declines on Bitcoin and Strategy preferred holdings.
For the wider market, the signal is narrow but meaningful. With corporate Bitcoin demand increasingly concentrated among firms that can still tap capital markets, Strive’s ability to keep selling SATA and converting those proceeds into Bitcoin offers a clear test of how long this financing flywheel can keep supporting demand.


The smartest crypto minds already read our newsletter. Want in? Join them.

Spot XRP exchange-traded funds have seen about $1.6 billion in net inflows since they began trading. Investment advisers are the category buying the most, second-quarter 13F filings show.
Seyffart puts the tally at $1.8 billion
The funds stretched their inflow streak to nine consecutive days, taking in $26.2 million on August 28, according to SoSoValue data. The products pulled in over $725 million over that nine-day window.
Cumulative net inflows reached about $1.6 billion. Total net assets were about the same. Daily hauls have been spotty, anywhere from about $2.4 million to over $28 million.
“XRP ETF flows have been surprisingly resilient. Money in the aggregate has mostly only gone one direction and now they have a total of $1.8 billion in cumulative net inflows. When you look at this compared $XRP price over this time period … its particularly impressive,” posted James Seyffart on X.
According to his own calculations, net inflows totaled about $1.8 billion, which is more than the SoSoValue number.
Compared to XRP’s price over the same period, Seyffart said the run was “particularly impressive.” XRP was the odd one out this week, still drawing in cash, as Bitcoin funds ended a nine-day streak of inflows.
Goldman Sachs holds about $87.4 million across XRP funds
Using 13F filings from the second quarter, Seyffart broke down the holders. Goldman Sachs has the most spot XRP ETF exposure, with about $87.4 million. Jane Street and Millennium Management are next, each with about the same amount of $16.6 million.
By type of firm, investment advisers held and moved the most money during the quarter, far ahead of hedge funds and brokerages. That composition points to buy-and-hold client portfolios.
To start, Canary Capital’s XRPC was used to launch US spot XRP ETFs on November 13. By mid-December, they had about $1.18 billion in assets and had seen 30 days of straight inflows.
Ripple’s fight with the US Securities and Exchange Commission ended in August 2025, when both sides dropped their appeals and left a $125 million penalty standing.
Judge Analisa Torres had imposed that penalty a year earlier, after ruling in 2023 that XRP was sold as an unregistered security to institutions.
According to CoinGecko data, XRP was trading at around $1.35, down 2.44% on the day and 7.07% on the week.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Britain’s National Crime Agency (NCA) has halted approximately $13.5 million (£10 million) that Sorare made to the Premier League, raising new questions regarding crypto companies cashing in on the global sporting event.
These funds were transferred by Sorare as a part of its four-year-long licensing agreement with the League and fall under civil-recovery investigation targeting “alleged third-party criminality.”
Sponsoring teams in football allows crypto businesses to connect with millions, but the freeze has proven the point that such payments may attract regulators’ attention as well.
Why one frozen payment rattles the whole sponsorship model
An order to freeze the accounts mentioned in the ICLG report was issued by the Westminster Magistrates’ Court in January of 2025 but has only been made public by The Sun this week. The money involved is in the amount of £10,024,041.33 and is deposited in a Barclays bank account that belongs to The Football Association Premier League Limited.
The NCA stated that the aim of the order is “to prevent dissipation of the funds while the NCA investigates any potential links between those funds and alleged third-party criminality.” The Premier League has not been accused of any wrongdoing and has requested the court to change this order, which is valid until September 14.
According to ICLG, this was the initial payment for a huge 2023 licensing arrangement which is estimated to be worth about £120 million over a period of four years. This agreement permitted Sorare to create digital cards highlighting all 20 Premier League teams’ players. Sorare state that the deal will end in 2025-26 season.
A company built crypto-first, valued at £3.21 billion
Sorare is important to the cryptocurrency ecosystem because it has transformed from a business built on blockchain foundations into one of the most renowned brands in football overall. Starting in 2018 with the creation of a marketplace built on Ethereum, in 2023 it introduced fiat payments through its Cash Wallet, and in late 2025 moved its digital cards to Solana, adding payments via SOL cryptocurrency.
The company’s investors and ambassadors include global football stars Lionel Messi, Kylian Mbappé, Zinedine Zidane, and Rio Ferdinand as well as tennis superwoman Serena Williams. According to The Sun, Sorare has been valued at £3.21 billion.
“Account freezing orders are made by magistrates on application and, importantly, the proceedings are civil rather than criminal in nature.” — Nick Brett, partner at Brett Wilson
Brett warned that the funds may point towards transactions made through Sorare or be of “an entirely separate origin,” which means that the order itself should not be taken as an indication of any wrongdoing.
Regulators are watching where gambling meets digital assets
Sorare’s difficulties with UK regulation do not end with the NCA freeze. The Gambling Commission accused the company of conducting gaming activity in the UK without the required license under the Gambling Act 2005. According to the latest update from the Gambling Commission dated May 18, the trial of Sorare will take place on June 7, 2027, at Birmingham Magistrates’ Court. The company pleads not guilty, claiming that the nature of the game is skill-based rather than dependent on luck.
Authorities “will be particularly vigilant where gambling and digital assets meet, regardless of the profile of the company involved.” — Chris Roberts, head of white-collar crime at Grosvenor Law
According to Roberts, greater amounts of money increase the chances that authorities will issue freezing orders in order to secure assets for restitution.
The FCA has already put clubs on notice
The case is part of a larger initiative in the UK to scrutinize sponsorships in football. On June 3, 2026, the Financial Conduct Authority (FCA) alerted various clubs about the dangers of working with unauthorized financial and cryptocurrency firms and recommended that appropriate caution be adopted in such instances. Cryptopolitan reported the warning at the time.
A later FCA freedom-of-information release showed it wrote to 21 clubs and identified 18 sponsorship arrangements involving 13 clubs and firms it had not authorised. The FCA stressed that lack of authorisation does not itself prove unlawful conduct.
FCA consumer-investments director Lucy Castledine said clubs should not allow unauthorised firms to exploit supporters’ loyalty by promoting potentially questionable products to millions of fans.
“A logo on a shirt means one thing: that firm paid for it.” — Lucy Castledine
The warning comes as digital-asset brands continue buying high-visibility football inventory, including Circle at Chelsea, OKX at Manchester City and Kraken at Tottenham.


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

Anthropic has signed a $35 billion cloud-computing deal with Lambda, an Nvidia-backed provider, for a Texas data center. Nvidia already holds the lease on that campus, which Hut 8 is building in Nueces County.
Some of the capacity will sit at Hut 8’s Beacon Point campus. Reuters put the Anthropic portion at about 350 megawatts, which is about half of the 704 megawatts Hut 8 has leased there.
Hut 8 has never named the tenant on 704 megawatts
Hut 8 is building out the site and has named its tenant only as a high-investment-grade counterparty. Lambda will install the chips, and Anthropic is the buyer.
Nvidia signed leases worth up to $50 billion for the campus, the Financial Times reported on July 28, citing five people familiar with the deal. That was five weeks before the Anthropic agreement surfaced.
The Wall Street Journal, which first broke the Lambda deal, said Nvidia would supply the chips and hold the lease, citing people familiar with it.
“Nvidia is working with ecosystem partners to accelerate the deployment of efficient AI infrastructure through the DSX AI factory architecture,” a spokesperson said, without confirming or denying the report.
Beacon Point is 525 acres with access to up to 1 gigawatt of power and a grid connection already built.
Getting hundreds of megawatts connected to the grid from scratch takes years. Miners already run large sites on cheap power and completed interconnections.
HUT popped on the news before pulling back, closing down 0.98% at $78.64 on Monday, August 31, and slipping further to close at $77.57 on Tuesday, September 1.
Riot withheld a tenant name on 191 megawatts too
On May 6, Hut 8 announced the first Beacon Point lease, a 15-year, 352 megawatt agreement with a base-term value of $9.8 billion. If all renewal options are exercised, that figure could rise to $25.1 billion, the company said.
That commitment brought the miner’s total contracted AI capacity to 597 megawatts and ~$16.8 billion in total base-term value.
A second 352-megawatt lease was signed July 20, doubling the same tenant’s campus capacity to 704 megawatts. It lifted the campus base-term value to $19.6 billion, or as much as $50.2 billion if renewals are exercised.
It’s a triple-net lease, so the tenant pays tax, insurance, and maintenance on top of the rent.
In August, Riot announced a 20-year, 191-megawatt lease at its site in Rockdale, Texas. The base term is worth $9.1 billion, and two five-year extensions would lift it to $16.1 billion.
Riot initially refused to identify the tenant, saying it was “one of the world’s leading frontier AI labs,” before the link was made to Anthropic, Cryptopolitan reported.
Riot expects to have as much as 96 megawatts available at Rockdale by December 2027 and fully deployed by June 2028.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Michael Nicolls, a senior Starlink executive, has been handed control of SpaceX’s AI infrastructure following a series of leadership departures.
SpaceX (NASDAQ: SPCX) faces growing pressure to maintain stable operations for its compute customers, including Google and Anthropic.
What went wrong with SpaceX’s data centers?
Michael Nicolls has been appointed to oversee xAI’s infrastructure after several senior data center executives left the company over the recent weeks.
The system Nicolls will be taking over is one SpaceX has bragged about building at a low cost. The company said its second data center brought its first two computing clusters online for roughly a quarter of what is typically paid in the industry. However, the company did not spell out which expenses were included in that estimate.
The Information reported that SpaceX compromised other factors in order to build its data centers quickly. For instance, xAI’s Macrohard site used temporary equipment to get running, including mobile gas turbines, Tesla Megapack batteries, and more than 100 portable chillers.
A person who worked on the project reportedly stated that this makeshift setup caused outages that disrupted AI model training. The site’s uptime consistently fell below the company’s internal goal of 99.9%.
Some SpaceX facilities in Tennessee and Mississippi reportedly ran for months with no backup cooling or power at all, due to the company’s race to stack up compute. SpaceX has since pushed back plans for data centers beyond its Memphis hub, including proposed sites in Texas.
The delays cast doubt on the company’s plans to generate more than two gigawatts of computing capacity by the end of 2026. SpaceX reported in June that it only has about 1.4 gigawatts online.
What is SpaceX doing to fill its vacancies?
Following the exit of several executives like infrastructure chief Jake Palmer in July, SpaceX is moving employees from its rocket and Starlink programs to fill the gap in its data centers.
For instance, Wesley Salandro, who used to be a senior production manager for the Falcon and Dragon rocket lines, has joined the data center team. Logan McConnell, a product-support director at SpaceX’s Starbase site in Texas, has also been moved to a new role.
SpaceX has also brought in an engineer from Neuralink, Elon Musk’s brain-implant company, and is now hiring for several civil-engineering positions.
What happens if SpaceX’s data centers fail?
Selling computing power to Anthropic and Google increased SpaceX’s AI revenue to $2.6 billion in the second quarter of 2026, compared to $737 million in the same period last year.
Google reportedly has a compute deal worth $920 million per month for about 110,000 Nvidia GPUs, while Anthropic has a lease on SpaceX’s Colossus system that costs as much as $1.25 billion per month, although that agreement can be canceled.
An earlier report from Cryptopolitan said that SpaceX’s Colossus contracts with Anthropic, Google, and Reflection AI are expected to generate over $28 billion each year.
If SpaceX cannot keep its systems running reliably, that revenue is at risk. The company must either spend more money to fix the issues or reduce the computing power available to its paying customers.
Despite these concerns, SPCX closed near $143.36 on September 1, just below the previous day’s close and roughly 36% below its June peak of $225.64. The stock has bounced back from an August low of about $104.83.
Elon Musk is trying to take control of more of the supply chain, confirming on Saturday that SpaceX is building a foundry in Bastrop, Texas, to make its own gas-turbine blades and vanes.
He said this move could speed up natural-gas turbine deployment by as much as 18 months, but the NAACP has accused SpaceX of operating turbines without proper permits at its Memphis site.
If you're reading this, you’re already ahead. Stay there with our newsletter.

President Donald Trump has attacked the growing backlash against AI data centers, warning in a Truth Social post that communities blocking the facilities risk becoming “backwards and poor.”
Trump called the data centers a “golden goose” and told Americans they will have themselves to blame if projects are canceled.
President Trump wrote on Truth Social on Monday, “If they want to be successful and rich, with far lower taxes and jobs all over the place, let Data Reign,” the president wrote.
He also claimed China welcomes the resistance, saying Beijing “could not be happier with this anti Data Center movement.”
The poll numbers Trump is pushing against
The posture of the president from his post shows that he is not in agreement with most of the country. This is because a Gallup survey conducted this spring found that 71% of Americans oppose building AI data centers near where they live. That figure has been well cited. In addition to that, more than six in ten Americans reportedly said they would strongly oppose such a proposal.
The opposition seems to also cross party lines, as both Republican and Democratic voters have grown skeptical about these facilities. The leading concerns are that the facilities consume large amounts of electricity and water, raising fears of grid strain, higher utility bills, and, for residents living nearby, the constant hum the buildings emit.
In Michigan, Republican Senate candidate Mike Rogers recently backed a one-year pause on new projects. Abdul El-Sayed, the Democratic Senate candidate for Michigan has called for an indefinite pause until federal rules require union jobs and protect water supplies.
Vance splits from the hard line
Trump’s own vice president struck a softer note the same day. JD Vance told reporters that data centers are “an important part of the AI economy.”
However, he conceded that most of the anger traces back to cost. Vance said, “I think probably 99% of the backlash to datacenters has come in areas where building a datacenter means higher utility and higher electricity for the people on the ground.”
His fix was pointed at the companies, as he said that developers should build power plants alongside their facilities and lean on federal deregulation.
“If you build a datacenter, you should be putting power back into the grid, not taking it out,” he said. “And if that is happening, I don’t think the datacenters are that controversial.”
A wave of moratoriums and blocked projects
The resistance is already affecting where these facilities get built; in some cases, it has paused ongoing projects.
Research group Data Center Watch found that grassroots groups blocked or delayed at least 75 projects worth around $130 billion in the first three months of 2026. Communities that fought back and won include Monterey Park near Los Angeles, Prince William County in Virginia, and Wake County in North Carolina.
In Pennsylvania, the governor signed an executive order on August 18 requiring local approval for projects. In July, New York Governor Kathy Hochul enacted a moratorium. All approvals in Texas have been paused pending an audit after Governor Greg Abbott issued a moratorium to that effect in August.
The pressure has reached inside Trump’s party. A National Republican Senatorial Committee memo warned AI firms that data-center anger was negatively affecting Senator Jon Husted in Ohio. Residential electricity rates in the state have gone up by 175% since 2005, according to U.S. Energy Information Administration data.
What the administration is doing to help builders
Trump signed an executive order last year fast-tracking federal permitting for AI data centers.
The administration has pushed for expedited grid connections and faster approvals of chemicals used at the sites. The Environmental Protection Agency (EPA) has been planning to scrap a rule requiring public notice and a comment window for air permits.
So far, that plan has courted widespread criticism.
To ease the uproar, the administration has backed a voluntary Ratepayer Protection Pledge with several major AI companies signing on to offset local grid costs.
Despite this, lawsuits continue to pile up on the facilities. Some of the lawsuits accuse firms, including SpaceXAI and Microsoft, of violating pollution-permitting laws, and a Microsoft-backed New Jersey site known as DataOne has been accused of running unpermitted gas-powered generators.
With the US midterm drawing closer, the fight is also making its way to the polls as AI-focused super PACs have ramped up their funding. They have reportedly raised $107 million and spent $55.5 million on federal races this cycle.
Pro-industry Leading the Future network, backed by Andreessen Horowitz and OpenAI President Greg Brockman, has raised $140 million. On the other side, the pro-regulation Public First Action has raised $80 million, including $40 million from Anthropic.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Anthropic launched a new model on September 1 called Claude Fable 5.1.
The model immediately claimed the number one spot on Artificial Analysis’s intelligence leaderboard with a score of 66 on the site’s index, dethroning Opus 5 in the process.
Fable 5.1 and Opus 5 sit atop the leaderboard
Artificial Analysis ranks over 250 language models on price, speed, and intelligence. It now ranks two Fable 5.1 iterations first and second on the table. The site scores the “max with fallback” at 66, while scoring the “xhigh with fallback” at 65.
Both tower above former leader Claude Opus 5 with a score of 63 in its max and xhigh modes.
That means Anthropic has the top four spots on a leaderboard that has models from top AI labs like OpenAI, Google, SpaceXAI, Alibaba, and DeepSeek. The closest to any Anthropic model is OpenAI’s GPT-5.6 Sol at max mode and SpaceXAI’s Grok 4.6, both scoring 61.
AI model leaderboard. Source: Artificial Analysis
The ranking comes from an independent party, giving it more validity than a lab’s own charts.
Where the coding and research scores landed
Anthropic internal numbers tell a similar story, even though they ought to be read as vendor-reported. Anthropic’s reporting ranks Fable 5.1 at 52.6% on Terminal-Bench-Science 0.1, which is a test of agentic scientific research. That figure is double that of Fable 5’s 24.7% and miles ahead of the 29% and 22.4% of Opus 5 and GPT-5.6 Sol, respectively.
Fable 5.1 scores 55.8% on the Terminal-Bench 4.0 coding benchmark, higher than Fable 5’s score of 42.0%.
The selling point is the ability of this new model to do work that runs for hours. Millennium told Anthropic that Fable 5.1 was able to trace a rare crash in its system to a bug that had proved too stubborn for its engineers for the past four to five years.
Browserbase said the new model completed 82% of tasks on its hardest browser-agent test, compared to 74% for Opus 5.
A 75% cut to cache-read pricing
There was no change in price, though. Fable 5.1 maintains Fable 5’s rates of $10 per million input tokens and $50 per million output tokens, way more than Opus 5, which costs $5 and $25, and Sonnet 5, going at $2 and $10.
The change occurs in the price of cached context. Anthropic reduced the cache-read price to $0.25 per million tokens, from $1.00, a 75% cut.
Anthropic estimates that the average workload will become 25% cheaper, while heavily agentic workloads will become 45% cheaper. This is as a result of agents’ ability to reread the same code, instructions, and conversation history.
Same model, two safeguard tiers
Anthropic launched a second name with Fable 5.1: Claude Mythos 5.1. They are basically the same models, but with separate safeguards. Fable 5.1 is available to the general public, but Mythos 5.1 is available only to vetted cybersecurity and life-sciences groups via Anthropic’s Project Glasswing.
That split comes after a tough period for Anthropic’s safety testing. As Cryptopolitan reported, Anthropic put a pause on external cybersecurity evaluations on July 23. This came after Claude got to real systems during tests meant to be sandboxed.
The company resumed external cybersecurity evaluations once it was able to add appropriate containment measures.
The smartest crypto minds already read our newsletter. Want in? Join them.

Amazon switched on “Update Me When” on Tuesday. Alexa for Shopping sends a notification when a product, show, book, or tour that a shopper follows goes live.
The assistant scans the web as well as Amazon’s catalog
Amazon announced the feature the same day as a batch of other shopping tools.
The assistant mostly waited for the user to ask, but the new alerts changed that. The software watches on its own and speaks first.
Rajiv Mehta, Amazon’s vice president of conversational shopping, wrote the company’s description.
The virtual assistant searches Amazon’s catalog and the web for relevant items and then sends a personalized alert, according to Mehta.
When a brand starts a new line, an author publishes their next book, a musician announces a tour, or a gadget maker sets a ship date, an alert goes off.
Shoppers set the alerts by asking in the Amazon Shopping app or on Amazon.
Examples from the company include “Update me when Reacher Season 5 drops on Prime Video,” “Update me when a new Kindle releases” and a request to be notified when cooperative board games for four or more players arrive at the store.
“Update Me When” expands Amazon’s Alexa for Shopping capabilities.
Amazon originally debuted its generative AI shopping assistant, Rufus, in February 2024 before expanding and rebranding its conversational agentic tools under the Alexa for Shopping umbrella in May 2026.
Amazon’s own illustration of the Update Me When flow, from its Alexa for Shopping page, updated September 1, 2026. Source: Amazon News.
Google and Walmart are building rival checkouts inside Search and ChatGPT
The system is already monitoring prices and will either notify a shopper or purchase an item outright once it drops to a target.
It also builds carts from previous orders, shows 30-, 90-, and 365-day price history, reads handwritten lists into a cart, and writes tailored shopping guides.
Amazon also spotlighted another feature, Scheduled Actions, which can replenish household staples on a recurring basis.
According to Cryptopolitan, Amazon introduced “Help Me Decide” in October 2025.
“Help Me Decide” analyzes browsing and purchase history to deliver a single, clear product recommendation when shoppers stall between similar items.
Google now lets people buy directly inside Search and Gemini from retailers including Walmart, Wayfair, and Etsy, Cryptopolitan reported in February 2026. The in-answer discounts are called Direct Offers.
Walmart struck a deal with OpenAI to let customers browse and purchase items via ChatGPT.
If you're reading this, you’re already ahead. Stay there with our newsletter.

New York’s financial regulator (NYDFS) has informed X, owned by Elon Musk, of its inability to pay interest on X money balances belonging to New York customers. NYDFS went on to say that the yield on such accounts will drop to 0.00% on October 1, 2026.
The announcement yanks one of X’s major perks from its users in what many would consider the largest financial market in America, barely weeks after X Money was launched across the country.
APY falls to zero as X never obtained a New York license.
Sawyer Merritt spotted the change and stated so in an X post on September 1, 2026. According to the post, the New York Department of Financial Services (NYDFS) informed X that it cannot pay interest on X Money balances and that New York accounts will receive no interest as of October 1.
X Money relies on yield as a marketing tool to attract users. When X Money was launched, it advertised a 6% APY for Premium Plus subscribers, while standard Premium users would enjoy such a rate once their deposits crossed a particular threshold or met a direct-deposit requirement.
Also, X offered users a metal Visa debit card linked to a user’s X handle, no fees on peer-to-peer transfers, and Apple Pay and Google Pay support.
For New Yorkers, however, the choice to earn interest is gone.
New York and Massachusetts not part of the nationwide launch
X Money was made available to all Premium and Premium+ subscribers with US accounts on August 31. This was announced by the company in its post. However, not all states were covered in the nationwide rollout.
When X Money was launched, X Payments LLC, a money-services subsidiary of X, had obtained money transmitter licenses in 41 states and D.C., with New York and Massachusetts absent.
Around 4.4 million Premium subscribers are eligible to use X Money, and as of its July rollout, New York had not made a decision, and X Money was still without a license.
Today’s decision did not come without precedent. State Senator Brad Hoylman-Sigal and Assembly member Micah Lasher encouraged the NYDFS to deny X a money transmitter license in a May 2025 letter. They cited Musk’s conduct at DOGE and worries about X’s management of consumer data.
NYDFS is hard to please
The pause in interest is consistent with the agency’s reputation. Cryptopolitan has previously reported on the strict operation that the DFS runs as a state-level financial regulator. It’s regarded as a pacesetter by banks and crypto issuers.
In July, after Circle had received a limited purpose trust charter from the DFS, Circle’s CEO, Jeremy Allaire, called the DFS “an international standard setter for digital asset regulation.”
The department has begun tightening rules. Cryptopolitan reported in June that the DFS proposed new stablecoin regulations intended to match the GENIUS ACT while it maintains its existing safeguards. This includes full reserve backing and independent audits.
The smartest crypto minds already read our newsletter. Want in? Join them.

Two unions that together represent roughly 10,000 Micron workers in Taiwan are preparing to organize a strike vote over the way the company calculates bonuses.
A strike could make the current chip shortage even worse and lead to higher prices for phones, laptops, and other electronics.
80% of Micron employees support going on strike
Micron Technology (NASDAQ: MU) in Taiwan, which is one of only three companies that control about 94% of the global DRAM market, could be facing a potential workers’ strike following complaints about how employee bonuses are calculated.
Taoyuan Union and Taichung Union, which together represent about 10,000 workers, are currently in talks with the company, but if negotiations fail, a formal strike vote could happen in September.
A recent poll revealed that about 80% of workers support a strike.
The complaints regarding Micron’s bonus system, called the “Incentive Pay Plan,” include that the system is not clear and makes it hard for employees to reconstruct. Workers want a simpler profit-sharing plan, like the ones used by its main rivals, Samsung and SK Hynix. The system used by the company’s rivals also pays more.
Samsung ties employee bonuses in its chip division to 10.5% of profit, while SK Hynix earmarks 10% of annual operating profit. Both figures are published and pegged to a published result, so a worker can check the math against the company’s own earnings.
However, at Micron, annual performance bonuses for its Taiwan staff are capped at 200% of target, worth about five months of pay at the ceiling, yet the payouts have actually landed near 2.6 months on average. The plan asks employees to trust a calculation they cannot see, which is the heart of the grievance.
What would a strike mean for the current chip shortage?
Strikes are very uncommon in Taiwan’s chip factories because the work is highly specialized, and the workers who run the production lines are not easy to replace. A strike at Micron now would cause serious problems for the company’s large DRAM chip-making and advanced packaging operations in Taiwan.
Not only can the work not be moved to other countries quickly, but there is also very little spare production capacity available elsewhere in the world.
Demand from AI data centers has pushed memory prices up a lot, greatly increasing profits for chip makers. Semiconductor shipments in South Korea went up by 209% compared to the same month last year, reaching $46.65 billion.
DRAM prices reportedly went up by 80% to 90% in just one three-month period due to supply being moved toward high-bandwidth memory, which is used for AI computers.
SK Hynix chief Kwak Noh-Jung commented in late August that he expects the chip shortage to continue until the end of 2030.
Apple’s Tim Cook, on his final earnings call as CEO, warned that memory prices would keep rising and squeezing profits.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Morpho’s outstanding loans have reached an all-time high of $5 billion. Crypto analytics firm Messari reported the metric on Tuesday, September 1.
This brings the decentralized lending protocol closer to Aave, which is the longtime leader in that space, and it also raises its profile among the fintechs and stablecoin issuers routing credit through it.
Lending protocol ranked by TVL. Source: Defillama
Stablecoins carry almost all of the borrowing
The composition of that $5 billion is lopsided, as Messari reported that 95% of the borrowing is denominated in stablecoins, with USDC accounting for 62% of it. It shows that most users are using it for dollar credit against crypto collateral and not leveraged speculation on volatile tokens
Morpho loan compositions. Source: Messari
On August 18, Messari flagged that Morpho Blue loans had reached their highest level since October 2025 and were approaching new highs, with Robinhood and Base together responsible for more than $650 million in fresh loans.
A week before that, Morpho’s markets generated over $4 million in interest in one week. It is the most since November 2025, and for one of the few times, its Base markets out-earned its Ethereum markets.
Base did the heavy lifting
Base has become Morpho’s engine. In an August 6 post, Morpho said its deposits on the Coinbase-built layer-2 had passed $5 billion, more than 70% of everything deposited on the network.
The protocol credits its “DeFi Mullet” model, the crypto-backed loan product Coinbase launched on Morpho rails in January 2025. According to Morpho, it has since driven around $1.3 billion in outstanding USDC borrowing backed by about $2.5 billion of cbBTC collateral.
Aave still leads, but the cushion is thinner
While Morpho is recording an increase, it is still not enough to unseat Aave, which leads the industry.
DefiLlama data shows that Aave has over $12.7 billion in active loans against Morpho’s $4.83 billion and $17.7 billion in total value locked, to Morpho’s $9.55 billion.
Aave also suffered losses after the $292 million KelpDAO exploit in April, losing more than $10 billion in total value locked. Aave’s backers rallied, and stakeholders publicly bet on a recovery.
Morpho, which had only minor exposure to the same April hack, has been growing into the space that opened up.
Named in DeFi’s strangest blowups
That growth has a shadow side. Morpho’s scale keeps surfacing in incidents that are not failures of its core code but of the markets and curators built on top of it. DeFi researcher DeFi Warhol pointed on August 26 to a Paxos PAXG oracle misconfiguration, where a decimal error overvalued collateral so badly that one borrower drew roughly $230,000 in USDC against about $350 of PAXG.
Marc Zeller, who used to be a prominent Aave contributor, needled the protocol on August 25, writing that user losses on Morpho were “an inch from a total of $150m” and jabbing, “They run our marketing dept.” Not everyone reads the losses as failure.
Euler cofounder Michael Bentley argued on August 27 that lenders can lose money and that the interest they earn is “compensation for taking” risk, pushing back on the idea that a protocol or curator must always be made whole.
A $175 million war chest
Morpho is not short on capital or credibility. The protocol reportedly raised $175 million in June, in a round led by a16z crypto, Paradigm, and Ribbit Capital, with Circle’s venture arm and VanEck also participating, valuing it at up to $2 billion.
Cofounder Paul Frambot, 25, started the project at 20 and initially built on top of Aave before pivoting to let anyone “build their own Aave” with custom risk parameters. Coinbase, Kraken, Anchorage Digital, and Galaxy Digital are among its users.
If you're reading this, you’re already ahead. Stay there with our newsletter.

DeFi Development Corp. (DFDV) said Monday it is seeking to raise up to $20 million through the sale of a new perpetual preferred stock.
The Nasdaq-listed Solana treasury company will pay a 13% annual dividend in daily installments, with the cash earmarked mostly for buying more SOL.
R.F. Lafferty holds a 30-day option on an extra 15%
Officially, it is called the Variable Rate Series C Perpetual Preferred Stock. The company dubbed it “CHAD Stock,” per its August 31 press release.
Each share has a stated amount of $10.00. The initial dividend rate is 13.00% per annum.
Instead of a quarterly payout structure, DFDV stated dividends will accrue cumulative value and pay out on each business day when declared by the board, with the initial payment scheduled for October 1, 2026.
The rate is variable, and the initial 13% may vary depending on the terms of the stock.
R.F. Lafferty & Co. is the sole book running manager. The company also granted its underwriter a 30-day option to purchase an additional 15% of the shares sold.
DFDV said it will establish a dividend reserve when the offering closes. It plans to set aside $1.30 per share, equivalent to a full year of dividends at the 13% rate, according to its filing.
The account will be funded from existing cash, financial instruments, or digital assets. The Boca Raton company is registering the sale under a Form S-3 shelf statement, the SEC declared effective April 27, 2026.
Net proceeds are for general corporate purposes, DFDV said. The list includes working capital, strategic transactions, growth initiatives, and buying more SOL and other digital-asset investments.
The company describes itself as the first U.S. public company built around accumulating and compounding Solana.
A ZeroStack sale helped fund 19,000 SOL at $98.14
In a separate statement on Aug. 27, DFDV said it had restarted buying SOL. It purchased about 19,000 tokens at an average cost of $98.14, leaving it with a treasury of about 2,333,432 SOL and SOL equivalents.
The company sold its ZeroStack position to help fund part of that purchase.
“DFDV is designed to provide investors with leveraged exposure to Solana,” Chief Executive Joseph Onorati said last week.
He said the company’s blend of “amplified SOL exposure, strong trading liquidity, and differentiated treasury yield is the DFDV model at work.”
Shares of DFDV climbed 8.03% Monday to close at $5.38, up 110% in the past month but flat for the year. SOL was trading around $101. That’s a long way from the 52-week peak of $18.99 the shares once touched.
Cryptopolitan reported the company expanded a share buyback program to $100 million as its market-value-to-net-asset-value ratio slipped to 0.9.
DFDV warned that there is no assurance it will be completed.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Cathie Wood’s Ark Invest bought 456,059 shares of Jack Dorsey’s Block Inc. on Monday. The stake is valued at about $37.4 million.
Block closed at $82.02, down 1.85%, according to a trading disclosure.
~$21 million August buy followed a 6.15% drop to $79
The buy was divided among three of Ark’s actively managed funds. They are the Ark Innovation ETF, the Ark Next Generation Internet ETF, and the Ark Blockchain & Fintech Innovation ETF, which trade under the symbols ARKK, ARKW, and ARKF, respectively.
Ark did not pool the purchase into one vehicle. Its disclosures cap any one holding at 10% of a fund, a limit intended to keep its portfolios diversified.
Ark has seen the name Block on tape repeatedly all year. In June, the firm bought 236,759 shares, or about $17.2 million worth, through ARKK.
A smaller deal made in July involved 19,029 shares through ARKW and ARKF at a close of $79.99, worth about $1.52 million.
Monday’s buy seems to be a housekeeping matter against Ark’s own rules. The firm’s ~$21 million Block purchase that month was routine rebalancing rather than a new bullish bet, Cryptopolitan reported in August.
That purchase followed a 6.15% decline to $79. Ark adds to positions that slip below their target weight and prunes those that run too hot.
Cryptopolitan reported that Ark bought Block in the same August session and sold 39,509 shares of Bullish through ARKW.
One crypto-linked stock bought and one sold the same day tracks portfolio weights, not sector conviction.
Cash App brought in $1.97 billion of second-quarter gross profit
Block’s fundamentals gave Ark a reason to keep the stock at target weight. Second-quarter gross profit gained 25% to $3.17 billion from a year ago.
Cash App tallied $1.97 billion and Square $1.16 billion. Operating income, adjusted, was $864 million on a 27% margin.
Adjusted diluted earnings came to $1.02 a share. Block lifted its full-year gross profit forecast to $12.51 billion from $12.33 billion previously, pointing to 21% growth.
Since axing about 40% of staff in February, the company has leaned hard into automation.
In June, Block introduced its Builderbot AI tool. This tool merges ~1,500 pull requests per week, accounting for about 15% of all production code changes across the company.
It said its internal Builderbot tool merged about 1,500 pull requests per week.
Mizuho flagged Block’s operating expenses even after the layoffs. The bank reckoned Block’s adjusted costs would increase from $4.48 billion in the first half of 2026 to $4.56 billion in the second, based on Block’s own guidance.
Mizuho retained an Outperform rating and a $100 price target, while saying it’s unclear if more spending is required to grow Cash App’s active users.
Ark’s Monday trading extended beyond Block (XYZ), purchasing 35,192 shares of Circle Internet Group (CRCL) valued at about $3.36 million as Circle surged 9.65% to $95.55.
Bernstein last week reiterated an Outperform rating and $140 target on Circle. The stock has soared 52.6% in a month.
If you're reading this, you’re already ahead. Stay there with our newsletter.

The Pentagon on Monday pushed military versions of OpenAI’s ChatGPT and xAI’s Grok to its GenAI.mil portal.
The two tools reach 3 million defense personnel. Anthropic’s Claude is locked out in a court fight over the safety limits it wanted.
1.7 million users signed on in nine months
GenAI.mil is the Department of War’s one-stop shop for commercial AI.
Staff can deploy frontier models without passing sensitive data through consumer apps, and military versions omit the data collection built into most consumer products.
Both tools received accreditation for Controlled Unclassified Information at Impact Level 5, the tier that covers sensitive government work that is not classified. That clearance allows them to deal with it at scale.
It launched about nine months ago with Google’s Gemini as its first model. The department says it has enrolled more than 1.7 million users out of a workforce of more than 3 million.
Source: U.S. Department of War.
With the addition of Grok and ChatGPT, the portal becomes a multi-model marketplace. The Pentagon framed that as a way to avoid getting locked into one supplier.
ChatGPT Mil is the result of a government program at OpenAI and an enterprise partnership forged by the two sides in 2025. It starts with chat, files, projects, and custom GPTs for document-heavy unclassified work across planning, policy, logistics, and administration.
The department points to Starshield AI. The secure-satellite unit of SpaceX runs on Starlink technology.
Its release lists deep-thinking inference, three modes of reasoning called Auto, Fast, and Expert, and reusable “playbooks” meant to capture institutional knowledge. Applications range from acquisition market research to supply chain work for logisticians.
The department said Grok enables forces to “execute missions faster and with greater precision.” The two releases are different, with the Grok one describing the execution of the mission and the ChatGPT one describing the desk work.
Judge Rita Lin saw a First Amendment problem
Although Anthropic won its federal court ruling to lift the national security blacklist, its Claude models remain excluded from the GenAI.mil portal. A parallel challenge is proceeding in the D.C. Circuit Court.
Late in February, Defense Secretary Pete Hegseth called Anthropic a Supply-Chain Risk to National Security, Cryptopolitan reported. This followed President Donald Trump ordering federal agencies to stop using the company’s technology.
Anthropic, which inked a $200 million Pentagon deal in July, wanted written assurances that its models would not be used in fully autonomous weaponry or mass domestic surveillance of Americans.
The Pentagon pushed for unrestricted access for all legal purposes and set a deadline of February 27 that was not met.
Anthropic called the label “unprecedented,” saying it’s normally reserved for foreign adversaries and vowed to fight it.
The company pointed out its history as the first frontier lab to operate models on classified U.S. networks and to support warfighters since June 2024. It pointed to a federal procurement law, saying Hegseth lacks the authority to prevent contractors from using Claude for non-defense work.
Although U.S. District Judge Rita Lin expressed First Amendment concerns during an early March hearing, she issued a final 59-page ruling on August 27, 2026, which officially struck down the Pentagon’s supply-chain risk designation.
Lin has not issued a final ruling. She said the government responded harshly after Anthropic went public with the dispute and said the crackdown “doesn’t seem to be really tailored to a stated national security concern.”
Anthropic’s lawyer, Michael Mongan, told the court that the designation had never been applied to a U.S. company before. The company is asking the court to lift the supply chain risk.
Defense staff currently gain access to OpenAI and xAI tools, whereas Anthropic remains locked out of the portal pending the outcome of the D.C. circuit court appeal.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Washington opened a two-day ministerial meeting in Chapel Hill, North Carolina on Tuesday by asking the world’s largest economies to hit the brakes on writing fresh AI rules or creating new bodies to police the technology.
The conclusions agreed upon in this G20 cycle is expected to fuel the debates in the leaders’ summit in Miami this December.
The Carolina Principles proposal
The framework from the US has a local name, termed the Carolina Principles. Michael Kratsios, who directs the White House Office of Science and Technology Policy co-hosted the gathering alongside Commerce Secretary Howard Lutnick and put the proposal to assembled officials.
Reuters reported that signatory nations would commit to “reserve new regulation for novel considerations” when they draft AI rules, ensure money goes into foundational research, and also further commercial openings for new technology.
Kratsios claimed restraint was simply the smart thing to do. “Policymakers do not need to approach each innovation in isolation and should not treat every emerging technology as a first-of-a-kind policy problem,” he told the group, according to Reuters. A White House official stated that member countries should also decline to create any new organization to oversee how AI develops.
Guest list at G20 summit
Ministers from Japan, Germany, France, India and South Korea all attended the summit. The industry’s biggest names were also present. Elon Musk, chief executive of Tesla and SpaceX, joined Tuesday’s session by video and used it to attack the European Union’s tech rules, saying the bloc’s policy “inhibits progress” and pressed leaders outside China to build new energy supplies for more data centers.
Meta’s appearance will be through CEO Mark Zuckerberg using video chat, standing in for company president Dina Powell McCormick who was slated to speak previously. Google DeepMind co-founder Demis Hassabis was added late to the video lineup, per a Google spokesperson. OpenAI’s Sam Altman and Nvidia’s Jensen Huang are both set to show up in person on Wednesday for sit-down talks with Lutnick.
Chapel Hill, home to the University of North Carolina, is one stop on a circuit of ministerial meetings happening in US cities ahead of December’s leaders’ summit in Miami.
Timing of proposal against the message
No federal AI law has gotten past the US Congress, and US states that have attempted to fill the void have felt pressure from the White House. Europe is already working towards building the exact thing Washington is working against, with the EU’s AI Act already in force, and the Digital Services Act being enforced against ChatGPT and other named platforms.
The Cyber Resilience Act also took effect this month, even though Brussels has delayed some compliance deadlines and softened parts of its data-protection regime.
A United Nations panel has also cautioned that AI is advancing faster than scientists understand it or governments can react. The summit also follows OpenAI’s disclosure of a testing incident in which autonomous AI agents reached the open internet and compromised systems at Hugging Face, a repository for AI software.
Washington could still face resistance
Canada has signaled that it will argue the opposite way, telling reporters its team will push to weigh AI progress against “public trust and safety,” a stance it holds despite an active trade conflict with the United States. China’s embassy in Washington has not said whether it would attend or not, Reuters reported.
Foundationally, there seems to be a contest for AI supremacy. Chinese open-weight models, whose core components are publicly available, are closing the gap on proprietary US systems from firms such as Anthropic and OpenAI. Vivek Chilukuri, a technology and national-security fellow at the Center for a New American Security, told Reuters the trend has “increased urgency for this administration to make sure that the rest of the world stays within the American tech ecosystem and doesn’t seek alternatives.”
Since the G20 runs on a consensus, a communiqué that does not touch on AI governance and regulation would already count as a win for the US’ position.


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

Twenty-one financial institutions including Goldman Sachs, Bank of America and Citi, have announced that they will form a shared company to issue a dollar-backed stablecoin, with a market launch targeted for the first half of 2027.
The venture will start with a single US-dollar stablecoin, according to a press release from PR Newswire and published by BBVA, one of the participants. The group of financial institutions wants to issue coins in additional G7 currencies over time and has named a stablecoin euro version as its next priority after the dollar stablecoin.
The new company still remains unnamed and is expected to be formed officially in the second half of 2026. The stablecoin is meant for wholesale, institutional and retail users, with cross-border payments and digital-asset trade settlement seen as early use cases.
The launch of a bank-backed stablecoin is a new development, as it places the coin into a market so far dominated by non-bank issuers.
Ten banks in October, Twenty-one now
This is an expansion of an effort that went public last October, when ten global systemically important banks (G-SIBs, the lenders regulators treat as too big to fail) said they were studying a 1:1 reserve-backed token that would live on public blockchains. Eight of those original ten are still in: Banco Santander, Bank of America, Citi, Deutsche Bank, Goldman Sachs, MUFG Bank, TD Bank and UBS.
Thirteen new names have now joined for the second phase of planning, with the roster now spanning five different regions. North America brings Capital One, Fidelity Investments, PNC Financial Services, Scotiabank, Wells Fargo and WisdomTree in addition to the earlier members. Europe adds BBVA, Commerzbank, Crédit Agricole, Lloyds Banking Group and Rabobank. East Asia is represented by MUFG Bank, Middle East by Abu Dhabi’s Sirius International Holding, and Standard Bank joins from Africa.
The rest of the group is made up of two US asset managers and the Abu Dhabi conglomerate subsidiary.
Why two rival stablecoin issuers joined in
The most interesting additions to this group of banks are Fidelity Investments and WisdomTree, with both banks already running their own stablecoins. Fidelity launched FIDD in January via a federally chartered national trust bank, while WisdomTree issues USDW under a New York trust charter.
Both banks’ decisions to back a shared token might point to a general conclusion within the industry that no single firm’s token or stablecoin is likely to acquire the reach and network needed to compete with the current non-bank issuers.
The press release from the group claims the initiative wants to comply with both the US GENIUS Act and the EU’s MiCA guidelines, as applicable, implying the stablecoin is being designed with two of the world’s stricter stablecoin regimes in mind.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Tron founder Justin Sun used X on September 1 to challenge how Bloomberg and Forbes value his crypto fortune.
His response comes just days after his split from Chinese actress Jing Tian became a public fight over a bride price and a $50 million demand.
How much is Justin Sun really worth?
In a series of posts published on X today, Justin Sun, the founder of Tron, argued that his personal balance sheet is nobody’s business but also nothing to hide. How much he holds “has never been” the industry’s concern, he wrote, and he added that he would not use the figure to answer his critics.
Sun then took aim at Bloomberg and Forbes, stating that they count his crypto holdings at only 20% to 30% of their value when tallying his net worth, because those assets are concentrated, volatile, and by the outlets’ standards, not “certain.”
Sun said he understands that logic, but it is the outlet’s methodology, and not his problem.
He wrote in a separate post that for 14 years, he has made the deliberate choice of keeping the bulk of his own assets in crypto. Forbes currently pegs the 36-year-old at $8.5 billion.
Why is Justin Sun suing Jing Tian?
Prior to the wealth posts, Sun published a lengthy account titled “My Girlfriend Jing Tian” on August 27. The post reportedly took him more than ten hours to write and was labeled “fictional.”
In it, Sun said the couple had moved toward marriage. He said he transferred a 30 million yuan bride price, about $4.5 million, to Jing’s parents after proposing in January, and that the two had agreed to have a child through surrogacy. Sun alleged that shortly before Jing was due to undergo an egg retrieval, she demanded a further $50 million.
At that stage, he turned to Anthropic’s Claude chatbot for advice and, acting on its recommendation, refused the money and ended the relationship. Sun is reportedly now suing the actress and her parents to recover the bride price.
Which other drama is Justin Sun involved in?
38-year-old Jing Tian, who is known for her roles in “From Vegas to Macau” and “Police Story 2013” and is followed by more than 26 million people on Weibo, rejected Sun’s account via the social media platform on August 28, insisting that she has not and will never sell her love.
Jing added that she still believed in the law and would let time settle the matter.
The episode blew up on Chinese social media and took over search results. Hashtags regarding the incident even trended on Weibo.
Some users have mocked the fact that a chatbot played a role in ending a romance, while others have been criticizing Sun for airing medical, financial and personal details.
In an interview with Sing Tao Headline, Sun also shared his doubt about the AI’s judgment, and like Jing, said he would leave the resolution to his lawyers and the courts.
Meanwhile, Cryptopolitan has reported that World Liberty Financial, the Trump family’s DeFi project, filed a defamation countersuit against Sun in May, alleging that he was involved in a coordinated smear campaign and short-selling after he sued the company for freezing his tokens.
Sun has called the WLFI suit a “meritless PR stunt,” but he also remains locked in a years-long fight with Hong Kong custodian First Digital Trust over TrueUSD reserves.
If you're reading this, you’re already ahead. Stay there with our newsletter.

OpenAI and Apple’s (NASDAQ: AAPL) legal battle in federal court escalated to a new level after the iPhone maker presented forensic data that OpenAI hardware had already started to run on trade secrets that its former employees were never authorized to share.
The “shocking evidence” pulled from a former engineer’s work laptop comes ahead of an October 1 court showdown and throws a wrench in Sam Altman and OpenAI’s push to enter the physical devices business.
Does Apple have evidence that OpenAI stole trade secrets?
Apple lawyers have told the court that they found dozens of confidential files, including a circuit schematic for a power converter on the work-issued MacBook that belongs to Chang Liu, a senior system electrical engineer who moved on to OpenAI in January.
Apple said that Liu did not just access those files, but he actually used it in simulations at his OpenAI role, according to the supplemental brief its lawyers filed in the Northern District of California.
Beyond the schematics, Apple lawyers also said they found evidence that Liu used a tool with a name identical to one of his former employer’s internal engineering programs. The lawyers warned that any AI agent that “learn[s]” can use that material to potentially create harm that is hard to reverse and keeps spreading.
Liu’s lawyers voluntarily submitted the old laptop for inspection earlier in August. Apple lawyers, on the other hand, have called the court’s attention to an alleged June attempt to wipe evidence as rationale to fast-track evidence gathering.
Their claim is that once Liu caught wind of the internal Apple investigation, he reached out to fellow OpenAI alum, Yu-Ting Peng to discuss a plan to “restore” the devices their former employer gave them.
By Apple’s interpretation of the exchange, restoring in this context meant wiping the machines of any evidence that could implicate them if the laptops were forensically analyzed.
Apple claims that more than 400 people have left the firm for OpenAI as of last count.
OpenAI sticks to negligence defense
OpenAI has not shifted its position. Since an August 3 blog post titled “Apple is getting this wrong,” the company has pinned the access on Apple’s own sloppiness, describing what Liu had as “residual access” left over because Apple often fails to shut off accounts when staff depart.
In that post, OpenAI said former Apple colleagues were the ones who messaged Liu asking for help finding files, not the reverse.
Apple tells a different story about how the access survived. It claims Liu kept his way in by exploiting “a rare, previously unknown authentication bug,” per TechCrunch, rather than through any oversight on Apple’s end.
What will happen on October 1 in the Apple vs. OpenAI case?
Apple has big stakes riding on how Judge Edward Davila interprets rules on two requests it made of the court, with both decisions due on October 1.
One is a preliminary injunction to make OpenAI pause any work on hardware using the iPhone maker’s technology while the case is still unsettled.
The second request is to expedite discovery so that Apple lawyers can depose witnesses and pull documents on a faster timeline.
The timing of the documents Apple wants to see adds a third name to the proceedings, Tang Tan. The request for material dated August 1, 2023, or later lines up with the six months before Tan, a design executive who spent more than 24 years at Apple, left for OpenAI.
Another name in the blast radius of this lawsuit is Jony Ive’s io Products.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.