Cryptopolitan

Over the next two years, Google will invest €13 billion in Finland on data centers, clean energy, and community programs. They will operate on nuclear power.
The company said Wednesday that it signed a 22-year agreement to help prolong the life of the Loviisa nuclear power plant, part of what it called its largest single investment anywhere in Europe.
The money for Finland comes as governments across the continent scramble to stop AI infrastructure moving to the United States.
Google doubles its previous German pledge in Finland
Hamina is the center of gravity where Google, 15 years ago, transmuted a defunct paper mill into a data center. The new spending is intended to keep pace with demand for Search, Maps, and Gemini, Google Global Infrastructure VP Bikash Koley wrote in the announcement.
Google said its Hamina operations drew on more than 600 Finnish suppliers for construction, operations, and fiber work from 2023 to 2025.
The new pledge overshadows the size of Google’s recent European moves. In November 2025, Cryptopolitan reported that Google pledged ~€5.5 billion, about $6.4 billion, to expand its German data centers in Dietzenbach and Hanau.
The figure for Finland is more than double that.
€31 million flows to Hamina, Kajaani, Muhos, and Vaala
In addition to the Loviisa deal, Google said it will add onshore wind capacity and enter into a contract for a 94-megawatt battery system that it said would serve to smooth prices during “cold, windless periods.”
Nuclear power offers a dependable baseload that wind and solar cannot, and that is a growing concern to operators of power-hungry AI facilities.
The project is expected to create more than 37,000 jobs across the country and bring €3.6 billion a year to Finnish GDP during construction in 2027 and 2028, according to Google.
Once the facilities are up and running, thousands of permanent roles will follow, from engineers to security and catering staff, the company said.
Google apportioned €31 million over four years to the communities of Hamina, Kajaani, Muhos, and Vaala. The budget covers AI upskilling for 4,400+ workers through local colleges.
It also funds training slots for 100 students aimed at future data center careers. The company said it will finance the regeneration of native forests and wetlands near its sites and construct recreational trails, public saunas, and fishing piers.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Suno replaced its entire AI music engine with a new model family, Suno v6, which has been trained on licensed material from Warner Music Group, BMG, and distributor Believe. The change is a direct response to the copyright suits that have shaken the startup over the past months, and shows that record labels are interested in building products with AI companies.
The launch includes three tiers, with the flagship v6 positioned as the most precise and polished option. The v6-wild model is designed as a more experimental counterpart aimed at less predictable results, while v6-mini is the faster model available on the free tier. Only v6-mini will be available to free users, while the other two models will be limited to Suno’s Pro and Premier subscribers, TechCrunch reported.
Everything before v6 has been decommissioned, and the company stated in its announcement that all previous models will be retired as Suno will move “entirely onto the v6 generation.”
Suno claims v6 can edit a selected section of a track using a text prompt and replace a single lyric without regenerating the entire song. The new model can also create mashups from multiple sources, isolate instruments for sampling, and use text, audio, images, or video as references.
Legal settlement required new models
Warner’s November 2025 settlement with Suno required a specific commitment to ship more advanced licensed models and retire the previous ones. The models that replaced the contested versions were co-developed with the same label that had taken Suno to court.
Chief product officer Jack Brody said the release was focused on remuneration for rights holders. “From the day it launches on September 9, this partnership starts generating revenue for our partners,” he stated, adding that the models open the door to remix and artist-participation features that would create “additional revenue streams” for participating artists.
Artists working with Believe can opt into the new Suno experiences, while Suno-made songs can become eligible for distribution through Believe and TuneCore, giving AI-generated tracks a path to conventional music platforms.
Suno discloses little about data
For a launch built on the line between licensed and unlicensed training of AI models, Suno said almost nothing about the data itself. When asked how the models were built, Brody refused to give any specifics, stating that “the v6 family has been trained from the ground up with a new set of data that differs from the data that went into our previous models,” including “licensed data from our partners and user data.”
Artist payments are also vague, and Suno says musicians can opt in and be paid when they take part; however, there are no definite payment structures in place yet.
Suno is valued at around $5.4 billion, and the company has raised more than $819 million.
Legal battles are far from over
Settling with Warner and BMG has not completely cleared Suno’s legal slate, with the company still facing suits from Sony and Universal Music Group, and artists including Jason Isbell. There are also still some suits from users who allege the AI model neglected security while chasing profit.
Europe seems to be on a totally different track, as a German court, in a case brought by the collecting society GEMA, ruled this year that Suno breached copyright terms. This is the very first such decision on the continent. Rival Udio, meanwhile, is also being sued by Sony over 30,000 songs, with $4.5 billion sought in payments.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Consensys Software Inc. announced on Wednesday, September 9, 2026 it will split into two separate companies. Its consumer wallet venture has morphed into MetaMask, while its Ethereum protocols and institutional software will become a new company but will keep the name Consensys.
The move is of particular importance to those who use MetaMask to hold crypto and to the banks that sit on ConsenSys infrastructure, as both companies will have separate leadership, balance sheet, and strategy going into the New Year.
Who’s in charge of MetaMask now?
The present company structure will now operate as MetaMask, and co-founder Joe Lubin will serve as chairman and chief executive. It will run the consumer wallet and consumer products.
The institutional aspect of the company, which included the Linea network, the Besu and Teku clients, will become the new Consensys. Mike Kriak will lead as CEO, while David Cunningham serves as president, and Lubin steps into the seat of executive chairman.
The companies are to run as two separate organizations, with the separation to be finalized by year’s end. Joe Lubin said the move is a reaction to a simple observation: the value of the consumer side of the business grew faster than the other parts of the firm.
A wallet that became a bank
MetaMask has seen over 100 million downloads from ~190 countries and has gone on to process trillions of dollars in cumulative transaction volume. This serves as the foundation for the “Open Money” platform, as Lubin calls it. It provides a single place to hold, spend, save, and grow money with crypto and traditional assets.
Money Account is the major sign of that shift. Launched in June, Money Account allows users to earn up to 4% APY on the mUSD stablecoin while using MetaMask to spend the same balance. The MetaMask Card is a Mastercard-supported debit product for trading through perpetual futures and prediction markets.
“MetaMask grew out of that work into the world’s most widely used self-custodial wallet, and today it’s becoming something larger: a platform where people don’t just hold their assets, but manage their money in its many diverse forms and aspects,” Lubin stated.
Banking on institutional tokenization
The second company is focused on banks and asset managers. Consensys claimed financial firms have pivoted from pilot projects to production use of tokenization, stablecoins, and blockchain settlement. Institutions like Citi, DTC, and BNY Mellon are currently running on its Besu infrastructure.
Lubin’s statement referenced a Citi report from June called “Tokenization 2030.” The report sees the market for tokenized assets between the $5.5 trillion and $8.2 trillion range by 2030.
The IPO and token questions left unanswered
The restructuring comes at a strange time for the firm’s corporate arc. Consensys was founded over a decade ago as a Brooklyn Ethereum incubator and moved to Texas in 2023.
The company sought to go public earlier this year but postponed it till fall 2026, when crypto markets took a nosedive, and Bitcoin went below the $80,000 mark. The company last had a valuation in early 2022, after a $450 million Series D raise. It was valued at $7 billion.
Lubin refused to state a future date for the IPO, even though the new company MetaMask could become public by Q1 2027. He also remained mute on the rumor of a MetaMask token, saying that the business and regulatory climate has caused firms to think twice about launching their own cryptocurrencies.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

In roughly 30 minutes after Hunter Biden’s meme coin, LAPTOP, went live on Base, over 95% of its value got wiped out, according to blockchain analytics firm Arkham.
Those who are still holding the token are now left with a fraction of what they paid.
$LAPTOP was a bloodbath!
80% of traders lost money:
• 2 lost $100K-$1M
• 100 lost $10K+
• 700 lost $1K+
• 11,000 smaller losses
Thank you Hunter Biden! 🤡 pic.twitter.com/kncqLeiZ4B
— Bubblemaps (@bubblemaps) September 9, 2026
A peak that lasted two minutes
LAPTOP went live at 8:02 p.m. and set its high two minutes later. At that point, Arkham put the token’s fully diluted valuation (FDV) at $144 billion, which is an eye-watering figure sitting on top of a liquidity pool holding just $48,000.
Thirty minutes after its first post, Arkham reported the FDV had fallen to $5 billion.
The token’s intraday high was $401.12, while the lowest it traded the same day is $1.20 the same day, per CoinMarketCap data.
LAPTOP token price. Source: CoinMarketCap.
LAPTOP currently trades around $1.5, down by over 95%. It is trading across two on-chain venues, an official Aerodrome pool holding $83,000 in USDC and a side Uniswap pool with $380,000.
Market maker cashing out on-chain
Part of the selling pressure was visible in real time. Trading firm Wintermute reportedly took in 2.5 million LAPTOP from the token’s Gnosis Safe. 2.09 million of it was moved across Wintermute’s own wallets, then the tokens were routed toward centralized-exchange deposit addresses.
Around 500,000 LAPTOP had already gone for about $2 million across 356 transactions, according to Onchain Lens.
For a token with a 1 billion total supply and 350 million circulating, that kind of flow from an insider-supplied allocation into sell orders is the mechanism a crash is made of.
The trader who turned $200K into $3K
Lookonchain spotlighted one buyer who pulled $250,000 out of Binance ahead of the launch specifically to chase LAPTOP. The buyer reportedly spent $200,000 on 919 tokens at a price of $218.
As of the time that Lookonchain published its post, that buyer’s stack was around $3,000.
Lookonchain wrote “Don’t FOMO,” a warning that arrived too late for anyone already in.
The warnings were on the record
Before the launch and subsequent crash, there have been voices that called out the setup, and one of them, Bubblemaps, has been a consistent voice in calling out projects like LAPTOP.
Bubblemaps posted what it called the distribution of LAPTOP “just BEFORE launch.” It called on the readers to check if they noticed anything.
In a subsequent post, Bubblemaps posted on X, writing that the token’s top holders are mostly fresh wallets. It wrote, “60% have no prior activity.” It also clarified what it meant by “fresh wallets,” stating that it means that the “wallet was funded in the last 10 days.”
The analytics platform also added that most of those wallets were funded the same day of the launch.
Analyst Benjamin Cowen greeted the collapse with a two-word verdict laden with sarcasm: “I, for one, am shocked.”
Hunter Biden, the son of former US president Joe Biden, is part of the founding team behind the token, which reserved 20% of supply for airdrops, in part to wallets that lost money on TRUMP.
Two in three TRUMP buyers, which accounts for around 988,905 wallets, were down a combined $3.81 billion by the end of June.
This is not the first time that a token linked to a popular or political figure will swing this way. The TRUMP token, which was launched in January 2025, lost around 55% of its value within minutes of the MELANIA token announcement.
There is also the LIBRA token that was linked to Argentine President Javier Milei, which crashed by over 94% within hours after launch.
Hunter Biden had spent the day before the launch defending the project. “Clear, strong, seven years sober, and with a lot to say,” he wrote on X on Tuesday, confirming the Wall Street Journal’s report that he was behind a meme coin called LAPTOP.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

By Kate L., Commercial Director of NOWPayments
If you run an e-commerce business, your next conversion win may come not from another campaign, but from giving ready-to-buy customers a payment method they can actually use.
As Commercial Director of NOWPayments, I work where crypto adoption meets business reality. One problem appears repeatedly: merchants can have demand and still lose the sale at payment. For this article, I reviewed checkout research alongside six months of NOWPayments e-commerce data to examine where conversion leaks, how crypto expands addressable demand, and why implementation determines the result.
This distinction matters because product demand and payment conversion are not the same thing. A customer can want the product, accept the price, and still leave because the available payment methods do not work for them.
This article is part of the Crypto That Works for Business series, a collection of materials about the commercial results behind crypto payments. We begin with e-commerce conversion. Future editions will explore other areas where better payment infrastructure can unlock revenue, reduce costs, and remove friction from growth.
The Conversion Leak Hiding in Checkout
Payment problems account for a meaningful share of abandoned checkouts.
When I went through Baymard Institute’s checkout research, the headline number was striking: the average cart-abandonment rate is 70.22%. Some of that is unavoidable. 42% of respondents said they were simply browsing or were not ready to buy.
The reasons that remain are more useful for merchants. After excluding those shoppers, Baymard found that 19% left because they did not trust the site with their card information, 17% found the checkout too complicated, 10% had their card declined, and 9% could not find enough payment options. These customers did not necessarily reject the product. The payment experience stopped them from completing the order.
Adding payment choices is not about decorating a checkout with more logos. It is about allowing more qualified customers to complete the purchase.
More Relevant Payment Methods Mean More Addressable Demand
Crypto can provide another such path – particularly for global customers, crypto-native audiences, privacy-conscious buyers, and users who cannot or do not want to rely on cards.
A large-scale payment-method study found that offering at least one relevant option beyond cards produced an average 7.4% increase in conversion and a 12% increase in revenue. A separate crypto-checkout case reported a 22% increase in checkout conversion after the payment experience was streamlined.
NOWPayments’ data provides another view. Over the last six months, 63.81% of payments created by e-commerce partners reached Finished status. Another 19.95% expired, meaning the invoice was created but not paid within the available window.
These figures begin later than conventional cart-abandonment metrics and are not directly comparable. Both, however, show that purchase intent can still be lost at payment.
Crypto Converts, but Implementation Decides How Well
Adding crypto creates a conversion opportunity, but the implementation determines how much of that opportunity a merchant captures.
Across two e-commerce implementations in the same six-month dataset, Finished rates ranged from 40.50% to 66.38% – a gap of 25.88 percentage points.
I would not present this as a ranking. The businesses serve different audiences and operate different customer journeys, and the data does not isolate one feature as the cause. The useful finding is the size of the variation.
In my experience, merchants get better results when they answer four questions before launch:
Are the right assets and networks available?
Are the amount, network, and payment window clear?
How quickly is the payment detected and confirmed?
Can the gateway handle payment errors and trigger fulfillment automatically?
A crypto button creates another route to payment. The quality of the integration shapes how well that route converts.
Coverage Has to Match Actual Demand
Crypto is not one payment method. Customers choose an asset and a network based on what they hold, which wallet they use, and the fees they expect to pay.
In NOWPayments’ e-commerce data, BTC accounted for 16.11% of successful payments, ETH for 3.88%, and USDC for 1.63%. Together, the three assets represented 21.62%.
In one partner sample, they represented only 9.67%: BTC accounted for 4.55%, ETH for 3.17%, and USDC for 1.95%. A currency mix that works for one merchant may therefore miss most of another merchant’s demand.
Demand also changes. A separate NOWPayments stablecoin analysis found that USDT accounted for 66.92% of stablecoin transaction volume in the first half of 2026. Meanwhile, USDC transaction count increased 209.02% year over year, while its volume rose 101.63%.
Merchants need broad underlying coverage without overwhelming customers at checkout.
CoinsBee reported greater payment flexibility and stronger conversion after expanding asset availability through NOWPayments. Genghis supports more than 300 cryptocurrencies across more than 98 networks and uses signed webhooks to connect payment confirmation with automatic delivery.
The longest currency list does not automatically win. The gateway must fit the merchant’s audience, product, and purchasing journey.
The Checkout Should Answer Before Support Does
More options help only when customers understand what to do next.
A crypto checkout should answer four questions immediately:
Which asset and network should I use?
How much should I send?
How much time do I have to complete the payment?
What happens after the payment is confirmed?
Unclear instructions and hidden timers create exit points. Delays after confirmation create another: the customer has paid, but the order still appears incomplete.
In my experience, a strong crypto checkout should feel uneventful. It presents clear instructions, detects the transaction quickly, and connects confirmation with the order flow.
Measure the Whole Payment Funnel
Crypto checkout should be measured as a conversion channel, not simply enabled as a technical feature.
The same five questions should guide the strategy from beginning to end:
Find the leak: Where do customers leave – at payment selection, invoice creation, or payment completion?
Offer relevant methods: Do the available payment options reflect real customer demand?
Evaluate the implementation: How many created invoices reach Finished status, and how quickly?
Know the audience: Which assets and networks produce completed payments for this specific merchant?
Remove friction: Are the instructions clear, and does a confirmed payment move the order forward?
These metrics locate the leak. If customers select crypto but do not create an invoice, the entry point may be unclear. Frequent expirations may point to instructions, asset availability, network selection, or timing. Completed payments that do not trigger orders indicate a fulfillment problem.
The business case is no longer limited to whether an online store should accept crypto. The real question is whether the store can make crypto one of its most effective payment options.
Global brands have already moved crypto checkout beyond the experimental stage. Tesla accepts Dogecoin for eligible products in its online store, while Balenciaga has introduced crypto payments through its US e-commerce site. Within the NOWPayments ecosystem, CoinsBee and Genghis demonstrate how the right integration converts that demand into sales. Merchants that delay may be losing customers they have already paid to acquire.
To me, mass adoption begins when crypto produces measurable commercial results. NOWPayments’ data shows the opportunity: 63.81% of created e-commerce payments reached Finished status. The 40.50% to 66.38% range between implementations shows why choosing and configuring the right gateway matters.
Businesses can explore NOWPayments’ e-commerce payment infrastructure to evaluate crypto as part of their checkout funnel.
Next: Where Else Is Payment Friction Costing You Revenue?
Checkout is only one place where money gets stuck. Similar friction can appear wherever a business accepts, moves, or pays out funds.
If this article finds its audience, another edition of Crypto That Works for Business will examine another overlooked point in the payment journey – and what businesses can change to keep more revenue moving.
Where is payment infrastructure quietly costing your business money?
Methodology note: The 22% headline figure is a provider-reported result following a crypto-checkout redesign, not a guaranteed outcome. NOWPayments’ e-commerce data covers the last six months. “Finished” and “Expired” are invoice-level statuses and are not directly comparable with cart-abandonment metrics. Stablecoin figures come from a separate platform-wide analysis covering the first half of 2026. E-commerce asset figures cover BTC, ETH, and USDC only. Partner-level variation does not establish causation by any single checkout feature.

Interior Minister Ulan Niyazbekov told a state blockchain council on September 5 that Kyrgyzstan’s police have opened 90 criminal cases involving cryptocurrency since the start of 2026.
Kyrgyzstan has drawn repeated Western sanctions over allegations that its banks and crypto platforms help Russia evade financial restrictions.
How are virtual assets showing up in crime?
Kyrgyzstan’s interior minister Ulan Niyazbekov disclosed at the third meeting of the National Council for the Development of Virtual Assets and Blockchain Technologies, held in the lakeside town of Cholpon-Ata, that the country’s police have opened 90 criminal cases that involve cryptocurrencies since 2026 began.
Niyazbekov said that virtual assets are being folded into the country’s criminal infrastructure like cyber fraud, embezzlement, money laundering, the concealment and movement of illicit funds, and the recruitment of “droppers,” or frontmen who lend their identities and accounts to launder proceeds.
The minister explained that tracing cryptocurrencies is especially difficult because transactions cross borders in seconds and pass through chains of wallets, mixers and other tools built to hide the trail.
He referred to a 2025 operation against a transnational online-fraud group, where investigators found a TRON-network wallet holding more than $3 million in digital assets on the suspects’ devices.
On June 5, the Interior Ministry and the National Bank signed a cooperation agreement to fight fraud across the financial system. Niyazbekov said that the agreement already lets banks freeze suspicious transactions quickly once a cybercrime alert comes in.
The ministry is also building a dedicated module inside the Unified Register of Crimes with the Prosecutor General’s Office, so that police, prosecutors and the central bank can share data on offenses committed in the digital environment.
Separately, it has proposed amending the Code of Criminal Procedure to add a “parallel financial investigation,” under which officers would trace criminal proceeds and their recipients at the same time as they pursue the people who committed the crime.
Kyrgyzstan’s authorities have a broader national framework against cybercrime in the works. Meanwhile, ministry staff have run more than 300 digital-literacy events around the country this year.
Why is Kyrgyzstan expanding into the cryptocurrency sector?
President Sadyr Japarov, who chaired the September 5 council, has told the interior ministry that he wants Kyrgyzstan to become a regional hub for virtual assets.
Government figures show that the value of transactions handled by licensed Kyrgyz crypto firms has jumped more than 500-fold, from roughly $60 million in 2022 to about $30 billion in 2025. The country now answers to the nickname “Cryptostan.” Chainalysis ranked it 19th worldwide for crypto adoption last year.
The state has so far issued two stablecoins, the gold-backed USDKG and the som-pegged KGST, and in August, KGST was among the three most-traded tokens for new Binance users in the country.
President Japarov has instructed the National Bank to build and test a digital platform by December 31, 2026, with real-world testing to follow in 2027.
Binance founder Changpeng Zhao, who now functions as a public adviser to Japarov, said the country had gone from having no crypto rules a year ago to a working framework with banks and exchanges already operating.
However, Cryptopolitan reported that Kyrgyzstan’s legal changes were made “amid reports of new EU sanctions.”
The European Union’s 21st sanctions package, adopted July 23, imposed a transaction ban on a Kyrgyz bank linked to Russia’s SPFS messaging system and on crypto platforms based partly in Kyrgyzstan and, for the first time created a mechanism for a full third-country ban on crypto services.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Anthropic alignment science lead Evan Hubinger says artificial intelligence has a greater than 10% chance of wiping out humanity within the next decade.
His estimate came Tuesday, hours after researcher Jacob Coxon said he had resigned from Anthropic and accused Anthropic and OpenAI of taking unacceptable risks as they chase stronger systems. Jacob wrote on X, “They are racing straight to self-improving superintelligence and gambling with our lives.”
According to Jacob, those who develop the technology see a risk of this happening even before 2030. He has pointed out that future AI can surpass human hackers very quickly and transform industries and have access to finance and resources.
Anthropic staff put a number on extinction risk as AI labs push toward self-improving systems
Evan replied that Jacob’s warning was correct and said Anthropic lacks a way to keep superintelligence aligned with human goals.
“Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to,” Evan wrote.
Evan later said the danger from today’s models is “low.” His concern is recursive self-improvement, where a system repeatedly upgrades its own abilities with little human help. That capability does not exist yet, although AI labs are working toward it. He said superintelligence emerging through that process is moving faster than expected.
Anthropic made a similar warning in June. It said “full recursive self-improvement also might increase the risks of humans losing control over AI systems.” The company added, “If systems are capable of fully building their own successors, the ways we secure them, monitor them, and shape their behavior all grow much more important.”
Anthropic and OpenAI are raising money while moving toward expected public listings. Neither company immediately responded when CNBC requested comment.
Elon Musk, CEO of Tesla (NASDAQ: TSLA) and SpaceX (SPCX), has warned for years that AI could threaten humanity. Researchers and academics have raised similar concerns about companies losing control.
Those fears rose again after an OpenAI model broke into Hugging Face, an open-source developer platform, in July. Jacob called incidents like that “warning shots” and said they make agreements between U.S. labs more realistic. That made him more hopeful about U.S. coordination, but he still expects a worldwide AI race to be hard to stop.
“I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities,” Jacob said.
Claude’s coding gains change how Anthropic tests software and checks its own engineers’ work
Anthropic said Claude completed 76% of its hardest open-ended jobs in May 2026, up 50 percentage points in six months.
One example began when a routine upgrade caused tens of thousands of training jobs to crash. An engineer gave Claude written context and access to the live computing cluster.
Claude checked running jobs, tested one environment setting at a time and traced the failure to an obscure debugging flag. It recreated the issue, proved the cause and confirmed a fix in about two hours. Anthropic said the same work would usually take a human engineer two to three days.
The company also measures whether Claude can write code that another engineer can understand and build on. Staff do not fully agree on the comparison. Many believed Claude’s code was below human work at Anthropic in late 2025. The company now says the quality is roughly equal and expects Claude to move ahead within a year.
That progress has changed Anthropic’s internal review process. Proposed code changes are now checked by an automated Claude reviewer before they can merge. It scans for bugs, security weaknesses and other defects.
A retrospective test found that reviewing every past change this way would have caught about one-third of the bugs behind earlier Claude incidents before production. Anthropic said the original code came from engineers it described as among the world’s best at building these systems, yet Claude found mistakes they missed.
Anthropic put the comparison this way: “Claude-written code was somewhat worse than human-written code at Anthropic in late 2025, is roughly at parity today, and we expect it to be strictly better within the year.”
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

The Pons launchpad on Robinhood Chain is minting fresh blockchain millionaires, as an Arkham-monitored wallet has turned a $2,600 July investment on the launchpad’s token into more than $1.2 million in paper profits as of this September report.
The epic ROI reported by the blockchain-tracking firm highlights a new cohort of winners from the speculative frenzy around Robinhood Chain since its hot summer launch.
How a trader turned $2,600 in PONS into $1.2 million
The trader that Arkham identified as @heylittlechef on X bought in on PONS across four purchases on launch day, back when the token’s entire market cap was around $160,700.
As of this report, @heylittlechef has cashed in on $77,400 and is still sitting on a PONS position worth around $1.08 million.
@heylittlechef’s portfolio. Source: Arkham.
Other traders have also taken big gains on PONS, including a wallet labeled 0x194 that turned $44,300 into around $811,000 in late July. That trader entered the trade at a $3.7 million market cap.
Interest has spread past retail, too. On-chain data shows market maker Wintermute has built a position worth about $3 million in PONS across multiple transaction tranches.
PONS has grown exponentially since launch, changing hands at $0.78 as of this report, with a market capitalization near $552 million. The token hit its $0.97 all-time high price on September 5.
The launchpad quietly minting fees on Robinhood’s chain
PONS is the token of the similarly named launchpad that allows users to create and trade fixed-supply tokens that graduate into deeper liquidity pools on Robinhood Chain. The Robinhood Chain itself is riding a hot hand after traders repurposed it as a de facto meme-trading base rather than the tokenized-stocks network it was presented as when it launched on July 1.
There is real substance behind the Pons run. The launchpad pulled in $90.93 million in fees and $16.83 million in protocol revenue over the last 30 days. Lifetime fees have already crossed $118 million and DEX volume has passed $1.5 billion, per DeFiLlama.
The Pons launchpad is earning more fees than Circle, Pump.fun, Polymarket and Hyperliquid. Source: DeFiLlama.
That activity has also helped Robinhood Chain’s baseline stats, helping the network set a $6 million single-day fee record on Friday, September 4.
What the project’s founder is telling holders
Much of the momentum ties back to Pons’ deflationary design, and its founder has been narrating the burns in real time.
Ozzy, who posts as @MEADGod, wrote on Tuesday that more than $208 million of PONS had been destroyed at current valuations, with circulating supply down to 699 million and shrinking “every 15 minutes.”
Roughly 80% of the revenue Pons generates is funneled into buying back and burning the token, DeFiLlama’s methodology notes.
Ozzy has also pushed back on a circulating complaint. In a Wednesday post, he said Pons does not allow taxes to be changed after a token launches, calling that claim “simply wrong,” and blamed some trading terminals for routing orders through high-tax pools rather than the ones Pons sets up.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

OpenAI’s partnership with Samsung Electronics on next-generation semiconductors continues to deepen, and the South Korean manufacturer has now agreed to produce a custom AI chip for the ChatGPT maker, OpenAI Korea confirmed at a Seoul press conference on Wednesday. The move is expected to draw Samsung deeper into a chip race that has so far been dominated by Nvidia and TSMC.
Harrison Kim, General Manager of OpenAI Korea, told reporters that chips were the area where the two firms had achieved the most in their partnership. “One of the areas where we have made the most progress and gained the most recognition with Samsung Electronics is our joint production and research on the next-generation chips we are developing,” Kim said, according to Reuters.
The collaboration runs across multiple fronts, including joint chip research, actual chip manufacturing, and a more expansive use of OpenAI’s enterprise software inside Samsung. OpenAI has not disclosed the chip’s name or its exact purpose, although reports believe it is likely to be an inference processor and a successor to Jalapeno.
Where the chip fits after Jalapeno
Jalapeno was OpenAI’s first custom chip, produced by Taiwan’s TSMC and unveiled in June. Developed with Broadcom, the chip is designed to handle inference workloads. Having Samsung lead the next design would give OpenAI a second manufacturing partner, reducing its reliance on a single supplier.
Memory seems to be the other half of the deal, with both companies planning to work on advanced memory chips to feed heavier AI workloads, an effort that could be connected to OpenAI’s Stargate data center project.
Stargate already involves Samsung’s domestic rival SK Hynix, and both Korean firms already signed letters of intent to supply memory for it last year.
Samsung’s memory push and ChatGPT footprint
Samsung has started shipping samples of its HBM4E high-bandwidth memory, becoming the first supplier to distribute that generation of AI memory. High-bandwidth memory sits inside AI accelerators such as Nvidia’s Rubin and Google’s Ironwood Tensor Processing Unit, and Samsung counts top chipmakers like AMD, Nvidia, and Google as part of its customers.
Kim also called Samsung one of the largest ChatGPT deployments anywhere, with staff in Korea and abroad using the AI model across research, marketing, and sales. OpenAI said ChatGPT Enterprise users across South Korean businesses and institutions increased 28-fold by late August, compared to the same month in 2025.
Late last year, no country outside the United States had more paying ChatGPT subscribers than South Korea, according to the company. In June, Samsung’s Device eXperience division cleared employees to use ChatGPT alongside Google’s Gemini and Anthropic’s Claude.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Brentwood, TN – (September 9, 2026) — Canary Capital Group LLC (“Canary Capital”), a digital asset investment management firm, today announced the launch of the Canary Staked TRX ETF (Ticker: TRXS). The Fund seeks to provide exposure to the spot price of TRX, the native utility token of the TRON blockchain network. In addition, the Fund also seeks to earn additional TRX through participation in the TRON network’s delegated proof-of-stake validation process, with net staking rewards reflected in the Fund’s net asset value.
“The Canary Staked TRX ETF brings investors exposure to one of the world’s largest blockchain settlement networks through a registered exchange-traded structure, while also enabling investors to benefit from potential staking rewards,” said Steven McClurg, CEO of Canary Capital. “As stablecoin adoption continues to grow globally, TRON has become a critical piece of the infrastructure powering digital asset payments and settlement. We believe investors are increasingly looking beyond digital assets themselves and toward the networks driving real-world blockchain adoption.”
TRON has emerged as one of the leading blockchain networks for stablecoin activity, supporting more than $94 billion in circulating Tether (USDT). The chain also processes the highest USDT transfer volume of any blockchain, totaling approximately $5.6 trillion year-to-date. Known for its speed, scalability, and low transaction costs, the TRON network serves as critical infrastructure for decentralized finance, global payments, and blockchain-based applications. The TRON network is governed by TRON DAO, the community-governed Decentralized Autonomous Organization (DAO) dedicated to accelerating the decentralization of the internet through blockchain technology and decentralized applications (dApps).
“The launch of the Canary Staked TRX ETF demonstrates the growing recognition of the TRON network as critical infrastructure for the global digital economy and provides institutional investors with a new way to access a network that is already powering real-world financial activity at scale,” said Justin Sun, Founder of TRON. “We appreciate Canary Capital’s leadership in bringing TRX to the ETF market and welcome innovations that broaden investor participation in the TRON network while advancing the integration of blockchain infrastructure into traditional financial markets.”
With the launch of the Canary Staked TRX ETF (TRXS), Canary Capital continues its mission to make digital asset investing simple, secure, and accessible while expanding investor access beyond Bitcoin and Ethereum into the next generation of blockchain infrastructure.
For more information on TRXS, click here.
Media Contacts
Canary Capital
media@canaryetfs.com
TRON
press@tron.network
About Canary Capital
Canary Capital is an investment management firm that blends rigorous risk management, strategic foresight, and innovative thinking to deliver private placement strategies, crypto hedge fund solutions, treasury management solutions, and publicly traded funds, with a focus on enterprise technology.
About TRON DAO
TRON DAO is a community-governed DAO dedicated to accelerating the decentralization of the internet via blockchain technology and dApps.
Founded in September 2017, the TRON blockchain has experienced significant growth since its MainNet launch in May 2018. Today, TRON hosts the largest circulating supply of USD Tether (USDT) stablecoin, which currently exceeds $94 billion. As of September 2026, the TRON blockchain has recorded over 403 million in total user accounts, more than 15 billion in total transactions, and over $28 billion in total value locked (TVL), based on TRONSCAN. Recognized as the global settlement layer for stablecoin transactions and everyday purchases with proven success, TRON is “Moving Trillions, Empowering Billions.”
TRONNetwork | TRONDAO | X | YouTube | Telegram | Discord | Reddit | GitHub | Medium | Forum

UK officials have started to question how much influence US President Donald Trump had in Anthropic not opening its newest Claude Mythos 5.1 model to Britain’s AI safety testing agency, the AI Security Institute (AISI), before it released it to the public.
A Cabinet Office spokesperson did not go so far as to confirm whether Mythos 5.1 had been withheld, telling the BBC that Britain “continues to collaborate closely with industry partners, including Anthropic, to make models safer.”
British officials are now probing whether this Anthropic launch was an exception or just the first of its kind in a changing relationship reflecting Washington’s tightening grip on advanced AI, per the Financial Times.
Do UK regulators test AI models before launch?
The AI Security Institute (AISI) is the British agency charged with assessing national security risks in frontier AI systems.
Notably, an unsettling precedent was set when the UK agency never had pre-release access to Mythos 5.1, according to the FT, which is the first time such has happened. Only agencies inside the United States got to see the model before public launch.
As Cryptopolitan reported last week, Anthropic released Mythos 5.1 as a narrow tool with looser safeguards than Fable 5.1, for approved professionals in cybersecurity and life sciences.
Why UK officials suspect Trump meddling
The concern in British quarters follows a pattern of Washington progressively ramping up its control of how advanced American AI is used abroad, rather than a single decision in this case.
Under President Donald Trump’s administration, the US has introduced export controls to temporarily limit foreign access to Anthropic’s most advanced models, including foreign nationals employed by US AI companies.
Just last month, OpenAI also agreed to limit some new systems to “trusted partners” at the government’s request.
Notably, the Trump White House, via an official, has already told CNBC that the administration does not formally sign off on products released by private-sector firms. They also stressed that the companies were responsible for the timing and scope of the voluntary testing and meetings with government experts.
What does the UK AISI do?
Set up as the AI Safety Institute in November 2023 and renamed the AI Security Institute in February 2025, the AISI runs on a £66 million annual budget and employs over 100 technical specialists from inside the Department for Science, Innovation and Technology. The institute has evaluated more than 30 frontier models, per the Ada Lovelace Institute.
However, the AISI cannot force a company to submit a model, cannot set the terms of access, and cannot stop a model reaching the market, as it is not actually a regulator. It can only deliver on its mandate of informing the British government of the models it tests via voluntary participation by model creators.
AISI found a sharp jump in Anthropic’s Claude Mythos model’s ability to run cyberattacks when it tested it in Spring 2026. It could not make any such evaluations this time around with Mythos 5.1.
The headache will have expanded into a migraine as of this week, after three Anthropic researchers publicly warned that advanced AI could threaten humanity. Alignment lead Evan Hubinger wrote on X that he sees a greater than 10% chance AI “could kill all humans” within the decade, adding that the company does “not yet have a plan to solve alignment for superintelligence.”
There is also recent history between Anthropic and AISI on cyber risk. In an August test, agents built on Anthropic’s Mythos 5 and OpenAI’s GPT-5.6 Sol went rogue, with the Mythos agent trying to slip malicious code into an open-source GitHub project and inventing fake identities to pressure the maintainer.
Anthropic paused external cyber testing on July 23 and resumed it on September 1 after adding new controls, Cryptopolitan reported.
The smartest crypto minds already read our newsletter. Want in? Join them.

BitMart appointed restructuring firm Alvarez & Marsal as its financial adviser and has announced that it will hand oversight to an independent third party.
Meanwhile, millions of exchange users remain unable to withdraw funds.
What is Alvarez & Marsal doing for BitMart?
BitMart has clarified that the restructuring firm Alvarez & Marsal (A&M) will work alongside its lawyers to review the company’s finances, sort through stakeholder questions and study how the exchange can resume user withdrawals in an orderly way.
The firm’s job, per the notice, is to weigh the options and help shape a near-term plan of action.
In the announcement, BitMart said all information on withdrawals, assets and next steps has to be approved by an independent adviser before they go public, which is why it hired A&M in the first place.
The exchange also named White & Case as its restructuring legal counsel in an earlier notice.
BitMart’s immediate actions will include building a dedicated feedback website as a single channel for users to submit views on the action plan. The company has also promised to publish the link within five working days.
Within three weeks, it said, it will roll out the feedback mechanism and the specifics of its plan.
The second commitment is that it will soon appoint an independent third party to supervise how the company runs and how it holds assets during this stretch in order to protect user interests.
Is BitMart still going to shut down?
The adviser appointment is the update BitMart said it would deliver no later than September 9, after it said in an earlier notice that it was considering a possible restructuring instead of the full shutdown it announced on July 26.
BitMart’s restructuring would include a phased, partial resumption of business paired with asset distributions to creditors.
Cryptopolitan reported that the earlier announcement that the company would wind down operations was due to operating conditions, the wider market and its strategic direction. The company stopped new account creation, deposits and fresh orders the same day, causing its BMX token to fall by about 60% within a day.
Prior to the restructuring talks, trading was supposed to end August 26, with full closure planned for January 31, 2027.
There is also some uncertainty about BitMart’s solvency. Cryptopolitan, citing CoinMarketCap, reported BitMart self-reporting roughly $5.36 million in reserves, most of it in BMX, against daily trading volume near $272.6 million.
The exchange promised a proof-of-reserves report back on May 23, but has yet to publish one. Founder Sheldon Xia said on August 8 the exchange had not run off with funds, but offered no figures.
BitMart’s gradual process is playing out against outside pressure. On September 3, Cryptopolitan reported that Echo Base, a firm specializing in distressed situations, had formed an ad hoc committee of affected customers and retained two law firms, Young Conaway Stargatt & Taylor and Ashbury Legal.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Sterling has made it through Britain’s latest political disaster and global tension without falling apart, but the pound is losing steam as the U.K. falls behind the change toward higher rates.
So far this year, the currency is up roughly 1.6% against the euro, 2.8% against the Swiss franc, 4.9% against the Swedish krona, and 1% against the Canadian dollar.
Against the U.S. dollar, it has barely moved, while the Japanese yen is ahead by 1.3%. Sterling is still beating several major currencies this year, even though that run is starting to cool right now.
Political events in Britain have not led to a meaningful sell-off in the pound. The resignation of Prime Minister Keir Starmer on July 20 paved way for the UK’s seventh prime minister in the span of 10 years. Labour acted quickly and got Andy Burnham to Downing Street.
Markets are interested in Andy’s fiscal plans, particularly whether the new government will continue to adhere to the “fiscal rules” as defined by Finance Minister Rachel Reeves.
The cost of borrowing for the British government has increased since Andy came into power, but bond yields have risen even in several other economies.
Oil and rising bond yields put fresh pressure on the sterling as markets shift toward higher rates
The UK’s economy grew 0.4% in the second quarter after expanding 0.6% in the first three months of the year. That meant that the U.K. was one of the fastest-growing developed countries during that period. Improved weather conditions, along with World Cup expenditure, boosted consumer demand. Companies were also operating amid all the geopolitics.
The pound got another boost when the Iran conflict started in April. Traders thought the Bank of England could respond more forcefully if higher energy costs pushed inflation up. That view supported U.K. rates and gave the currency extra support.
Now that picture is shifting as oil gets more expensive and yields rise elsewhere. U.S. stock futures were softer early Wednesday after Tuesday’s losses. Dow Jones Industrial Average futures fell 0.24% by 5:47 a.m. ET. S&P 500 futures were up less than 0.1%, while Nasdaq 100 futures dropped 0.14%.
Europe was mostly red too. The Stoxx 600 fell 0.69%, while Britain’s FTSE 100 dropped 0.32%. Germany’s DAX was down 0.68%, France’s CAC 40 lost 0.95%, and Italy’s FTSE MIB slid 1.27%.
Asia was more mixed. Japan’s Nikkei 225 closed 0.19% lower, while South Korea’s Kospi jumped 1.40%. Australia’s S&P/ASX 200 fell 0.11%, while mainland China’s CSI 300 finished 0.30% higher.
Oil drove much of the pressure. Brent crude futures rose more than 2% and moved above $100 a barrel for the first time since July. Tensions between the U.S. and Iran are getting worse, raising worries that energy supplies from the Middle East could face more disruption.
Crude had already climbed during Tuesday’s session and pulled U.S. stocks lower. Markets were shut Monday for Labor Day, so Tuesday was the first session of the shortened week. The Dow fell 1.2%, its worst day in almost three weeks. The S&P 500 lost 0.6%, while the Nasdaq Composite ended 0.3% lower.
Bonds added more pressure. The 10-year U.S. Treasury yield briefly moved above 4.8% on Tuesday as rising oil prices added to inflation worries. Higher yields then hit stocks again and added to the broader rate pressure now weighing on Sterling.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

OpenAI claimed that it has cracked the Navier-Stokes existence and smoothness question, one of the seven Millennium Prize Problems declared by the Scientific Advisory Board of the Clay Mathematics Institute of Cambridge, Massachusetts (CMI). However, the harder problem to solve might be convincing a public gallery that has grown more skeptical of the giant AI lab’s methods.
OpenAI walked straight from solving one of the hardest unsolved mathematics problems of our lifetime to fielding questions from two mathematicians asking whether their unpublished work somehow helped the lab arrive at its answer.
However, as this drama plays out publicly, OpenAI is a few legs behind in the perception lane, as expected when a trillion-dollar AI lab implicated in a rogue agent security hack and an internal “safety reckoning” over risk handling goes up against two relatable academics.
Which math problem did OpenAI just solve?
The Sam Altman-led AI lab said it has written proof and a machine-checkable version in the Lean proof language that a smooth, initially calm fluid can, under a smooth force, speed up without limit and break down in finite time.
The work, which OpenAI claims was done by an internal model more capable than its recently released GPT-6 Astra model, solves one of seven million-dollar problems the Clay Mathematics Institute listed in 2000.
OpenAI said it spent more on compute bill than the $1 million prize that the CMI is offering, according to head of research Mark Chen. Sebastien Bubeck, a researcher at the lab, said about 10,000 agents were working on the problem at one point.
The Clay Institute has not recognized the OpenAI proof yet. The institute requires solutions be published on a qualifying outlet, be available to the public for two years, and receive broad acceptance from mathematicians before consideration for the prize.
Notably, OpenAI has already said it will not seek the prize money.
Did OpenAI use the work of mathematicians in their chat log?
The main debate about OpenAI’s work is not even about the math. The real drama is around whether the information that Tristan Buckmaster of NYU and Levent Alpöge, a mathematician employed by Anthropic, fed into AI models, including OpenAI’s Codex, was siphoned.
Tristan directly asked the lab’s researchers whether his and Levent’s Codex sessions were being used when he first heard that the AI lab had directed major compute capacity at cracking the Navier-Stokes nut.
Typical of a mathematician, Tristan stopped short of making any accusations without proof, writing in a public statement that “I do not know whether our data was used.”
Tristan did not directly accuse OpenAI of stealing his and Levent’s work. Source: NYU Courant Institute
However, the same document accused OpenAI of floating “proposals” that included removing Levent’s name from an announcement that attributed the solution to OpenAI’s internal model.
Tristan claimed OpenAI offered to cut out Levent from the Navier Stokes publication. Source: NYU Courant Institute
Bubeck has since refuted that claim, writing on X: “I never ever asked for Levent to be removed from authorship of his own work,” he wrote.
I would like to clarify a few things:
1) The screenshot is my reaching out to Levent to coordinate our releases. I hope it’s clear from the message that we came in with the best possible intentions.
2) I never ever asked for Levent to be removed from authorship of his own work… pic.twitter.com/yRli0hLNuM
— Sebastien Bubeck (@SebastienBubeck) September 8, 2026
Sam Altman also weighed in to back his team.
Unlikely but not impossible is not a strong enough denial for skeptics
OpenAI’s core defense is flat: neither its people nor its agents saw the pair’s prompts, drafts, or user data, and they only encountered the work once it went public.
Yet in its own blog post, the company added a line that Levent himself flagged: “While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models.”
“i mean props to them for straight coming clean,” Levent posted.
Tristan’s account also complicates OpenAI’s framing. He says he was initially told “very little human input” went into the result, but that during calls, it emerged that an entire team had been involved, that easier problems were tackled first, and that even the prompt shown to him had itself been generated by prompting Codex.
He credited the underlying program not to any AI but to mathematicians Diego Córdoba and Luis Martínez-Zoroa, calling the moment a “Deep Blue-Kasparov” turning point for the field.
Tristan has not filed any lawsuit. OpenAI maintains that its proof differs substantially from the pair’s and that no user data was accessed to produce it.
As more research runs through AI systems owned by companies building rival research tools, the harder question is not who owns a finished idea but what a platform is allowed to do with what gets typed into it, and how anyone could ever prove influence after the fact.
If you're reading this, you’re already ahead. Stay there with our newsletter.

What Gates Actually Wrote
On August 26, Bill Gates published an essay on Gates Notes titled “The turbulent AI era is here. The choices we make now are critical”. For a man who has been an AI optimist for the large part over the past few years, the shift in tone here is striking.
The line that stood out was this one: “AI will either be the greatest equalizer ever invented, or the worst source of injustice.” Throughout the essay, he does not claim to know which but he does reiterate that nobody is steering. “Right now, we are not preparing for it,” he wrote. “I don’t see evidence that leaders, experts, and communities are confronting the challenges adequately. There is no plan to ease the entry into the AI era.”
It’s important to note that he distinguishes his position from a general anti-AI stance. Gates is genuinely bullish on AI’s capabilities in fields like medicine, education, agriculture and science. His trepidation is on distribution and pace rather than capability.
The comparison he uses is with his own experience. Gates helped build the personal computer and the advent of this technology completely changed the way people worked over a generation. Compared to that, he believes, AI has the power to transform work and life way faster. AI performs cognitive work, spreads through infrastructure that already exists, and requires no specialised skill to operate. As he put it, “AI will take on work in law, customer service, medicine, software and manufacturing, and it will hit those industries over the course of a decade rather than a few generations.”
The Three Things He Is Worried About
Jobs comes first. He stresses that AI will have a significant impact on jobs and that White-collar roles are already being hit modestly. He also states that the biggest risk is among entry-level and mid-level roles, particularly where AI systems operate without a human checking each step. He cites a Stanford Research paper by Brynjolfsson, Chandar and Chen which compared two groups of young workers aged between 22 and 25. Those in roles that were heavily exposed to AI and those in jobs that weren’t exposed. The second group kept growing while the first ended up around 19% smaller than it should have been. The jobs are not disappearing so much as never being posted.
Malicious use comes second. Gates argues that crimes which used to require real skill and tact are becoming increasingly available to anyone. Cyberattacks, fraud and disinformation can spread rapidly and easily when a model can do the hard parts for you. He extends the same worry to engineering biological threats. He also flags the risk of models acting against human interests in ways that become hard to reverse.
Children come third, and this is the section that reads most personally. He states that talking to AI now is simply too easy. There is no risk of being misunderstood with no real effort required from the other. That comfort is the concern and when you strip out any sort of real friction, he argues that these tools can become addictive, particularly for teenagers and that relying on them stunts critical thinking and social development.
Running underneath all three is a charge aimed squarely at his own industry. Gates has said the tech sector downplays these risks in public while acknowledging them privately, driven by competitive and fundraising pressure. His line on this is direct: it is good that some AI companies propose solutions to problems their own technology created, but we should not expect them to lead the charge.
What He Wants Done About It
Three proposals, and the first two got most of the coverage.
A tax on AI tokens and robots. The logic is an asymmetry in the existing tax code. Hire a person and you pay payroll tax on their earnings. Buy a robot and you write it off as a business expense. Gates argues the system currently nudges employers toward machines, and that a targeted levy would slow that nudge slightly while funding retraining and a stronger safety net. He floated a version of this in 2017 and it was received as eccentric. He is candid that it remains a hard sell, telling Axios it would be a larger change to the tax system than any in his lifetime, at a moment when politics is more polarised than any in his lifetime.
“Human Reserved” jobs. The more unusual idea, and the one he seems least certain about. Gates borrows the logic of nature reserves: land societies could develop but choose not to, because the loss would be permanent. Applied to labour, certain roles would be deliberately kept in human hands even where AI could technically perform them. Some protections would be permanent, like a doctor delivering a terminal diagnosis. Others would be transitional, and here he offers a concrete case. You cannot tell a 55-year-old who has spent a career in construction to move into elder care and expect that to work. He has floated getting the reserved share of work up to around 40%, using shorter workdays and earlier retirement to spread what remains. He told GeekWire he worked parts of this through in conversation with Claude, which is either encouraging or slightly absurd depending on your priors.
New institutions, domestic and international. The least detailed of the three, and effectively an admission that the first two cannot happen without machinery that does not currently exist.
Gates is upfront that the Human Reserved concept has holes. Who decides what qualifies. How rules survive across borders. How you stop firms routing around them. These, he writes, will need to be worked out in public. The International Federation of Robotics has already dismissed the tax proposal as solving a problem that does not exist.
Reader Pulse: The Poll
What the Split Reveals
The useful context here is not who these newsletter readers are but what they have told us before. This is an audience that follows AI closely enough to read about it daily, and across three polls this year it has shown a consistent pattern: high literacy, low personal enthusiasm. In May, 38% said they had never used an AI agent, with another quarter unsure what one was. Later that same month, the largest group said they do not consult AI before making life decisions. Now a majority is agreeing with a warning about the technology from a man who has no obvious incentive to deliver one.
Read together, those results describe a group that watches this sector attentively and keeps it at arm’s length personally. That makes the 56% less surprising than it first appears, and arguably more meaningful.
Yes (56.05%): A decisive majority, and our read is that the specificity did the work. Gates did not warn vaguely about the future. He cited a named Stanford study, identified where the damage shows up first, and proposed policies concrete enough to argue with. Warnings from senior industry figures usually arrive as abstractions about existential risk, which are easy to nod along to and impossible to act on. This one came with a mechanism and a timeline, and readers who track this space daily would recognise the difference.
Maybe (~23.6%): Almost a quarter hedged, and the hedge reconstructs easily. You can accept the diagnosis while rejecting the treatment. Someone might agree entirely that entry-level hiring is contracting and that children spending hours with chatbots deserves attention, while thinking a token tax is unenforceable and Human Reserved is a category nobody can define. The poll offered one box for agreement, and this group used the middle option to register partial support.
No (~20.4%): One in five disagreed, and the strongest version of that position is not techno-optimism. It is scepticism about the messenger. Gates built a company that consolidated extraordinary control over how the world computes, spent years in antitrust proceedings over it, and is now proposing constraints while Microsoft holds one of the largest positions in AI. Whether that read is fair is debatable. It is coherent, and roughly a fifth of this audience appears to hold it.
So What Happens Now
Gates is not asking for a pause. He says so plainly, and his reasoning is that competition between firms and between countries makes any coordinated slowdown unworkable. His proposals go after the fallout instead.
Which puts the three ideas in an awkward spot. He admits the tax is a harder sell than anything he has attempted before. Human Reserved is closer to a sketch than a policy, and he says as much when he writes that the details will need to be worked out in public. The institutions he wants do not exist and there is no obvious process for building them.
So the proposals are not really the thing to watch. The claim underneath them is. Gates says the labour market moves faster than governments can respond, and that gets tested every quarter whether anyone is paying attention or not.
The Stanford numbers are where it shows up first. Right now the effect sits with workers aged 22 to 25 in the most exposed roles, which is narrow enough that reasonable people are reading it as a blip. Watch whether it stays there. If the same pattern turns up among 30 and 40 year olds over the next couple of years, an essay that currently reads as pessimistic starts to look like it was published early.
Our readers have already made their call on the diagnosis. The harder question, and the one Gates cannot answer either, is what anybody does about it.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

Reflection tokens are the latest hot trend in crypto, with platforms on Robinhood Chain and Solana. The recent high-profile launches are raising the question of whether they are a short-term trend on its way out or just starting to grow.
At their most straightforward, reflection tokens reward holding by airdropping additional passive earnings. The idea has been tested in the crypto space during previous cycles, with the original Reflection Token and projects like SafeMoon. In those cases, users gained more of the same token for holding.
In 2026, reflection tokens integrate several trends in the crypto space, leading to a new wave of liquidity and hype. The attraction of passive income and short-term meme gains is one of the solutions for the otherwise range-bound crypto market.
What do new reflection tokens do?
The tickers and presentation of reflection tokens are tied to memes. In this way, they resemble older projects like BONK, which also rely on the holding of a meme to distribute rewards.
Reflection tokens also appeared at the peak of the trend of fee redistribution. The potential for passive income boosted assets like HYPE and PUMP, based on their platforms’ fee generation.
New reflection token projects also built their own tools for fee generation, mainly by levying a tax on each transaction.
Reflection tokens also combine tokenized equities with their tradable memes. This means the rewards don’t come in the form of additional tokens, but in tokenized equities. The stock-based reward taps the most hyped equities, bridging over the stock-trading culture of WallStreetBets.
As a combination of rewards and trading, the reflection tokens of 2026 present a potentially more reliable reward system based on stocks.
Have reflection tokens peaked?
Despite the attempt to encourage holding, reflection tokens still had examples of rapid price drops and the risk of rug pulls. Previously, projects like Reflect Finance had built a wide user base, including influencers and whales. Despite this, Reflect Finance crashed to virtually zero after the initial hype wore off.
New reflection tokens are also volatile and have wide daily fluctuations. The token reward model still depends on having enough dedicated holders and new buyers to generate fees and use the fees to buy some of the underlying rewards.
Reflection tokens also rely on trends in crypto blue chips, stocks, and other assets to build up their reputation. One of the leading reflection tokens, ZCat, based its rewards on ZCash. As ZEC trades above the $1,200 range, ZCat is also in the spotlight. The token’s own fluctuations also depend on the ZCash narrative.
The most popular ticker is NVDA, already reflected by 5,207 meme launches as of September 9. For stock-reflecting tokens, the long-term incentives may bring additional investments. For tokens based on other crypto assets, the relationship may be riskier. Even for tokens reflecting Solana, the results may vary by thousands of percent, depending on liquidity and the presence of willing traders.
Even when based on a blue-chip asset like SOL, reflection token performance varies for each newly minted meme. | Source: Stonkfun
As of September 9, reflection tokens may be just starting out. BonkBot announced it has already included reflection tokens from StonkFun in its automated trading. Additionally, Coingecko is already clustering memes based on which equities they reflect in its new ‘Stonks’ dashboard.
Other chains also set up reflection token tools years ago and may benefit from the new wave of launches. In the best-case scenario, reflection tokens may become a new source of liquidity for RWA trading and tokenized assets. The riskiest model may repeat the losses from meme trading, while the reflected assets also lose their bullish trend.
As of September 2026, reflection tokens are still unregulated. There is no standard of stock ownership and custody, and no limits on how many memes can offer stock ownership.

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

According to the latest data from Token Terminal, Euro-denominated stablecoin supply reached $848.1 million as of September 7. Since the start of the year, its supply has grown by around 22.6%, up from $691.7 million on January 1. While the percentage growth in itself is not something to write home about, in absolute terms, euro stablecoins added roughly $156 million thus far this year.
Dollar-based stablecoins added about $159 million during the same timeframe, moving from $298.54 billion to $298.699 billion. This is the interesting part because these are two markets separated by a factor of 350 in size but still saw almost the same net new supply over the course of the last eight months.
Despite this growth, the dollar dominates the sector with a market share of 99.5% while euro stablecoins come in second with 0.3%.
Two Issuers Hold 82% of Euro Stablecoin Supply
When looking at the market cap by asset within the euro stablecoins sector, it’s clear that the split isn’t really broad. EURC holds 62.6% market share while EURCV holds 19.6%. That is 82% of all supply with two issuers. EURI sits at 4.5%, EURe at 3.9% and the remaining 22 assets make up under 6%.
EURCV is the one to keep close tabs on here. The asset is issued by SG-Forge, Société Générale’s digital asset subsidiary, which holds electronic money institution approval from the ACPR under MiCA. A fifth of the entire euro stablecoin supply is now in the hands of a licensed European bank subsidiary. Nothing like this exists on the dollar side, where Tether and Circle continue to dominate.
Ethereum Took Nearly All The Growth
In terms of the growth in market cap by chain, Ethereum is the leader of the pack, growing from $463.4 million to $588.7 million, an increase of $125 million since January 1. The chain now holds a commanding 69.4% share of the euro stablecoin space. Behind Ethereum is Solana with a 14.7% market share and went from $94.9 million to $124.9 million. Combine the two chains and that comes to $155 million, or effectively every dollar of the year’s growth.
Base went the other way, sliding from $73.9 million to $58.7 million. Gnosis picked up a few million to reach $22.3 million and BNB Chain climbed from $4.1 million to $10.4 million.
New compliant issuance is landing where institutional liquidity already sits.
Dollar Supply Has Not Moved Since January
The dollar segment opened the year near $298.5 billion and sits at $298.7 billion now, a change of 0.05%. For a market that spent 2025 expanding on the back of the GENIUS Act, eight months without net growth looks like an expansion that has run out of room rather than one taking a breather.
That reframes what the euro number actually means. Growth is happening inside a stationary market, which makes it a share shift among currencies rather than an expansion of the category.
Issuance Is Running Ahead of Demand
MiCA gave European banks and licensed EMIs a legal path to mint, and they are minting. What has not arrived is a reason to hold the tokens in size. Offshore demand for synthetic dollars is enormous and it is what built the dollar market in the first place. Europeans already hold euros, and euro pairs stay thin across DeFi lending pools and perps collateral.
Until that gap closes, the euro segment grows because issuers are supplying it, not because users are asking for it.
If you're reading this, you’re already ahead. Stay there with our newsletter.

OpenAI is expanding its advanced models into chip design, life sciences, and finance, arguing that a model with a higher upfront cost can pay off by performing a task with fewer tries.
Reuters reported on September 9 that CFO Sarah Friar made this argument at Goldman Sachs’ Communacopia + Technology Conference in San Francisco on September 7.
For businesses trying to decide between high-end and low-cost models, this alters the debate. Instead of asking which model has the lowest token price, OpenAI would like customers to consider which model offers the cheapest completed task.
Chip design as a wedge, not an EDA replacement
OpenAI is not attempting to supplant electronic design automation systems that engineers are currently employing to develop chips. Its solutions, in fact, assist in enhancing those processes by helping the engineers analyze problems, evaluate approaches, and speed up certain jobs.
According to Friar, the demand for AI developed for specific tasks is growing. As has been reported by Reuters, OpenAI is also testing a business model where pricing depends on how effective the AI is for the given company rather than on consumption.
This aligns with the post made by Friar on July 17 that suggested the importance of “Useful Intelligence per Dollar” over token price. While a cheaper model could require several retries and reviews as well as human intervention, an expensive model may prove to be a more cost-effective option as it gets to the result quicker.
Jalapeño as the in-house proof point
The clearest demonstration of OpenAI’s ingenuity is Jalapeño, their very first custom inference chip. The company stated in a post on August 25 that its AI technology facilitated Jalapeño’s transition from the stage of conception to tapeout within only nine months.
Thanks to AI, the cycles of design, measurement, and verification were shortened. AI was also used to improve the performance of the arithmetic circuits and program the finished chip.
OpenAI claimed that the performance of the Jalapeño chip was 1.5-1.9 times higher when it came to AI work per watt in comparison to Nvidia’s GB200 and GB300 equivalent devices used for performing the calculations by GPT-OSS 120B, DeepSeek R1, and Kimi K2.5.
According to the information provided by Cryptopolitan on August 25, the Broadcom-designed chip is meant strictly for OpenAI’s purposes and not for selling at the market. These statistics are therefore only OpenAI’s numbers. SemiAnalysis observed the InferenceX runs but did not independently reproduce the full test suite.
The cost-per-task argument, checked against outside data
Friar stated that OpenAI’s decision to slash the cost of their lower-tier product Luna has contributed to an almost tenfold growth in its use, as well as increased the number of Codex’s users to 25 million, according to Reuters.
Independent testing provides some evidence for the pricing declaration. Artificial Analysis compares the pricing of GPT-5.6 Luna with Z.ai’s GLM-5.3 and establishes the price of Luna at around $0.18 versus GLM-5.3 at $2.01 for a standard task.
However, this is not exactly a one-sided comparison. GLM-5.3 scored 45 on the Intelligence Index of Artificial Analysis, which gives Luna a score of just 38. Thus, OpenAI has an advantage in pricing per standard task but is not necessarily the best-performing AI.
OpenAI Jalapeño Chip: AI Efficiency, Latency, Model Costs and $31.6T Infrastructure Outlook
A field OpenAI is entering late
AI-assisted chip design is already well established. Synopsys says its generative-AI copilot can cut information-retrieval time by 40% and reduce time-to-solution by 10 to 20 times.
Cadence said on June 1 that its autonomous ChipStack AI engineer can reduce some RTL validation cycles from five weeks to less than a day. Google’s AlphaChip has also been used in advanced chips across Alphabet.
OpenAI’s real differentiator, then, is not the claim that AI can help design chips. It is the argument that its proprietary models can do so at a lower cost per successful outcome.
Why the stakes keep climbing
The market behind that argument is enormous. PwC said on September 2 that global AI-infrastructure capital expenditure could reach $31.6 trillion through 2050, with annual spending rising from about $800 billion in 2026 to $1.8 trillion in 2050. Recurring chip and server upgrades are expected to drive much of that spending.
The OECD’s July 10 report on AI markets adds the counterpoint: quality-adjusted AI model prices fell nearly 80% between January 2024 and April 2026, even as compute, chips and other critical inputs remained highly concentrated.
OpenAI is betting that tighter integration across models, software and silicon will turn that concentration into an efficiency advantage. If it succeeds, cost per completed task could matter more than token price—and become another force pushing AI toward full-stack players.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Block, the payments company founded by Jack Dorsey, wants federal regulators to let it custody bitcoin and stablecoins through a chartered trust bank. On Tuesday, the company said it applied to the Office of the Comptroller of the Currency (OCC) to establish Builders Bank & Trust, N.A., an uninsured, non-deposit-taking national trust bank.
If approved, Builders Bank is ready to offer custodial and fiduciary services for bitcoin and stablecoins, and some of Block’s existing digital asset activities will fall under a single federal regulatory framework.
The combination of Block’s experience with digital assets and Square Financial Services positions Builders Bank well “to support Block’s broader vision of economic empowerment,” said Lee Woolley, proposed president and CEO of Builders Bank.
Why custody is the prize, not deposits
One of the main attractions of an OCC national trust bank charter is custody. Davis Wright Tremaine describes custody as “the foundation of any digital-asset business”. A national trust bank charter can bring cryptocurrency companies under the federal umbrella and provide them with a certain level of comfort in dealing with various state regulations, as well as give institutional clients the comfort of using recognized oversight.
This is key for institutions that prefer to work with regulated counterparties when putting assets on the blockchain. In some cases, a federal charter can also help in achieving qualified-custodian status.
For a long time, Anchorage Digital was the only cryptocurrency company doing business in this way. But that is no longer the case, as more companies start applying for federal charter and the focus of their competition shifts toward security, compliance, tokenization services, and institutional scale.
A charter rush measured in numbers
The OCC says it received 40 new-bank charter applications over roughly the past 18 months. Twenty-three, or 57.5%, involve some form of digital-asset activity. The Block reports that 21 of the 40 have been approved and two denied.
OCC Bank Charter Applications: 40 Filed, 23 Involving Digital Assets
According to Comptroller Jonathan Gould, the 23 digital asset initiatives mark an increase of eight times over the past four years, indicating that cryptocurrency activities are gaining stronger foothold in the sphere of regulated banking.
Despite sharing the same purpose, not all approvals are alike. The OCC conditionally approved the charter application or conversion requests for BitGo, Paxos, Fidelity Digital Assets, Ripple and Circle’s proposed First National Digital Currency Bank.
Stablecoin issuer, Coinbase, received its preliminary conditional approval for establishing a national bank on April 2, 2026. Circle, in turn, received its final approval on July 10 to set up Circle National Trust bank.
Revolut’s approval on September 2 is part of a wider phenomenon related to fintech charters, but what OCC provided was permission to do business as a full-service insured national bank, rather than as a national trust bank.
Timeline of Major OCC Crypto and Fintech Bank Charter Approvals, 2025–2026
What these charters do not grant
National trust banks are not just commercial banks that put on a crypto label. According to Block, Builders Bank will not be able to take deposits or make loans unlike a commercial bank. Trust companies are often uninsured and have to comply with the approved fiduciary and other powers.
This distinction lies at the centre of the political opposition. In a letter to Gould written on May 18, Senator Elizabeth Warren pointed out that approval is an abuse of the National Bank Act and added:
These companies are effectively crypto banks that want to evade the fundamental safeguards and obligations that come with being a bank.
Moreover, a trust charter does not guarantee access to Federal Reserve payment systems. In May, the Fed proposed a special-purpose “payment account” for organizations that meet legal requirements, but this proposal does not extend the number of institutions eligible for Federal accounts.
The stablecoin backdrop and a global nerve
The charter rush is unfolding alongside the GENIUS Act, enacted July 18, 2025. The OCC proposed implementing rules in February covering reserves, redemption, custody, risk management and issuer applications.
The implications go beyond U.S. licensing. The BIS warned in August that wider use of dollar stablecoins could fuel “digital dollarisation” and weaken monetary sovereignty in some countries. The IMF has similarly argued that tokenization could accelerate cross-border capital flows and currency substitution.
For crypto markets, more federally chartered custodians could make bitcoin, stablecoins and tokenized assets easier for institutions to hold and move under familiar supervision. But the same infrastructure could also extend the reach of dollar-based tokens abroad, turning a U.S. custody race into a much bigger contest over who controls the rails of tokenized finance.
The smartest crypto minds already read our newsletter. Want in? Join them.