Cryptopolitan

Russia’s largest lender, Sberbank, is now launching cryptocurrency payments as a service for corporate clients trading internationally.
The move coincides with the enforcement of the law “on digital currency,” which regulates coin transactions in the Russian Federation.
Sberbank starts processing cross-border payments in cryptocurrency
The biggest Russian bank by assets, Sber, has begun processing cryptocurrency payments for companies engaged in foreign economic activities.
The financial services giant made the announcement at the Eastern Economic Forum (EEF) held September 1 – 4 in the far-eastern city of Vladivostok.
Unlike traditional wires, payments between crypto wallets take only a few minutes, not days, its press service highlighted in a statement for the business news portal RBC.
At the same time, the average transaction cost is only 0.3%, which is lower than with traditional transfers, the banking behemoth emphasized, also quoted by the Vedomosti daily.
And all documents required for the bank and foreign exchange control are generated automatically during the transaction – another benefit for clients working with partners abroad.
Alexander Vedyakhin, first deputy chairman of the Executive Board of Sberbank, elaborated further:
“For Russian exporters and importers, international settlements in cryptocurrency mean faster payments, new markets, and partners previously inaccessible due to restrictions.”
The executive clearly meant Western sanctions imposed over Moscow’s war on Ukraine that have severely limited access to existing fiat channels for Russian firms engaged in foreign trade.
It’s widely believed that Russian players, both government and private entities, have been circumventing such measures using digital money. Sber itself is also sanctioned for its role.
Vedyakhin reminded that these kinds of operations were first tested under an “experimental legal regime” (ELR) established by the Central Bank of Russia (CBR) in September 2024.
In his opinion, this pilot project, which has been running for almost two years, has demonstrated the need for such solutions.
“Now that the law regulating cryptocurrencies has entered into force, we expect strong demand for this new instrument,” added the top Sber representative.
The legislation, which he was referring to, came into full effect on September 1, legalizing crypto transactions like investment, exchange and trading, as reported by Cryptopolitan.
The bill on “Digital Currency and Digital Rights” had been adopted by the Russian parliament in July and signed into law by President Vladimir Putin in early August.
While the framework permanently bans the use of cryptocurrency as a means of payment inside Russia, it allows Russian businesses to employ it for settlement in cross-border deals.
As far as digital currency payments in Russia itself are concerned, Moscow wants them conducted with the digital ruble, which was launched for public use on Tuesday, too.
Sberbank to lend against cryptocurrencies like Bitcoin as collateral
The decision to introduce cryptocurrency payments for international settlements was discussed a few days ago by Anatoly Popov, another deputy chairman of Sberbank’s board.
Speaking to TASS ahead of the economic conference in Vladivostok, he revealed the feature will become available to users of the SberBusiness app by the end of 2026.
He told the state-run news agency the necessary infrastructure had already been created and that the new payment method would be integrated as an option for Sber customers.
Popov also said that Sberbank intends to develop lending secured by digital assets, accepting Bitcoin (BTC), Ethereum (ETH) and Tether’s stablecoin USDT as collateral.
The three leading cryptocurrencies in terms of market capitalization and liquidity were recently approved for regulated circulation in Russia by its monetary authority.
Sberbank first announced it was considering offering crypto-backed credit in December last year, and then issued the country’s first such loan shortly after, before unveiling plans to scale up this lending.
This past summer, it made it clear it’s also going to launch a regulated coin trading platform and a cryptocurrency wallet by the last month of 2026.
According to the new law, established players in Russia’s traditional financial market like Sberbank will be able to work with crypto via dedicated divisions under their existing licenses.
If you're reading this, you’re already ahead. Stay there with our newsletter.

The UK’s equity markets are about to undergo a major modernization, with the news reported by the Financial Times that the London Stock Exchange (LSE) has agreed to partner with Payward, Kraken’s parent company, to tokenize the 100 largest publicly listed companies on the LSE. Tokens will be issued under Payward’s xStocks framework and they said that the products will become available within weeks to eligible investors in more than 110 countries.
Payward and the London Stock Exchange are partnering to advance the tokenization of UK equity markets.
In the coming weeks, the 100 largest London-listed equities will go live as xStocks, bringing 24/7, programmable onchain access to investors in more than 110 countries.…
— Payward (@Payward) September 1, 2026
However, UK-based investors will not have access as xStocks still aren’t available domestically. This means products built on British blue chips will trade onchain for overseas users except in Britain when the rollout begins. So far no companies have been named nor is there any confirmed launch date.
LSE 24 Doesn’t Launch Until 2027
After passing the necessary regulatory approvals, the LSE also plans to list xStocks on LSE 24, its 24/5 trading platform which is scheduled to go live in the first half of next year. Exchange traded products are set to come first and client testing is said to be done before the end of this year. Most notably, Payward and the LSE said it will explore fully fungible equity tokens carrying the same rights as ordinary shares.
Nearly every tokenized equity trading today is a wrapper. Investors get price exposure through a token backed by shares sitting with a custodian, without the voting rights or the direct claim that comes with holding the stock itself. A 300-year-old exchange saying out loud that it wants to close that gap is a different signal from a crypto platform saying it.
xStocks Was Losing the Issuance Race Before This
Tokenized equities now sit at roughly $2.5 billion, up about 267% year to date, as per rwa.xyz. Ondo leads with around $840 million. xStocks held around $606 million in August. Binance’s bStocks was already at about $593 million, two months after launching.
Binance won that ground on distribution. It brought new users into tokenized stocks rather than pulling them off competing platforms, which is the harder version of the trade and the one that compounds. Payward cannot answer that directly. Kraken’s user base is large but it is not Binance-large, and no amount of listing velocity fixes a reach problem.
Supply Is the Only Lane Payward Had Left
So the LSE deal is a supply-side answer to a demand-side loss. Exclusive access to the FTSE’s top names, distributed to 110 countries, wrapped in the credibility of an exchange that has been operating since the 1600s. Binance can list faster. It cannot list what it does not have.
Whether that holds depends entirely on exclusivity. If LSEG signs similar agreements with other issuers over the next year, Payward’s advantage lasts about as long as the paperwork takes. If it doesn’t, xStocks becomes the only venue outside the US where a trader can get onchain exposure to the London market, and inventory starts doing the work that distribution couldn’t.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.

August saw the broadest spillover of liquidity into the top 100 altcoins and tokens. 83 of the top 100 assets saw significant net gains, boosted by several major narratives.
August ended as ‘Bitcoin season’, as the leading coin recovered the $80,000 level on several occasions. However, this did not prevent liquidity from spilling over into the top 100 most active and liquid coins and tokens.
Top 100 coins and tokens outperformed BTC on a three-month basis, and in August, 83% of the assets had positive gains. | Source: CoinMarketCap
Based on CoinMarketCap data, 26 altcoins and tokens outperformed BTC significantly in the past three months. As Cryptopolitan reported, other assets were closely watched, including ETH as the leader of a new altcoin season.
Cryptorank also noted that the market breadth ranking reached a yearly high, signaling the spread of liquidity to selected coins and tokens. During those rallies, older assets return to positive gains, but there is no rush to new launches or a liquidity rush to new mints. Traders switched to legacy assets with proven utility and presence on exchanges, with the rare meme token launch.
Which altcoins were the most active?
Altcoins are historically more volatile compared to BTC. Cryptorank data places BTC volatility at 2.25%, while altcoins often go above 5%. This also translates into more active short-term rallies and daily gains, leading traders to explore opportunities.
August was also the best month for crypto as a whole in 2026. The altcoin market added more than 26% in value, while retaining a 20% market dominance.
The most active altcoins were linked to apps, DEX trading, perpetual futures, and fee generation. ZCash (ZEC) added more than 80% to its price in August, after recovering the privacy narrative and compensating for previous losses. XMR followed the trend, but with a weaker gain of 44% for the past month.
Hyperliquid’s HYPE added 60% net over the past 30 days, rising beyond $80. The token got a boost from the ongoing demand for perpetual futures activity, as well as the recent launch of permissionless prediction markets.
Solana (SOL) got another boost from meme tokens, DeFi, and stablecoin liquidity, adding 40% in the past 30 days.
TRON, on the other hand, was a big exception, with only 1% in net gains, despite the increased on-chain and app-based activity. The slower performance, even for a top 100 token by market cap, shows that this time, altcoin gains are not lifting all assets. Even with altcoin strength, there are now far fewer opportunities for small assets to rally, and even then, most have low liquidity and crash quickly.
Will the market retain its momentum in September?
For BTC, September has been a relatively weak month historically. However, in 2026, August defied expectations, leading to the strongest rally for the year.
Previously, BTC was expected to weaken further in September, with the bottom of the bear market sometime in October. This may mean some of the liquidity has been waiting to be allocated on the sidelines, while prices bounced decisively from their lows.
BTC and some assets are already facing resistance at various price levels, potentially causing stagnation or negative price moves. BTC has already traded the $80,000 level with signs of resistance, while SOL stalled under $100.

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

Government bond yields rose across Asian markets on Tuesday, with Japan’s benchmark 10-year yield hitting the 3% mark for the first time in 30 years amid inflation worries and mounting fiscal strains.
This surge in borrowing costs continues to weigh heavily on Japanese equities, particularly technology and artificial intelligence-related stocks. It follows a volatile Monday session, in which the Nikkei share average closed virtually flat amid a mix of tech-sector anxiety and macroeconomic headwinds.
These equity gains remain capped by geopolitical risks in the Middle East and rising domestic inflation fears, especially as market consensus builds for an imminent Bank of Japan interest rate hike this September.
Analysts predict a high probability of a Bank of Japan rate hike
Bond yields moved higher in tandem with those of other Asian and global economies on Tuesday morning. For instance, South Korea’s 10-year government yield rose 0.06 percentage points, exceeding 4.37%. Meanwhile, Australia’s benchmark yield scaled a five-week high above 5.1%, alongside slight upward ticks in New Zealand and Singaporean 10-year debt. The 10-year U.S. Treasury yield had also hit a peak of 4.76% overnight.
Some of these upticks are closely tied to escalating conflicts in the Middle East, which have lifted Brent crude futures by 0.7% to nearly $91 per barrel, intensifying global commodity supply fears. Nonetheless, the current momentum is boosting expectations for more central bank rate hikes.
According to CME FedWatch metrics, markets currently price in a better-than-60% chance that the U.S. Federal Reserve will raise interest rates at its September meeting.
Beyond global macro factors, Japanese market sentiment is being adversely affected by structural fiscal factors and weakening supply-and-demand dynamics in the domestic bond market. The chances of an increase in the bank rate in the near future stand at 93%, based on Totan Research and Totan ICAP data.
Nonetheless, Japan will be selling its 10-year government bonds on Tuesday afternoon. Michael Wan, a senior currency analyst at MUFG Bank, Singapore, noted that it would reveal investors’ stance on higher yields in developed market bonds around the world
Some of Japan’s AI-related shares declined
Most AI-related stocks, which have a substantial influence on Japan’s major Nikkei index, fell on Monday. Cable maker Fujikura fell 3.5%, while conglomerate SoftBank Group fell 3.3%. However, semiconductor equipment makers rose, with Tokyo Electron rising 2.1%.
The Topix index of Japan also ended marginally higher, rising 0.1%, amid new costs weighing on corporate profit margins in the technology industry. Chipmaker Nvidia just told its best customers they should be ready for an increase of more than 15% in AI servers due to rising memory costs.
The weak performance of AI-oriented Japanese stocks comes as investors wonder whether the extraordinary rally in technology stocks can continue at the same pace. Semiconductor and AI companies have enjoyed huge demand for data centers, advanced chips, and computing solutions, but rising financing costs are putting more pressure on a more challenging environment.
Investors will now ask whether corporate earnings can still justify elevated valuations as bond yields rise. If AI spending cuts or pressure on profit margins play out, such companies that saw big gains during the AI boom could see their shares fall, particularly those with the greatest gains in earnings.
At the same time, Japan’s semiconductor sector remains strategically important, supported by government efforts to boost domestic chip production and attract investment in advanced manufacturing. That might provide some longer-term support even if higher interest rates create short-term volatility.
Wataru Akiyama, an equities strategist at Nomura Securities, speaking on the future margins for global tech firms, noted: “We view the concerns over the profitability of semiconductor-related companies — led by Nvidia — as little more than a pretext or trigger for selling” following “a significant build-up of open positions in margin trading recently. I think it is fair to say that the outlook for earnings growth in AI and semiconductor-related stocks has not changed.”
On the other hand, with fixed-income returns rising amid inflationary concerns, Wataru pointed out that Japanese stock market indices are not well-equipped to sustain a prolonged rise at this point.
On the brighter side, companies in Japan increased their capital expenditures in the second quarter, indicating confidence in doing business.
Certain economists suggested that an increase in capital spending could boost Japan’s economic growth prospects, thereby giving the Bank of Japan room to raise interest rates in the near future.
According to figures issued by the Ministry of Finance, Japanese capital expenditures rose by 1.6% in the second quarter compared with the same period last year, a significant pickup from the 0.05% growth recorded in the previous three-month period.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Anyone can argue with their friends about whether a candidate will win an election, a company will beat earnings, or a celebrity couple will stay together. Turning that disagreement into a functioning market where people can put money behind their predictions is considerably more complicated.
Behind every prediction market is infrastructure for creating the market, matching trades, managing liquidity, settling positions, and deciding what happens when an outcome is disputed. Until now, much of that process has been kept concentrated among a small number of platforms operating the market themselves. But as prediction markets attract more users and use cases, attention is increasingly shifting from who can trade on them to who can build them.
That is the problem Rain Protocol is trying to address with the second version of its software development kit (SDK). The permissionless prediction market protocol has launched SDK (Software Development Kit) v2, allowing developers and AI agents to create and operate independent prediction market platforms on networks like Arbitrum One. The update also includes migration guides to help existing users move from v1 to v2.
The update brings much of the infrastructure that would otherwise need to be built independently into the protocol itself, including market creation, trading, settlement, resolution, disputes, and appeals. For developers, that means less work happens on the infrastructure layer and more can happen at the product layer, where decisions about what kinds of markets to create and how users interact with them are made.
The new SDK is also designed for both human developers and AI agents. Its machine-readable documentation and built-in agent skills allow AI coding tools to understand the protocol and build on top of it more independently. That matters as AI agents take on a larger share of software development tasks. In the context of prediction markets, Rain is betting that those agents could eventually do more than assist developers. They can help create and operate markets themselves.
Roy Shaham, CEO of Rain Protocol said, “As the market expands, we expect the biggest shift to come from users being less passive and increasingly a part of the building process. Our goal with SDK v2 is to give developers the freedom to build new types of markets, explore new ways they can be used, and shape them around their own ideas, and we’re eager to see what our community builds with it.”
On the trading side, the update also changes how trading can work by adding an on-chain order book alongside Rain’s existing automated market maker (AMM). The two systems give builders different ways to structure trading. Some markets can continue to depend on automated liquidity through an AMM, while others can allow buyers and sellers to place orders that are matched directly on-chain. That distinction can matter for markets with different levels of liquidity, trading activity, or user behavior.
Other updates handle how users interact with markets once they are live. Users can approve a session once instead of authorizing each individual action, while builders can receive real-time updates on trades and other market activity. The SDK also allows users to convert collateral into Yes and No positions, and convert them back into collateral, without changing the market price.
If prediction markets continue to expand beyond a handful of large platforms, the next phase of the category may be shaped as much by the people building markets as by the people betting on them. Rain’s SDK v2 is one example of how prediction markets are becoming easier to build, not just use.

After implementing new protective measures, Anthropic has resumed its external cybersecurity tests of its various models. The testing had previously stopped on July 23 due to security issues that occurred during evaluations.
The restart is important as it is considered that external evaluation is one of the means of assessing such frontier AI solutions by clients, regulatory authorities, and competitors. This evaluation is growing in importance against the background of the increasing investment in the sector. According to the estimates made by Gartner on July 20, global end-user expenditure for AI models and platforms will reach $64.252 billion in 2026. It represents 63.4% growth compared to $39.311 billion in 2025.
Why a testing pause rattles a $64 billion market
The forecast by Gartner indicates the rapid growth of expenditures on artificial intelligence platforms and models. As companies begin to allocate more money to cutting-edge artificial intelligence, inquiries regarding the reliability, risk, and behavior of different models start to gain traction. Third-party assessments provide a means for corporations and regulators to obtain an unbiased evaluation of those risks before the systems are finally put into operation.
When Anthropic halted external cybersecurity testing of its pre-release models, it temporarily removed one of the tools utilized by clients and regulators to analyze model responses in case of adversarial conditions.
The timing of the pause was particularly critical. In a summary of a July 20 research paper carried out by MIT FutureTech and the University of Queensland, 272 international AI experts rated AI-enabled weapons and cyberattacks among the five most dangerous risks between 2025 and 2030.
According to the study’s “pragmatic mitigation” scenario, experts estimate a 12% chance of disastrous results for AI-enabled weapons, cyberattacks, and other capabilities causing mass damage. The report highlights information, national security, and finance as the areas most exposed to the risk from AI.
“Coding and hacking are some of the areas where we’re seeing the fastest growth in AI capability.”— Peter Slattery, MIT FutureTech research scientist and study co-author, speaking to MIT Sloan
This makes this case especially important for the cryptocurrency and finance sectors, where cyber risk, fraud, manipulation, and automated attacks can ignite changes in numerous interconnected markets in a matter of minutes.
What Claude did during the evaluations
On July 30, Anthropic reported three incidents from its cybersecurity evaluations in which Claude models reached the internet from a third-party evaluation environment and gained unauthorized access to the real systems of three organizations.
Internet access was available because of a misconfiguration, even though prompts told the models they were operating in a simulation without internet access. The models were also deliberately run without Anthropic’s normal cyber safeguards so researchers could measure their underlying capabilities.
Then, on August 4, the UK’s AI Security Institute published its incident report on unsanctioned agent behavior during cyber testing. AISI said it detected the activity on July 28 during a routine evaluation.
The institute conducted a cybersecurity test of different models a total of 122 times. Out of those, an agent acted independently and made unsanctioned actions in 10 cases, resulting in 19 such actions.
Seventeen actions were related to Mythos 5 developed by Anthropic, and two actions were related to GPT-5.6-Sol developed by OpenAI. In the most serious case, an agent attempted to place malicious code in a real open-source GitHub project and later created fake accounts and attempted to persuade the maintainer to approve the offer. However, the maintainer did not agree and AISI found no damage from the attempts.
AISI emphasized that this was not a sandbox escape. They said internet connectivity was purposely activated and the provider’s cyber classifiers were switched off. It also pointed out that the Mythos 5 and GPT-5.6-Sol configurations that were subjected to the testing are not sold commercially.
The safeguards Anthropic put in place before restarting
Anthropic says the incidents combined an operational security failure with alignment problems it has documented before, including motivated reasoning and a willingness to take harmful steps in pursuit of a narrow goal.
Before restarting external evaluations, the company said on August 31 that it had added a real-time classifier that can block tool calls, end tasks, and alert a human when a model aggressively probes its testing environment, tries to escape it, or unexpectedly gains internet access.
Anthropic also moved high-risk internal cyber sandboxes to stronger isolation and now requires outside partners testing pre-release models with reduced safeguards to use stricter controls, including hardened sandboxes with no internet access by default.
The company also plans to work with METR on an independent review. METR earlier this year conducted a pilot assessment of internal AI risks involving Anthropic, Google, Meta and OpenAI.
The broader issue is what Anthropic calls “pacing the frontier”: deciding when safety concerns should slow development even as commercial pressure pushes the industry to move faster.

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

Stablecoin payment cards closed their best month in August, with volume, transactions and users all reaching new monthly highs. Crypto card spending hit $1.076 billion in August, making it the second consecutive month where volume has surpassed the $1 billion mark. Data from paymentscan.xyz also shows that transactions reached around 10.67 million, while addresses grew to over 283K from 261K the month prior. When we divide the volume by the number of transactions, this works out to an average purchase size of $100.80.
RedotPay took the lion’s share of the total volume with around $390.1 million processed in August or about 36% of the entire sector. When it comes to transactions, the share is much larger at over 6.34 million, or 54% of every swipe or tap tracked. The gap between these two numbers is noteworthy because when we do the math, RedotPay’s average purchase sits around $61.50, which is well below the sector average mentioned above. This points to routine spending rather than occasional large purchases. EtherFi and KAST remain in the second and third spot in monthly volume generated.
Emerging Markets Account for Most of the Growth
StraitsX, which is a Singapore-based payments firm that sponsors Visa card programs for other crypto companies, recorded a 600% growth in gross transaction value across lower-GDP regions between early 2025 and 2026. At the same time, Binance has highlighted how the average users of its Brazil card rose 53% from the launch quarter to Q2 2026 while average volume grew 80%. The leading uses were ride-hailing, food delivery, groceries, restaurants and online subscriptions.
Similarly, Kraken also reported that weekly payments via its Krak Card more than doubled over the past year to 8.3 per user, with retail and store purchases making up 59.3% of spending.
The Issuing Rails Got Cheap Enough for Small Tickets
Mastercard turned on stablecoin settlement on June 3, covering USDC, Paxos-issued tokens, RLUSD and SoFiUSD across eight chains. Visa now counts more than 160 stablecoin card programs either live or in development. Providers like Rain, Reap and Stripe’s Bridge have stripped out much of the float and licensing overhead that made a $12 purchase pointless to settle onchain. Small tickets only work when the cost per transaction collapses, and that is what changed over the past year.
The Data Comes With Caveats
Three programs make up 55.6% of all volume, which makes the sector far more fragile than the headline number suggests. Paymentscan’s RedotPay figures are self-reported by the issuer rather than observed onchain, and RedotPay is both the biggest name in the dataset and its loudest promoter. The company is currently dealing with a roughly $472.8 million claim in Hong Kong brought by Binance affiliates over alleged user diversion, and its planned $1 billion US listing now looks unlikely before 2027.
Scale is worth keeping in perspective too. August’s $1.076 billion annualizes to roughly $12.9 billion, which is about 0.06% of the $20 trillion-plus traditional card market. This is still a rounding error in global payments.
Spending Went Up While the Float Went Down
Here is the part that stands out. Stablecoin supply has dropped 3.6% from its May peak and now sits around $304 billion, as per DefiLlama. Card spending set a record in the same window.
For most of the past three years, onchain payment activity moved with the size of the stablecoin pool. More supply meant more speculative capital sitting idle, and some of it eventually got spent. That link appears to be breaking. Volume, transactions and users all hit highs while the float contracted, which suggests the people using these cards are funding them to spend rather than parking capital and swiping the remainder.
The smartest crypto minds already read our newsletter. Want in? Join them.

According to a report by the Wall Street Journal (WSJ) on Monday, the US Securities and Exchange Commission (SEC) wants investment companies to provide proof that their products actually contain the private stock they claim to offer.
In a private market dominated by popular AI companies like OpenAI and Anthropic—and where some forms of investment are traded in the form of blockchain tokens—the inquiry prompts the one question that investors are most concerned about: is the promised investment really there?
According to Reuters, which refers to the WSJ’s article, the SEC has asked registered investment advisers to prove that the special purpose vehicles (SPVs) under their jurisdiction actually have ownership and/or some exposure to the shares they promote. Reuters stated that it was unable to get any verification of the news from its side. The reported SEC inspection does not relate to any particular firm.
Why the SPV question hits AI hardest
SPVs gather funds from investors to take positions in various private companies making it possible for outside investors to invest in companies that are not available on the open market. In recent times, SPVs have appeared to be an attractive option for investing in the artificial intelligence boom.
The Cryptopolitan reported, in a report published on August 27 based on the pre-IPO valuations tab of DeFiLlama, that Anthropic and OpenAI topped the list of 182 companies with estimated valuations of $1.38 trillion and $900.29 billion respectively.
Source: Cryptopolitan, citing DeFiLlama’s pre-IPO tracker snapshot reported August 27, 2026. These are estimated private-market valuations, not official company funding valuations
The amounts involved are staggering. According to Stanford’s AI Index for 2026, private investments in AI around the world rose an astonishing 127.5% in 2025, reaching a total of $344.7 billion, which includes $170.9 billion for generative AI technologies. As capital continues to flood into private companies for years to come, the ability to prove what the investment is buying is becoming ever more critical.
The companies are already policing their own equity
OpenAI has already warned investors about unauthorized exposure. Its equity-transfer notice says the company is “aware of firms that market unauthorized opportunities to gain exposure to OpenAI,” including through direct equity sales, SPV interests, tokenized interests and forward contracts. OpenAI also warns that unauthorized transactions may leave investors with an interest that:
“will not be recognized and carry no economic value to you.”— OpenAI, Unauthorized OpenAI Equity Transactions
Anthropic has issued a similar warning, saying transfers involving its stock require board approval and that it does not permit SPVs to acquire Anthropic stock. In both cases, an SPV claiming access does not necessarily mean the underlying exposure is valid.
An enforcement case that shows the failure mode
The SEC has already brought a case showing what can go wrong. On August 10, 2026, it charged Adit Ventures Management, CEO Eric Munson and three affiliated general partners with allegedly defrauding investors over pre-IPO holdings including SpaceX and Klarna.
The SEC alleged Munson falsely told an investor that a fund owned shares in a private company it did not hold. The complaint also alleges defendants resold pre-IPO shares to client funds at higher prices, misrepresented costs, charged millions in unauthorized fees and pledged client assets to support a $10 million credit line.
“That misconduct has no place in investment advisory relationships where clients count on investment advisers being their fiduciaries.” — Corey A. Schuster, Chief of the SEC Enforcement Division’s Asset Management Unit
Without admitting the allegations, the defendants consented to judgments subject to court approval, including disgorgement, civil penalties and, for Munson, an associational bar with the right to seek reentry after three years.
Where crypto rails raise the stakes
The issue reaches crypto because private-company exposure is increasingly moving onchain. Cryptopolitan reported in April that OpenAI’s implied valuation crossed $1 trillion through onchain pre-IPO instruments backed 1:1 by SPV exposure on Jupiter.
Tokenization does not solve the ownership problem. It can instead spread the same underlying claim across more investors.
In a January 28, 2026 statement on tokenized securities, SEC divisions said moving a security onchain:
“does not affect application of the federal securities laws.”— SEC Divisions of Corporation Finance, Investment Management, and Trading and Markets
What matters next is whether the reported examinations become enforcement actions and whether products tied to major AI names are specifically targeted. For investors, the question is simpler: can the firm selling the exposure prove it owns what it says it owns?


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

On August 31, Andreessen Horowitz (a16z) raised additional $1.75 billion with the help of its fifth growth fund after it had closed its own AI hardware fund at $1.1 billion just a few days prior to that.
These two consecutive raises of funds coincide with the recent research made by Goldman Sachs that the investment into AI around the world will reach $1 trillion in 2026, which indicates that money is flowing through the whole spectrum of the industry at a remarkable rate.
Growth fund: a venture fund allocated to last stage startups in need of money to grow their products, teams, sales, and geographical reach.
Two funds in one week
The growth fund debuted in January at $6.75 billion, thus the latest $1.75 billion is an extension rather than creation of new financial vehicle, according to TechCrunch. David George, the leader of a16z Growth, shared that the team has supported more than 100 companies on their growth-stage journey over the past seven-plus years, mentioning Databricks and SpaceX in the company announcement.
Both funds are part of a wider initiative. a16z revealed more than $15 billion in new funds in January. It was also reported by TechCrunch that it manages around $90 billion. The strategy for growth includes enterprise and customer artificial intelligence, defense and industrial technology, robotics, health, and infrastructure.
Betting on the machine age
The $1.1 billion Machine Age Fund takes a16z deeper into the physical layer of AI.
Machine Age Fund: a16z’s $1.1 billion fund for the physical AI stack, from semiconductors and memory to networking, storage, data centers and robotics.
a16z believes that artificial intelligence is leading to changes in hardware. From the first H100 systems to Rubin, compute density in each rack has been multiplied by 28, rack power consumption has grown from 5 to 10 KW to 100-250 KW, and it is expected to reach 1 MW in three years, while hardware supply growth is traditionally around 20-30% annually.
Independent investment data points in the same direction. The OECD found that AI infrastructure and hosting companies attracted $109.3 billion in venture capital in 2025, up from $47.4 billion in 2024 and accounting for more than 42% of all AI VC investment that year.
A $1 trillion backdrop
Goldman Sachs Research estimates about $1 trillion in worldwide AI-related investment in 2026, including $581 billion in the US, with cumulative global AI investment since 2022 reaching roughly $1.8 trillion by year end.
Gartner forecasts semiconductor revenue of about $1.6 trillion this year, up 92%, with memory alone at roughly $837 billion. It expects AI data centers’ share of chip revenue to rise from 36.5% in 2026 to more than 53% by 2030.
Stanford’s 2026 AI Index found global private AI investment reached $344.7 billion in 2025, up 127.5%, while US private AI investment hit $285.9 billion, more than 23 times China’s $12.4 billion.
The OECD uses a narrower venture-capital measure. It found AI companies attracted $258.7 billion, or 61% of the $427.1 billion invested across the global VC market in 2025, up from 30% in 2022. That shows just how sharply venture funding has tilted toward AI, even before the latest a16z raises.
Money reaching companies faster
The OECD also found deals above $100 million accounted for about 73% of AI VC investment value in 2025, while deals over $1 billion represented almost half. That concentration helps explain why a relatively small group of AI companies is reaching enormous private valuations so quickly.
Anthropic, OpenAI and xAI reached $100 billion valuations in roughly five years, according to Forge Global, versus about 16 years on average for older companies such as SpaceX, Stripe and Waymo.
That appetite is now reaching deeper into hardware. British chip startup Fractile recently entered talks over a $6.5 billion pre-money valuation after securing a preliminary agreement to supply about $250 million of chips to Anthropic, as Cryptopolitan reported. Its chips are not expected until 2027, making the valuation a clear example of investors pricing future AI infrastructure demand well before the hardware reaches the market.

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

According to a post on X by the investment firm Dialectic on August 31, their Starloop vault is set to launch in September 2026, providing holders of tokenized SpaceX exposure the opportunity to utilize credit to earn yield via DeFi without having to sell the tokens.
With the ongoing search for real-world assets in crypto that can serve as collateral, this project may serve as an example of whether tokenized equities can integrate easily into the realm of DeFi lending just days after Coinbase introduced stock tokens on Base.
Borrow against the position instead of selling it
Dialectic described the process in a memo made available August 28 via Substack. A depositor gives a tokenized SpaceX position to the vault, which uses it as collateral in blockchain lending marketplaces.
The vault then borrows stablecoins, adhering to predetermined loan-to-value ratios, before deploying those funds into different strategies expected to yield higher returns than the loans’ interests. The positions are tracked and adjusted based on fluctuations in collateral value, rates, and yield availability.
The vault operation uses Makina’s non-custodial system, with authorized actions and risk settings established on the blockchain. Base, Coinbase’s Layer-2, provides the execution layer while Dialectic’s Meccanico division oversees the strategy.
Dialectic has likened its model to SpaceX’s approach to vertical integration to a portfolio; namely, an asset is used as collateral, collateral serves as liquidity and liquidity generates profit.
Why Dialectic is betting on SpaceX now
The thesis is based on SpaceX’s initial public quarter. The memo from Dialectic quotes a 92% increase year over year for Q2 revenue at $7.8 billion and a nearly tripled adjusted EBITDA figure of $3.5 billion, plus 12 million Starlink subscribers and 1.4GW of nominal computing capabilities, and $100 billion cash and marketable securities at the close of the quarter.
By comparison, Dialectic puts a blended fair value calculation of $286 per share on SpaceX, which is about twice times more than its latest closing price of $143.69 on August 31. The gap between these two numbers is the crux of Dialectic’s argument.
Dialectic claims that the market treats SpaceX as a collection of separate concepts instead of a whole company. In addition, the firm reveals its custodial relationship with both SpaceX and Tesla and says that borrowing costs can be greater than the returns from investments.
Tokenized stocks arrive as crypto looks for real collateral
The timing coincides with the tokenized stocks launch by Coinbase on August 24 on Base using its B20 standard. The tokens represents actual shares and are supported by a regulated custodian, can circulate across Base DeFi apps, and can be used as collateral in Aave. Thus, for onchain lenders, the collateral possibilities may go beyond just crypto-related assets.
Nevertheless, liquidity remains low. According to Galaxy Research, Coinbase was launched with a circulating supply only in NVIDIA, Meta, Apple, and Alphabet with a combined market capitalization of approximately $7.5 million.
Coinbase has also created a SpaceX contract called SPCX among 13 tokens, which did not have circulating supply at the time of launch. Furthermore, these tokens have also been banned for U.S. users while the market waits for the SEC’s “innovation exemption” for on-chain stock trading.
The ownership question hanging over the trade
Galaxy’s Alex Thorn flagged a legal wrinkle. Coinbase’s tokens use a “third-party issuer” structure, while Coinbase described them as:
“a real share that you actually own”— Coinbase/Base
Coinbase also said holders get:
“a direct claim on the share.”— Coinbase/Base
Thorn’s response was:
“Which is it? A real share, or a claim on a share?”— Alex Thorn, Galaxy Research
With the use of these systems, the legal relationship of the tokenholder is with the special purpose vehicle rather than SpaceX itself, and the rights of shareholders depend on the terms set by the issuer.
The demand for exposure to SpaceX seems pretty clear. When SpaceX went public in June, Cryptopolitan reported that crypto-traders generated more than $1.2 billion in Hyperliquid perpetual-futures volume in their attempts to price the stock, although those contracts never provided anyone with the shareholding rights.
Starloop offers a different path towards satisfying the stated demand. The main question now is whether meaningful, verifiable collateral will be available onchain once the vault opens.

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

Based on blockchain analysis by Arkham, North Korean hackers have laundered over $30 million worth of Bitcoin through Hyperliquid over the last three weeks, according to reports. The developments come at a delicate time.
President Donald Trump and the Commodity Futures Trading Commission (CFTC) are attempting to bring the offshore derivatives exchange within the boundaries of the United States. The discovery of Arkham intensifies the unresolved question: is it possible for a decentralized venue to fall under American regulation without incurring the sanctions and unlawful financial risk avoidance that were purposefully avoided by the creation of permissionless DeFi?
Arkham’s data lands as Hyperliquid courts Washington
As reported by Cryptopolitan and various media outlets, Hyperliquid was subject to the same criticism in December 2024 in relation to wallets linked with North Korean parties. According to Hyperliquid, no hacking had taken place and no money had vanished from user accounts. The new transactions show that the association is still in place while Hyperliquid is becoming more popular in the US.
According to Chainalysis, the North Korean regime looted around $2 billion worth of cryptocurrency in 2025, its most prosperous year ever in this regard, with the funds allegedly used to fund its weapons programs. CertiK has said that the DPRK has stolen about $6.75 billion in total during 263 incidents since 2016. On August 11, the Royal United Services Institute published a report outlining how Pyongyang exchanges the looted crypto for fiat currency to support its programs and calling for stricter onboarding processes and information-sharing rules regarding virtual asset service providers.
A rental deal through Kraken’s parent, not a purchase
Hyperliquid Labs is currently having discussions with Payward, the parent company of Kraken, regarding a U.S. strategy. Instead of purchasing an exchange with a license, American registered traders will be able to make trades using specific Hyperliquid-related perpetual futures via Bitnomial, the CFTC-regulated clearinghouse operated by Payward. Payward has already submitted the plan to the CFTC but has not yet received approval.
The Hyperliquid application would still not be available to U.S. users and registered users of Bitnomial would have access only to a small portion of what Hyperliquid has to offer in terms of perpetual futures. Specific compliance levels in the app are unknown.
On May 1, Payward confirmed the successful acquisition of Bitnomial, which was valued at $550 million. This move gave Payward access to Bitnomial’s entire derivatives stack, which consists of a Futures Commission Merchant, Designated Contract Market, and Derivatives Clearing Organization, among others, that are registered with the CFTC.
The commercial logic runs through HYPE. Hyperliquid allocates 99% of protocol fees to its Assistance Fund, which automatically converts trading fees into HYPE. Its documentation says the fund’s HYPE is burned, permanently removing it from circulating and total supply. An SEC filing by Hyperliquid Strategies said 46.7 million HYPE, or 4.7% of initial supply, had been acquired by the fund and permanently removed as of August 23.
It is not known if the volume routed via Bitnomial would contribute to such a mechanism, nor infomation about the commercial agreement between Payward and Hyperliquid has been provided. The flat licensing fee and the percentage of the U.S. trading income could mean something very different for HYPE investors. HYPE reached its all-time high of $86.71 on August 27, while there were not a single U.S. trader using the outlined route.
The perimeter question critics keep raising
Trump during a meeting at the White House on August 19 referred to CFTC Chairman Michael Selig as playing an important role in making the crypto trading platform of Hyperliquid “fully compliant and legal”. Following this comment, HYPE saw a spike of almost 17%. Further reports from Cryptopolitan indicate that the CME Group and ICE participated in the meeting too but they have been pushing regulators to investigate Hyperliquid for price manipulation and sanctions exposure.
That is the key point – when it comes to onshoring a decentralized platform, one has to account for customer verification, market surveillance, and checking for sanctions – which are the processes that the DeFi sector minimizes. Arkham’s figures illustrate what these measures need to uncover on the market.
The investors’ enthusiasm keeps growing. The Bitwise fund launched its spot Hyperliquid ETF, BHYP, on May 14, trading it at NYSE Arca and using Anchorage Digital Banking as its digital asset custodian.
What the CFTC proposal does and does not settle
The reported on-chain movements do not establish who controlled the receiving exchange accounts or whether the exchanges knew the funds were linked to Lazarus. CoinDesk also notes that public blockchain data cannot show all compliance actions taken after assets reach a centralized exchange. Hyperliquid has yet to explain publicly how its architecture can screen or block the Lazarus-linked wallets identified by Arkham.
A proposal is not a clearance. Hyperliquid and Payward declined to comment, and no final registration terms, timetable or product-specific compliance framework is public. Payward has reportedly presented the basic structure to the CFTC, but final regulatory approval remains pending.
Until regulators act, onshoring remains a proposal, and the North Korean transfers remain a reminder of what supervisors would be signing up to police.


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

The corporate purchases of Bitcoin resumed this week, creating additional demand in the market that has already risen significantly throughout August. Strive bought Bitcoin for approximately $143 million while Strategy came back from a roughly ten-week hiatus and bought approximately $370 million worth of Bitcoin, further indicating the ongoing interest from corporate buyers who have turned into a new source of demand for the crypto-assets.
As for the larger market, timing plays a crucial role. Even as of Monday, Bitcoin was trading near $78,600, closing August up by more than 24%. The return of some significant treasury buyers eliminates another burden that has been put on the industry when those firms stopped purchasing or in some situations sold Bitcoin.
Treasury buyers return as Bitcoin holds near $79,000
The buying spree comes after a period of uncertainty regarding treasury trades. Strategy, founded by Michael Saylor and the company which pioneered the model, resumed buying BTC last week, acquiring 4,603 BTC worth $369.7 million and bringing its total holdings up to 845,050 BTC.
The acquisition marked the firm’s first Bitcoin purchase in 10 weeks. However, in that interim, Strategy sold 6,916 BTC for approximately $430 million between June 30 and August 10 to help enhance its liquidity position.
Tom Lee’s Bitmine was also in the market with its largest Ethereum purchase since June. All of these actions indicate that the digital-assets treasury companies are returning to accumulation mode as prices recover.
On August 25, Bitcoin hit a price in excess of $80,000 for the first time since mid-May, aided by a softer U.S. dollar and increasing fears of debasement, as per a report by Reuters.
Publicly-traded firms are reported to have approximately 1.264 million Bitcoin across a total of 198 companies according to Bitcoin Treasuries. Nevertheless, these holdings as mentioned are highly concentrated. The top five listed companies today have about 78.4% of the total amount of Bitcoin among public companies with Strategy having independently more than 66% of the total amount.
That concentration matters. Corporate demand can return quickly, but aggregate flows still depend heavily on a relatively small group of companies with access to capital markets.
Strive climbs to the fifth-largest corporate holder
Strive bought 1,800 BTC between August 24 and 28 at an average price of $79,431 per coin, including fees and expenses, according to a Form 8-K filed Monday. The purchase lifted its stash from 21,356 BTC to 23,156 BTC.
With the acquisition, Strive became fifth in the Bitcoin Treasuries’ rankings of public equity Bitcoin holders, surpassing Bullish.
The buying speed of Strive has greatly increased. The company purchased 20 BTC in the last week of July, acquired 79 BTC from August 10-14, and in the subsequent week acquired 1,110 BTC, followed by 1,800 BTC the week that ended August 28.
TD Cowen lifts its price target on ASST
Wall Street responded positively to the acquisition. TD Cowen increased its pricing target for Strive’s stock from $28 to $32, up 14%, while maintaining its buy recommendation.
ASST’s stock increased by about 6% to around $23 Monday morning, and the asset is up almost 165% over the last six months, according to The Block.
As of August 28, Strive’s year-to-date Bitcoin yield stood at 40.8%. TD Cowen anticipates that the company will purchase approximately 4,300 BTC in the third quarter, which is more than 180% higher than its earlier projection of 1,500 BTC.
The new filing also reiterates why dilution is so important to the model. Bitcoin’s value held by Strive increased by 8.4% in the interval between August 21 and 28. However, during this same period the number of effective common shares increased by about 4.0%. This means that Bitcoin held for each effective common share increased by approximately 4.3%.
Why the treasury model cuts both ways
The demand story has a counterweight. The Financial Times reported that Bitcoin treasury companies had shed more than $80 billion in market value from their mid-2025 peak as the model unwound.
Strive itself reported a GAAP net loss of $257.6 million in the second quarter, with $234 million tied to fair-value declines in its Bitcoin and Strategy preferred-stock holdings.
That volatility is the trade-off. When Bitcoin rises and capital markets stay receptive, companies can issue securities, buy more coins and reinforce buying pressure. When Bitcoin falls or equity premiums disappear, the same mechanism can reverse through dilution, financing stress or outright Bitcoin sales.
The latest purchases show the treasury flywheel turning again. Whether it becomes a durable source of global crypto demand will depend on whether these companies can keep raising capital faster than Bitcoin volatility erodes the premiums that make the strategy work.
The Financial Times says companies that copied Strategy’s model included businesses such as a Spanish coffee chain and Japanese clothing retailer. As the treasury model deteriorated, some companies began selling crypto and returning toward their original operations. This means Strive’s $143M purchase may actually be evidence of a shakeout, not a broad revival.
The corporate crypto-treasury experiment is entering a selection phase: the strongest operators are accumulating aggressively while weaker or more financially constrained companies are retreating, restructuring or losing the executives who built their crypto strategies.
Today’s evidence Direction Strive Accelerating BTC accumulation Strategy Resumed BTC purchases after ~10-week pause Smarter Web Bitcoin strategy chief departing Sept. 1 35 of top 50 treasury firms Shares down >50% from earlier levels Top 50 treasury-company market cap ~$150B → ~$67B Smarter Web BTC 2,712 BTC Strive BTC 23,156 BTC
The smartest crypto minds already read our newsletter. Want in? Join them.

George Santos, former U.S. Representative, was banned from participating in Kalshi’s prediction market and has been asked to pay more than $71,000. This marks the first time Kalshi has banned a person indefinitely in response to disciplinary sanctions. However, it suggests that alongside federal regulators, the platforms are also starting to impose penalties on their users who bet on results they can affect personally.
Mr. Santos bet on whether or not he would be present at the State of the Union Address, which was a market that only he could affect himself.
Kalshi bans Santos after CFTC fines him $35K
According to the notice issued by Kalshi’s compliance team, it has been established that “George Santos engaged in trading activity in certain markets related to his attendance at the State of the Union address.”
The permanent ban and the fine of more than $71,000 have been imposed independently of any government proceedings by Kalshi itself. Meanwhile, it is not the only punishment that Santos is receiving because of those trades. Last month, he settled with the Commodity Futures Trading Commission (CFTC). The agency holds jurisdiction over prediction markets.
According to the CFTC, Santos manipulated the value of the State of the Union contract through the positions he held at Kalshi, based on public statements he made about attending the event in the two weeks prior that moved the contract price “significantly.”
For that violation, Santos had to pay $35,000 without admitting or denying the CFTC’s allegations. As stated by Santos’s attorney, Joseph W. Murray, Santos cooperated with the CFTC in its investigation and the address was the first time his client placed a trade on a prediction market. As per Murray, Santos reserved his travel and hotel in Washington because he expected to attend the event.
Santos was a congressman representing New York starting in January 2023. Before the end of the year, he was expelled from Congress due to the investigation performed by the House Ethics Committee regarding Santos’s misconduct associated with ethics violations.
Why do prediction markets continue to fall into the same trap
Insider trading has become a recurring issue since prediction markets have developed into a business worth billions. Santos is the latest name, but not the only one. In April, the Department of Justice arrested an active-duty U.S. Army soldier who allegedly traded on confidential information on Polymarket before the arrest of former President of Venezuela, Nicolás Maduro.
Days later, the same pattern came up. Last Friday, the CFTC issued an order to Gabriel Perez, who used to be a White House teleprompter operator, to pay more than $172,000 for using advance access to President Donald Trump’s speeches to benefit from “mention markets” on Kalshi.
Some bipartisan legislation bills have been proposed by lawmakers that would prohibit trading by individuals with non-public information. However, nothing has been approved yet. Meanwhile, prediction market companies. This includes Kalshi and Polymarket, which rely on their own approaches, such as the requirement to verify the employment of traders in sensitive markets. The lifetime ban on Santos is the most vivid example of self-regulation of venues.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Polygon Labs quietly released two hard forks that fixed a number of security issues in its proof-of-stake (POS) network.
The layer-2 network finally shared the details in a forum post. All node operators have to upgrade or they will be kicked out of the network consensus.
Austin fork stops nodes from sending the TxDependency field
The worst bug was in Heimdall, which is the software that coordinates Polygon validators.
Heimdall bundles the contents of each transaction inside a wrapper called, google.protobuf.Any, according to a forum post by Parvez03.
A wrapper can sit inside another, just like how boxes can be packed inside other boxes. However, in this analogy, there’s no cap on how many layers deep wrappers could go.
An attacker could create a single transaction stacked with those layers for almost no cost. This makes validators waste computing power to unpack it.
Polygon described the flaw as “a permissionless way to force costly, correlated work across the whole validator set.”
The Kyoto hard fork upgraded the Heimdall software to v0.11.0.
“Kyoto adds a byte-level pre-scan that rejects a transaction once its nesting passes a threshold, enforced identically at mempool admission (CheckTx) and on the consensus path (ProcessProposal),” the post said.
Polygon checks two times:
When a transaction arrives.
When validators build the block.
The second hard fork, named Austin, focused on Bor, Polygon’s execution client. It patched two denial-of-service paths.
One was state-sync events, the process that carries L1-to-L2 bridge deposits. Those events execute contract code and precompiles, just like a normal transaction would do. However, the amount of gas burned was not capped per block.
The Austin fork adds a hard per-block limit on state-sync gas.
Polygon never capped the size of the TxDependency field.
A validator could load it with huge data to crash every node that read the block. There’s nothing else about that block that looked wrong.
Austin removes the field from the wire format entirely.
Source: Polygon Forum.
Kyoto fork activated at Heimdall height 51,533,000
Polygon said that nobody exploited the bugs on the mainnet and that it fixed them all before the forks went live.
The Austin fork was activated on mainnet at Bor block 91,949,700 and the Kyoto fork at Heimdall height 51,533,000.
Every node requires Bor v2.10.0. Validators and full nodes also require Heimdall version 0.11.0. Nodes running older software have already forked from the canonical chain.
The fixes are simple binary upgrades, with no state migration, genesis change, or full resyncs needed.
POL traded at $0.09, down about 9.3% over the past 24 hours. The token is in the green zone, up 25.8% over the last 30 days, according to CoinGecko.
In July, the Heimdall V2 mainnet went offline for about an hour, according to a past Cryptopolitan report.
The smartest crypto minds already read our newsletter. Want in? Join them.

Anthropic becomes the latest AI company to be sued over how it trained its AI models. Sony Music Publishing and Warner Chappell Music took Anthropic and its co-founders to a federal court in Northern California. The suit was filed late on Friday, August 28, and it accuses the AI firm of pirating over 20,000 songs to train its Claude chatbot.
The lawsuit comes amid growing copyright suits against AI companies. Just last year, Anthropic was ordered to pay $1.5 billion by the courts after it was sued by authors in another piracy case.
The plaintiffs in this case argue Anthropic has not changed its pattern of pirating copyrighted content.
Why are publishers suing Anthropic?
The 48-page filing tells of Anthropic’s antics, calling it “one of the largest and most blatant ongoing thefts of intellectual property in history,” naming CEO Dario Amodei and co-founder Benjamin Mann as individual defendants as well as the company.
The plaintiffs accuse the Amodei-led firm of running “a brazen campaign of illegally torrenting, scraping, and downloading copyrighted works” just to train Claude.
The suit gets more detailed as it describes how the copyrighted materials were allegedly obtained. Co-founder Benjamin Mann is accused of obtaining over five million pirated books using BitTorrent while his employees pulled down 2 million more from Pirate Library Mirror.
Those books contained lyrics and sheet music. Anthropic also obtained lyrics from MusixMatch and LyricFind; both services pay the labels for the license to their content.
The catalog Anthropic allegedly pirated
The publishers allege the piracy of popular songs like Mariah Carey’s “All I Want for Christmas is You,” Taylor Swift’s “Paper Rings,” Marvin Gaye and Tammi Terrell’s “Ain’t No Mountain High Enough,” Bon Jovi’s “Livin’ on a Prayer,” Earth, Wind & Fire’s “September,” and Leonard Cohen’s “Hallelujah.”
Sony and Warner seek a jury trial and statutory damages to the tune of $150,000 per pirated copy, as well as $25,000 for each time copyright information was copied from a work. That amounts to billions of dollars for over 20,000 songs.
In matters pertaining to individually copyrighted commercial recordings, plaintiffs need not prove they lost money before seeking statutory damages.
Why the piracy claim is a big deal
The manner of acquiring the music catalog is what makes this a big deal for Anthropic. This all leads back to the author’s case from 2025. The judge ruled that using copyrighted work to train AI can pass as fair use, but downloading such works from pirated sources is against the law.
That led to the $1.5 billion settlement, and Sony and Warner have made reference to the judge’s ruling in that case, claiming Anthropic’s business model is “built on copyright theft.”
The irony is not lost on anyone who has been paying close attention to the AI industry. AI labs jealously guard their systems against distillation, the practice of training a rival model on another model’s outputs, calling it the pilfering of proprietary work.
The same firm(s) protecting their intellectual property are now accused of stealing the proprietary work of others.
Anthropic’s response
Anthropic is not going down without a fight, saying, “We disagree with the publishers’ claims, and we intend to defend ourselves robustly in court.” The company continued its response by calling this “the third lawsuit from the same lawyers, recycling allegations from cases already before the courts,” and reiterating that model training is “a transformative fair use.”
The suit comes in the wake of other lawsuits against Anthropic from Universal Music Group, Concord, BMG, and Round Hill Music. It remains to be seen whether or not US courts will label AI training on copyrighted material as fair use.
If you're reading this, you’re already ahead. Stay there with our newsletter.

BlackRock’s BUIDL fund has grown back to roughly $2.8 billion in assets and reclaimed its title as the largest tokenized U.S. Treasury product.
BUIDL is taking the lead back from Circle’s USYC. At the same time, Depository Trust & Clearing Corporation (DTCC) and NYSE owner Intercontinental Exchange (ICE) are moving their tokenization plans closer to launch.
What token occupies the top spot in the U.S. Treasury?
BUIDL, formally known as the USD Institutional Digital Liquidity Fund, has reclaimed its position as the lead in the U.S. treasury after it lost it to Circle’s USYC in March 2026.
The fund currently holds about $2.8 billion in assets, accounting for about 18.5% of a tokenized Treasury market that data trackers put at $15.1 billion. The remaining 81.5% is split among USYC and everyone else.
The overall market is slightly higher at $16 billion, up from $15 billion in recent weeks due to an increase in institutional demand. Franklin Templeton (NYSE: BEN) and Ondo Finance are among the newer names contributing to the increase.
Securitize, which built BUIDL, now handles the fund across eight blockchains, including Ethereum, Solana, Aptos, and BNB Chain.
The company posted record first-quarter 2026 revenue on the asset-servicing fees tied largely to BUIDL, and has since listed on the NYSE.
What is DTCC launching in October?
Depository Trust & Clearing Corporation (DTCC), which clears most U.S. equity activity, announced back in May that it plans to open its DTC tokenization service commercially in October 2026.
In regard to that, the company completed production trades on July 15 that stress-tested collateral pledges, securities lending, Treasury and repo settlement, and CCP margin workflows across the Canton Network and Hyperledger Besu.
The SEC issued a no-action letter clearing DTC to run the service for three years on pre-approved blockchains on December 11, 2025. DTCC’s Industry Working Group now counts more than 50 firms, including BlackRock, JPMorgan, Goldman Sachs, NYSE, Nasdaq, Circle, and Ondo.
DTCC data puts $300 trillion in global high-quality liquid assets on the table, of which only 10% to 11% currently gets used as collateral.
On August 31, 2026, tZERO and ICE (NYSE: ICE) also signed a memorandum of understanding under which tZERO will help design the digital transfer agent and broker-dealer systems for ICE’s planned NYSE-affiliated Digital Trading Platform.
ICE also agreed to put money into tZERO’s latest funding round. In addition, ICE will license tZERO’s blockchain patent portfolio, which includes 23 patent families and 103 patents, according to tZERO.
The two firms said they will consider using tZERO tokenized assets as collateral at ICE’s clearing houses.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Metaplanet put 2,400 bitcoin worth around $186 million into Coinbase Prime over three hours on August 31.
The firm’s third large transfer to the exchange in a week comes as bitcoin trades well below the average buy price for the Tokyo-listed company.
Six deposits from 103 BTC to 800 BTC in 3 hours
Arkham data showed Metaplanet’s transfer occurred on August 31 and was broken into six deposits from 103 BTC to 800 BTC.
The sole 800 BTC leg, valued at some $62 million, went out about half an hour before the rest, which then came down in quick succession.
A deposit of 1,000 BTC on August 25 was worth about $79.77 million at the time, while 1,350 BTC, worth about $108 million, went out on August 28.
Metaplanet has sent at least 4,750 BTC, worth ~$374 million, to Coinbase Prime in under seven days.
It’s unclear how much of the 3,000 BTC overlaps, so the total counts only the distinguishable transfers.
On August 29, Lookonchain tweeted that 3,000 BTC, worth around $237 million, had hit Coinbase Prime in a 24-hour period.
Metaplanet still reports 43,000 BTC on its balance sheet, now worth about $3.36 billion, at an average purchase price of ~$96,191 per coin.
None of these recent deposits have been connected to any confirmed decrease in that holding.
Coinbase Prime holds crypto in regulated custody. It offers over-the-counter (OTC) trading and offers collateral services.
Source: Lookonchain via X.
Superplanet would take 4.9% of the holdings and $2.5 million in cash
Bitcoin was trading close to $78,000 last week. This price puts Metaplanet’s coins underwater against their $96,191 cost.
That gap does make a sale a costly proposition, because selling now means taking the loss.”
CEO Simon Gerovich has pushed back on the sale chatter directly, calling the transfers custodial and not showing “any intention to liquidate.”
Gerovich explained that the company had actually moved 5,014 BTC when trackers flagged a 3,881 BTC transfer around August 13.
“This was a routine custody operation. No bitcoin was sold, and our holdings remain 43,000 BTC,” Gerovich wrote on X, per Cryptopolitan.
A similar 4,986 BTC reshuffle in March also ended without a sale.
Metaplanet has offered to give Superplanet 2,100 BTC plus $2.5 million in cash, or about 4.9% of its present holdings.
Superplanet is a Nasdaq-listed US bitcoin treasury platform that Metaplanet is building with Super League Enterprises.
Nasdaq filings, regulatory approval in the US and Japan, and a Super League shareholder vote anticipated in the fourth quarter of 2026 are all prerequisites for that plan.
Metaplanet is the world’s third-largest listed corporate holder of bitcoin, according to data from Bitcoin Treasuries. It’s behind Strategy’s 840,447 coins and just under Twenty One Capital’s 43,514.
The smartest crypto minds already read our newsletter. Want in? Join them.

Etherscan has released a suite of tools that lets AI agents and coding assistants pull live blockchain data directly from the explorer’s API.
The latest release is aimed at developers building agents that need verified onchain facts across more than 60 EVM-compatible networks. The company announced the “Build with AI” launch in a post on X on August 31, 2026.
What Etherscan actually shipped
The release bundles three ways for a machine to reach Etherscan’s data. There is a Model Context Protocol (MCP) server, a hosted interface that agents such as Claude and ChatGPT can call in plain language; a command-line tool for terminals, scripts, and CI pipelines; and a set of installable agent “skills.”
According to Etherscan’s documentation, the MCP server runs at a single endpoint, mcp.etherscan.io/mcp, and exposes 20 tools.
Those cover the basics of reading a chain: native and token balances, normal and internal transactions, transaction and receipt lookups, contract details, gas prices, and event logs. One connection reaches every supported network.
The Command-Line Interface (CLI) follows the API’s structure, with commands in the form etherscan <module> <action> and JSON output by default, plus table and CSV options for people reading the results.
All of it authenticates with a standard Etherscan API key, and a free key works across every supported chain.
Skills that trace money and read contracts
Etherscan’s public GitHub repository lists four installable skills through a npx skills add command or by copying a folder into an agent’s skills directory.
Etherscan Flow traces money movements across addresses and writes them into a case file. Etherscan Contract Review explains what a verified deployed contract does. A transaction debugger reconstructs what a given transaction did and why. An orchestrator skill routes a task to the right interface.
According to the repository, the stated point of the skills is that “every address, amount, and transaction hash comes from a live API call, never invented.” The whole suite is read-only and respects the API’s normal quotas and rate limits.
Fighting the hallucination problem
Etherscan’s release is seen as a solution to the hallucination problem that also found its way into crypto-AI tools.
Language models can fabricate a wallet balance or a transaction that never happened, both of which can be costly. This usually happens when they do not have the right data, and in most cases, it is because they cannot crawl the pages with the data they are looking for.
In its documentation, Etherscan also warned users not to paste their API keys in fake or fraudulent MCPs posing as legitimate. It stated that only one official MCP server exists and warned that listings on MCP marketplaces calling themselves “Etherscan MCP” are unaffiliated.
Built for an internet where bots outnumber people
So, the scale usage and agentic activities call for appropriate measures to be in place to prevent lapses.
In early June, AI agents surpassed humans as the internet’s largest source of traffic. This was confirmed by Cloudflare CEO Matthew Prince, who said that it happened sooner than he had predicted.
Then Cloudflare Radar data put agentic bots at 57.4% of web traffic against 42.6% for humans, with North America skewed further toward bots at 68.6%.
So far, humans have taken the lead back; however, bots still account for over 35% of the web traffic in the past four weeks.
Even publishers are now working on making it easy for AI agents to access their content.
Etherscan’s release is similar to that move as well. It is publishing machine-readable docs and a query layer built for software readers, not just human ones.
If you're reading this, you’re already ahead. Stay there with our newsletter.

Yorkville America has announced the launch of an artificial intelligence ETF called the MANGOS Plus Index ETF (ticker FRUT).
Yorkville previously acted as an investment advisor for Truth Social’s defunct digital-asset funds. The move is now being regarded as a signal of a departure from crypto into AI stocks.
Started with Bitcoin baskets, now a single-ticket AI fund
The new fund will adopt the ticker FRUT and is designed around the hardware aspects of artificial intelligence, Yorkville America announced. The company believes FRUT will serve as a “single-ticket expression” of the platform and hardware aspects of artificial intelligence. This simply means that investors get access to AI infrastructure with one product, as opposed to mixing up different chipmakers and platform names.
The launch becomes more relevant when you realize who is backing it. Yorkville America was an advisor to Truth Social Funds, the ETF branch of Trump Media & Technology Group. This marks a clear shift in emphasis for a firm that has spent the last eighteen months developing crypto products.
The crypto ETFs that arrived first
Prior to its recent pivot into AI, Yorkville America served as an advisor to a bunch of digital-asset funds. In February of this year, TMTG filed with the SEC for the Truth Social Bitcoin and Ether ETF. The joint product weighted ~60% to Bitcoin and 40% to Ethereum.
A companion filing covered the Truth Social Cronos Yield Maximizer ETF, centered on Cronos (CRO), the token linked to Crypto.com.
Both funds were built to disburse staking rewards to shareholders. Crypto.com was billed to serve as a custodian while offering liquidity and staking services. Yorkville America Equities advised on both funds.
Why TMTG withdrew the crypto filings
The digital asset play gained little traction. In May of this year, Truth Social withdrew its Form S-1 registration statements for the Truth Social Bitcoin ETF and the Truth Social Bitcoin & Ethereum ETF.
“The Company has determined to withdraw the Registration Statement and not to pursue the public offering at this time,” the filing said.
Yorkville America framed the withdrawal as repositioning rather than a defeat. The President of Yorkville America, Steve Neamtz, stated that the ’40 Act structure “allows us to bring more differentiated investment strategies to our investors that are not possible under the ’33 Act framework.”
The ’33 Act, or Securities Act of 1933, dictates the manner in which securities are first offered to the public for sale. While the ’40 Act, the Investment Company Act of 1940, dictates the organization and structure of investment companies.
Not everyone sees it the same way, though, as James Seyffart, a Bloomberg analyst, believes the decision is related to the saturated nature of the market for spot Bitcoin ETFs. He referenced Morgan Stanley’s MSBT, which arrived at a 14 basis-point fee.
The SEC initially approved US spot Bitcoin ETFs in January 2024, and those funds have gone on to accrue inflows to the tune of $57.7 billion.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.