Anthropic has proposed new AI policy frameworks as advanced systems gain stronger capabilities.

The company wants governments to set rules for frontier models and prepare workers for AI’s economic impact. Its plan covers dangerous deployments, independent testing, cybersecurity, and public resilience.

Anthropic seeks stronger AI safety powers

Anthropic introduced two proposals under its “Policy on the AI Exponential” plan. The Advanced AI Framework focuses on powerful models, while the Economic Policy Framework addresses workers and shared financial benefits. The company argued that AI now moves faster than current policymaking systems. It also said governments need authority to block or deter dangerous model deployments.

Under the plan, civil penalties would tie to global annual revenue. Repeat violations would bring higher penalties, based on the proposed framework. The framework also calls for frontier developers to test models before release. Developers would publish summaries, safety frameworks, and system cards for powerful AI systems.

Independent evaluators would review model tests and risk reports. Anthropic also wants developers to maintain strong security programs for model weights and training systems. The proposal supports transparency laws in states such as California and New York. However, the company argued that public disclosure alone no longer matches the speed of AI development.

The framework targets catastrophic AI risks

The proposed rules would apply only to the most advanced AI systems. Anthropic set the threshold at models trained above 10²⁵ floating-point operations. The framework would also cover companies earning more than $500 million in AI-related revenue. Firms spending more than $1 billion on AI research and development would also fall under it.

Anthropic named four main risk areas in the proposal. These include biological risk, cyber risk, loss of control, and automated AI research. For biological risk, the company warned that unsafe systems could help attackers develop harmful viruses. It also noted that similar AI tools can support drug discovery.

For cyber risk, frontier models can find serious software flaws at large scale. Anthropic said those capabilities raise concerns for hospitals, energy grids, and other key systems. The company also highlighted risks from systems acting outside developer control. Automated AI research could increase biological, cyber, and control risks if safeguards fail.

Developers face testing and security duties

Anthropic wants frontier developers to publish regular risk reports. These reports would describe the developer’s overall risk posture and model safety work. The framework also calls for at least one qualified independent evaluator. That evaluator would review company evaluations and publish findings on model risk reports.

Governments and industry would also set standards for those evaluators. The proposal says evaluators need funding and access to frontier models. Security rules form another major part of the framework. Developers would protect their full development environment from outside attackers and insider threats.

Companies would describe their security programs publicly at a high level. They would also share more details with a designated government agency when requested. Anthropic said policymakers could start with lighter rules and adjust them over time. The framework says regulation should follow model capabilities and evaluation standards.

The proposal includes resilience measures

The second part of the framework focuses on public resilience. Anthropic recommended stronger planning for biological, cyber, and control-related AI risks. For biology, the proposal includes gene synthesis screening and early-warning biosurveillance. It also mentions protective equipment stockpiles and tools to reduce airborne transmission.

For cyber, the framework calls for stronger internet software and support for critical infrastructure operators. It also recommends replacing legacy systems in essential infrastructure. Governments should also track frontier cyber capabilities through a dedicated function. Anthropic proposed joint work between government and industry on model safeguards.

The company said work on loss-of-control and automated research risks remains less developed. It called for better tools to detect, contain, or shut down unsafe systems. Anthropic urged policymakers to act as model capabilities continue improving. The company said AI governance must keep pace with the technology.

Decentralized exchange Raydium has reportedly suffered a $1.3 million exploit that saw attackers drain the firm’s old liquidity pools.

Crypto investigator “Specter” spotted what they believed to be a Raydium exploit at 3pm GMT+1, claiming that the funds have been bridged to Ethereum and are now being laundered via Tornado Cash. 

They also shared what they believe to be the attacker’s addresses:

• 0x0EaBAAb9a56011c6158D4aA7f2E49A82fB34E609
• 4WnPebowR4HHfumvNPaDjG6Pa5Hi1jxLm6xmmBq33QVk.

Since Specter’s post, Raydium official “Infra” has revealed that the firm is aware of the exploit and is conducting a security review to determine what happened. 

Infra says, “No current users of Raydium are affected by this exploit or would have been able to interact with these pools through the UI since their deprecation.” 

They claim that after an initial review, 150,177 RAY, 5,603 SOL, and 893,700 USDC have been stolen. Together, the stolen funds are worth roughly $1.34 million.

Read more: One laptop: How poor security ruined Humanity Protocol

The cause of the attack has been attributed to a vulnerability associated with “insufficient validation of the LP mint,” and an exploit with Raydium’s legacy AMM V3 program that was phased out in 2021.

Infra said, “Because the program did not properly verify the LP mint address, an attacker was able to create a new mint and use it as the LP token, bypassing the intended proportion checks.”

It added that all other Raydium mainnet programs avoid this vulnerability as they use a “virtual supply mechanism for proportion checks and correctly verify the LP mint along with all other relevant account information.”

Infra ruled out any key compromise or authority-level issue, and claimed that the attack was caused “by a self-contained logic flaw.”

According to Infra, affected users will be fully compensated by Raydium’s treasury.

Got a tip? Send us an email securely via Protos Leaks. For more informed news and investigations, follow us on X, Bluesky, and Google News, or subscribe to our YouTube channel.

Since 2024, Bitcoin (BTC) has posted four major corrections after interest rate hikes by the Bank of Japan (BOJ), with declines ranging from 18% to 28%. This dynamic places renewed attention on the BOJ’s June 16 policy decision. 

Data currently point to a variety of pressures on BTC, with BTC whale distribution and exchange inflows possibly carrying more weight than Japanese monetary policy.  

BOJ hikes and Bitcoin drawdowns: Will history repeat?

The relationship between BOJ policy and Bitcoin has gained attention because each rate increase since Japan ended its negative interest rate policy has been followed by a sizable correction. 

Following the March 19, 2024, hike, Bitcoin corrected by 18%. The July 31, 2024, increase preceded a 18.5% decline. 

After the Jan. 24, 2025, hike, Bitcoin fell nearly 25%, while the Dec. 19, 2025, decision was followed by a 28% drawdown. 

Across the four events, Bitcoin’s average decline was 22.4%. 

BTC/USD, one-week chart. Source: Cointelegraph/TradingView

The sell-offs did not occur under identical conditions. The March 2024 correction followed Bitcoin’s breakout to new all-time highs during the spot Bitcoin exchange-traded fund (ETF) cycle. The July 2024 decline followed months of consolidation below peak levels and coincided with the sharp unwind of the yen carry trade, which affected global markets. 

The January and December 2025 drawdowns followed extended rallies and periods of contraction for both BTC spot and futures 30-day demand. 

BTC: spot and perpetual futures demand growth contraction. Source: CryptoQuant

The relationship between BOJ policy and Bitcoin is often linked to the yen carry trade. For years, investors borrowed yen at low rates and deployed that capital into higher-yielding assets, including stocks and cryptocurrencies. 

When the BOJ raises rates, some of those positions can be reduced, weighing on risk assets. The July 2024 hike coincided with one of the largest carry-trade unwinds in recent years and a sharp sell-off across global markets, not only BTC. 

The influence of that particular condition appears smaller today. The BOJ has already raised rates to 0.75% from -0.1% in March 2024, while Japan’s 10-year government bond yield climbed to 2.68% from 0.63% over the same period. 

Japan’s 10-year bond yield increase since 2024. Source: TradingEconomics

With Japan’s borrowing costs already higher than during the negative-rate era, each additional hike represents a smaller policy shift than the BOJ’s initial move away from ultra-loose monetary policy. The June 16 meeting would extend an existing tightening cycle rather than introduce a new one. 

Likewise, market analyst Cryptic Trades noted that concerns about a renewed yen carry-trade unwind are overblown, arguing that Japan has effectively moved away from its deflationary policy framework in 2024. The analyst added,

“The Yen Carry Trade has been dead ever since 2024. It is also a BIG nothing burger for the markets.”

Related: Bitcoin price may slide toward $30K as institutions dump 450% of daily BTC supply

BTC whales add to the pressure

While the BOJ meeting is a macro event that traders may monitor, onchain data points to a more immediate source of pressure.

Crypto analyst MorenoDV noted that Binance has recorded rising BTC inflows from wallets holding 100–1,000 BTC and 1,000–10,000 BTC since the sell-off began in early June. As a result, the exchange’s 30-day whale inflow sum has climbed to $6.6 billion. 

Bitcoin whale to exchange flow. Source: CryptoQuant

The pressure is already visible in realized activity. Short- and long-term whales have collectively locked in more than $2.5 billion in losses during the decline, indicating that some large holders have actively reduced exposure.

Short-term whales appear particularly vulnerable. The cohort is carrying roughly $16 billion in unrealized losses after briefly returning to profit for around 10 days in early May. Those positions now sit close to break-even levels, creating a potential source of supply during rebounds. MorenoDV said, 

“Taken together, these three readings describe the stress profile of a late-stage bear market: capitulating whales, distribution into weakness, and a fragile short-term cohort with its finger on the trigger.”

Related: Bitcoin may act as a ‘canary in the coal mine’ as risk-off pressure spreads: Bitwise

Mastercard (MA) is betting that AI agents will soon become active participants in the economy and wants its payments network to sit at the center of that shift.

The payments giant on Tuesday unveiled Agent Pay for Machines (AP4M), a service enabling AI agents and software systems to make payments to each other securely and at scale. The platform supports automated transactions across cards, bank accounts and stablecoins, while providing identity verification, spending controls and guaranteed settlement through Mastercard’s network.

The service comes as companies across technology, payments and crypto race to build infrastructure for what many are calling agentic commerce, where AI systems complete tasks, purchase services and coordinate transactions on behalf of users. Agents could be involved in trillions of dollars worth of transactions by the end of the decade, according to some estimates.

“We are already seeing a number of services and agents popping up to provide a range of products and services,” Raj Dhamodharan, Mastercard’s executive vice president of blockchain and digital asset products and partnerships, told CoinDesk. Those services range from booking travel and building websites to creating artwork and completing other digital tasks.

Dhamodharan said the next challenge is creating trust between those systems.

Businesses and consumers need confidence that agents are interacting with legitimate counterparties and operating within authorized spending limits. Service providers, meanwhile, need assurance that they will be paid.

“These are problems that we’ve solved before in the B2B world and the carded world for decades,” Dhamodharan said. “We’re bringing the same level of trust and ability to find the right set of agents, ability to convey that you’re actually going to complete the payment and to make sure that people can get paid.”

The platform is designed to address those concerns through credentialing, permissioning and settlement services. The company said the system can authenticate agents, enforce spending rules and settle payments across multiple payment methods, including stablecoins.

More than 30 companies have joined the initiative, including Coinbase (COIN), Stripe, Adyen, Checkout.com, Cloudflare, RippleX, Polygon Labs, Solana Foundation and OKX. Mastercard said permissions and credentials associated with AI agents will initially be recorded on the Polygon, Solana and Base blockchains.

Although large-scale machine-to-machine commerce remains nascent, Dhamodharan said Mastercard is already seeing signs of demand. He pointed to increasing activity around HTTP 402, an emerging internet payment standard, where automated transactions often fail because no payment method is available.

“There are already transactions happening,” he said. “There are already many declines happening because there is no payment option available. That is a leading indicator in our view.”

Mastercard said it plans to expand access to Agent Pay for Machines later this year.

If bitcoin and Ethereum had been invented on the same day, nobody would have heard of bitcoin. I sold every bitcoin Bit Digital held and deployed the proceeds into Ethereum. I have built one of the largest corporate Ethereum treasury positions in the world and said, on the record, that we will never sell it. People have asked me to articulate the single strongest argument for that conviction. On March 30, 2026, that argument arrived. Last month, Citi confirmed it.

In a research note published on May 18, Citi analysts warned that quantum computing advances have shortened the timeline for practical attacks on digital assets, and reached a conclusion that should give every institutional bitcoin holder pause: bitcoin faces significantly greater quantum risk than Ethereum, and the gap between them comes down not just to technology but to governance.

That finding echoes the landmark paper released in late March by Google Quantum AI in collaboration with Stanford University and the Ethereum Foundation, which found that the computing resources required to break bitcoin’s foundational cryptography are approximately 20 times lower than previously estimated. A sufficiently advanced quantum computer, operating with fewer than 500,000 physical qubits, could derive a bitcoin private key from its public key in roughly nine minutes. That machine does not exist today. But the window to act responsibly is narrowing faster than most institutions realize. When Google raises the alarm, and Citi confirms it in the same quarter, this is no longer a fringe concern. This is the silver bullet. And it points directly at bitcoin.

Why bitcoin is exposed

Bitcoin’s security rests on elliptic curve digital signature algorithms. When you spend bitcoin, your public key is briefly exposed onchain. Under classical computing, reversing that to obtain a private key is infeasible. Quantum computers running Shor’s algorithm can, in principle, do exactly that during the brief window a transaction is broadcast. The Google paper doesn’t merely confirm this theoretically; it quantifies it with a precision that removes comfortable ambiguity.

Nic Carter, co-founder of Coin Metrics and one of the sharpest minds in digital assets, has been sounding this alarm for months. In a series of essays beginning in October 2025, Carter called quantum computing “the biggest long-term risk to bitcoin’s core cryptography” and accused developers of “sleepwalking towards collapse.” He estimates a quantum computer could meaningfully break elliptic curve cryptography as early as 2028. Approximately 6.9 million BTC could be vulnerable at a sufficient quantum scale, including legacy wallets and Taproot outputs, which already represented more than 21% of all bitcoin transactions in 2025.

Bitcoin’s governance problem

One might ask: can’t bitcoin simply upgrade? Yes, in theory. In practice, this is where the risk compounds.

Bitcoin’s governance is intentionally conservative and consensus-driven, which makes it extraordinarily slow. SegWit took roughly 8.5 years from conception to widespread adoption. Taproot took approximately 7.5 years. The current quantum proposals, BIP-360 and BIP-361, are still at the draft or early testnet stage as of 2026. A full base-layer transition to post-quantum signatures would be the most contentious change bitcoin has ever attempted. As Carter documented, most bitcoin Core developers have expressed limited concern about urgency, a disposition that is, at minimum, a serious governance liability for any institution holding bitcoin in treasury. A quantum breakthrough does not politely wait for committee consensus.

Ethereum has already acted

This is where the picture diverges sharply. Ethereum’s approach to quantum resistance is not a reactive scramble. It is a structured road map already in execution, built on the NIST post-quantum cryptography standards finalized in August 2024.

The Pectra upgrade, which shipped on Ethereum mainnet in May 2025, introduced EIP-7702, a critical stepping stone toward full account abstraction. Rather than requiring a single network-wide hard fork, Ethereum’s architecture allows individual accounts to choose their own signature verification and switch to quantum-safe signatures voluntarily. The upcoming Hegotá hard fork, planned for the second half of 2026, embeds this further at the protocol level. The Ethereum Foundation has set structured milestones targeting completion of core post-quantum infrastructure by approximately 2029, with active interop devnets already running across multiple clients.

The contrast with bitcoin’s governance paralysis could not be more stark. Ethereum was designed, in ways bitcoin simply was not, to accommodate exactly this kind of foundational upgrade. That is not an accident. It is architecture.

The institutional calculus

For corporate treasurers and sovereign wealth managers, quantum risk is no longer a tail scenario to be footnoted and dismissed. Governments are already treating it as operational. U.S. federal agencies faced an April 2026 deadline to submit post-quantum cryptography transition plans under National Security Memorandum 10. The EU has set a 2030 quantum-resistance target for critical infrastructure. The G7 Cyber Expert Group published a coordinated financial sector road map in January 2026. This compliance architecture will, over time, extend to digital asset treasury holdings.

The question for any institution holding bitcoin is whether they are comfortable with an asset whose quantum-resistance road map is still in draft, whose governance moves at geological speed, and whose developer community is divided on whether urgency is even warranted.

The question for any institution considering Ethereum is whether they want the asset with a structured, transparent, and already in motion upgrade path.

Ethereum is the more adaptive, more capable, and more durable asset. I have put the balance sheet of a Nasdaq-listed company behind that conviction. The Google paper is what finally gives that conviction a single, undeniable, technically grounded answer to the hardest question in digital asset treasury strategy: which asset is built to last?

Ethereum is not a perfect asset. No asset is. But in the context of quantum risk, it is the asset whose architecture was built to survive what is coming. If Carter and Google are right, that distinction will matter enormously, and sooner than most people expect.

TLDR

• RippleX launched the XRP Ledger AI Starter Kit on June 10.
• The kit enables autonomous payments for AI agents on XRP Ledger.
• XRPL now supports the X402 protocol for web-based software payments.
• AI agents can pay for APIs, model inference, and digital services using XRP and RLUSD.
• Ripple-backed startup t54 helped integrate XRPL into X402.

Ripple Labs moved to link blockchain payments with the fast-growing AI economy through a new developer release. The company introduced the XRP Ledger AI Starter Kit on June 10 to support autonomous transactions. The rollout aligns with efforts to connect software agents with direct web payments.

XRP Ledger integrates X402 for autonomous web payments

RippleX launched the XRP Ledger AI Starter Kit to help developers build agent-powered applications. The first phase supports tools that enable autonomous payments on the XRP Ledger network. RippleX said the network design supports fast settlement and predictable costs.

“The XRP Ledger was built with many of these qualities in mind,” the announcement stated. It added that agentic payments now require speed and low transaction fees. The release supports X402, an open protocol that enables software to send payments without human approval.

Through support from t54, XRPL now operates as a supported chain on X402. Ripple backed t54 during a seed funding round to expand payment infrastructure. As a result, AI agents can use XRP and RLUSD for web-based transactions.

The system allows AI agents to pay for API calls and model inference services. It also enables payments for other digital services across web platforms. Developers can integrate these features into applications that require automated billing.

The starter kit also provides tooling for AI coding agents. A dedicated Model Context Protocol server supports queries to XRPL documentation. Clients, including Claude Code, Claude Desktop, and Cursor, can access the documentation directly.

XRP and RLUSD power Mastercard’s agent payment framework

The launch occurred on the same day Mastercard Inc. introduced Agent Pay for Machines. Mastercard designed the framework to support autonomous payments across digital services. More than 30 partners joined the initiative, including Ripple and t54.

RippleX senior vice president Markus Infanger outlined the role of XRPL and RLUSD. He said the network provides a settlement layer that clears in seconds. He added that the system offers predictable costs and built-in compliance.

“XRPL and RLUSD give Mastercard’s framework a settlement layer that clears in seconds,” Infanger said. He also referenced programmable compliance and a full audit trail. Mastercard listed Ripple as one of the core blockchain partners.

The framework supports high-volume and low-value transactions between software agents. It enables machine-to-machine settlements without manual approval. RippleX confirmed that XRP and RLUSD serve as payment rails within the system.

Ripple Labs continues to expand infrastructure through RippleX initiatives. The XRP Ledger AI Starter Kit now rolls out in stages across developer channels. The company released the tools on June 10, alongside Mastercard’s AP4M framework.

Stand With Crypto UK, representing about 286,000 crypto enthusiasts and professionals, is pressing its network to challenge banks that block or restrict transfers to cryptocurrency exchanges. The campaign cites a UK Cryptoassets Business Council report indicating that 40% of crypto transactions are blocked or restricted by banks, with many restrictions applying to exchanges registered with the Financial Conduct Authority and not accounting for individual customer risk profiles.

The campaign outlines that one exchange recorded nearly £1 billion in declined transfers over a 12-month period, and 80% of surveyed platforms reported an uptick in blocked or restricted transfers. To push the issue forward, Stand With Crypto UK has launched a complaint-tool on its website that auto-generates letters challenging these restrictions; the organization says bank responses will help shape the next steps. Their tagline captures the spirit of the effort: “Your money. Your choice.”

Key takeaways

• According to the UK Cryptoassets Business Council, around 40% of crypto transactions are blocked or restricted by banks in the United Kingdom.
• One exchange reported nearly £1 billion in declined transfers over a year, illustrating the scale of friction for on-ramps and off-ramps.
• Approximately 80% of surveyed exchanges and platforms said they have seen higher rates of transfer blocking or restriction.
• The Stand With Crypto UK tool enables supporters to generate complaint letters to banks, with regulator-facing responses expected to influence the campaign’s next moves.
• Policy context in the UK centers on stabilizing and regulating stablecoins, with ongoing scrutiny of how such assets fit into the domestic financial system.

Bank access under scrutiny and the push for targeted risk controls

The campaign argues that blanket transfer bans hinder access to digital assets and stifle competition in a sector that is increasingly regulated. By urging members to file complaints, Stand With Crypto UK aims to convert anecdotal friction into formal regulatory and industry dialogue. In its public communication, the group emphasizes that many restrictions appear to apply broadly, regardless of a customer’s risk profile or the specific platform involved.

Industry voices calling for risk-based solutions

Industry observers have long warned that broad prohibitions can hamper legitimate crypto activity. In commentary to Cointelegraph, Mark Fairless, CEO of UK clearing bank ClearBank, underscored the need for a risk-based approach to crypto-related payments rather than sweeping blocks across the sector. “Interventions should be targeted and proportionate, as broad blocks risk undermining competition and the ability of regulated firms to operate effectively in the UK,” Fairless said.

Regulatory backdrop: UK stablecoins and the broader digital asset framework

The Stand With Crypto UK initiative arrives amid a wider regulatory effort to shape the UK’s stablecoin regime. In early May, a House of Lords committee examined proposed stablecoin regulations, questioning industry executives about bank-run risks, anti-money laundering controls, and the potential impact on traditional banking. Later in May, the Bank of England signaled a softer stance on proposed caps and reserve requirements as part of its review of the pound-denominated stablecoin framework. The objective is to foster a domestic stablecoin market while limiting risks to bank funding and financial stability; non-dollar stablecoins currently account for a small share of the global market. In June, the Lords committee warned that certain proposed stablecoin requirements could limit the viability of pound-denominated tokens and urged regulators to avoid measures that would hinder sector growth.

Beyond stablecoins, broader digital-asset initiatives are gaining momentum. In May, the central bank floated extending operating hours for settlement infrastructures to support tokenized markets, while the Financial Conduct Authority proposed on June 8 allowing some retail-focused investment funds to allocate up to 10% of their portfolios to crypto exchange-traded products.

What to watch next

As the UK weighs its stablecoin framework and the broader integration of digital assets into mainstream finance, the outcomes of this campaign could influence how banks calibrate risk and how regulators balance access with safety. The coming weeks will clarify whether targeted interventions replace blanket blocks and how exchanges and users adapt to evolving policy expectations.