If you’ve been tracking crypto markets, you’ve likely seen analysts flag bitcoin’s BTC$62,731.27 200-week moving average (200W MA) as a key inflection line where the broader trend can turn decisively higher. This line represents bitcoin’s average closing price over roughly four years.

Strategy is now tracking the same indicator, along with bitcoin’s premium (or discount) to that average. The firm’s Founder, Michael Saylor, announced it on X on Sunday, reinforcing the long-term average as a pivotal level for traders and investors alike.

“We’re now tracking Bitcoin’s 200-week moving average and its premium to that level on http://Strategy.com. Since the 200W MA became available, Bitcoin has traded above it 92% of the time. Today, it sits almost exactly on the line,” Saylor said Sunday on X.

However, in the hours since the post, prices have come under pressure, likely over concerns about a delay in the passage of the long-awaited Clarity Act, which is expected to unlock a significant institutional bid for digital assets. According to reports, the Senate did not list the Clarity Act in Monday’s agenda.

BitGo CEO Mike Belshe challenged Anthropic’s Claude models on Aug. 1 by publishing a Bitcoin address holding 100 BTC and inviting the AI system to move the funds. 

BitGo’s official website identifies Belshe as its co-founder and chief executive.

Belshe framed the wager as a response to Anthropic’s July 30 disclosure that Claude models accessed three real organizations during cybersecurity evaluations. He wrote that there had been “enough with the ‘we created a hacking monster’ games” and called for a real-world demonstration.

BitGo’s 100 BTC wallet remains untouched

The address received 100 BTC on July 31, according to publicly indexed blockchain reports. On-chain checks cited through Aug. 2 showed the full balance remained at the address with no outgoing transaction.

Either @AnthropicAI is terrible at building sandboxes… or excellent at marketing. (or both)

But enough with the “we created a hacking monster” games.

Do it for real.

Advertisement

I put this in an @BitGo wallet for you. Go get it.

100 BTC:… https://t.co/RhvivRk9YK

— Mike Belshe (@mikebelshe) August 1, 2026

The absence of movement does not establish that Claude tried and failed. Belshe’s post did not describe an evaluation setup, grant access to BitGo systems or identify which Claude model should participate. It instead created a public, observable target whose balance can be monitored through the Bitcoin network.

Meanwhile, Anthropic said it found three incidents after reviewing 141,006 evaluation runs. The cases involved Claude Opus 4.7, Mythos 5 and an internal research model. A misunderstanding with evaluation partner Irregular left test machines connected to the internet, even though prompts told the models they were inside sealed simulations.

The models then used basic techniques, including weak passwords and unauthenticated endpoints, against real infrastructure they treated as part of capture-the-flag exercises. Anthropic said the models did not deliberately escape or pursue independent goals. It described the episodes as closer to an “operational failure” than a model alignment failure. The models also lacked the standard classifiers and monitoring used in Anthropic’s publicly available products.

In the most serious case, Opus 4.7 obtained credentials and reached a database containing several hundred production records. Mythos 5 separately published a malicious package to the public PyPI registry. Anthropic said the package remained available for about one hour and ran on 15 real systems before its removal.

The Claude challenge is not a like-for-like test

Publishing a Bitcoin address does not provide the credentials needed to spend its funds. Bitcoin transactions require valid cryptographic signatures produced with the relevant private keys. BitGo’s technical documentation says its Bitcoin multisignature wallets generally require two of three independent keys to authorize a transaction.

Claude would therefore need access to a signing environment, key material or an exploitable operational weakness. The address alone supplies none of those. Moreover, the exact signing policy behind this particular unspent output cannot be confirmed from Belshe’s post alone, even though he identified it as a BitGo wallet.

That makes Belshe’s challenge a test of whether an AI-enabled attacker could breach BitGo’s wider controls, rather than whether Claude can derive private keys from public blockchain data. A controlled comparison would also require agreed rules, authorized access, activity logs and independent verification of any attempted attack.

As previously reported, BitGo has tested post-quantum MPC signing for institutional custody. In related coverage, crypto.news examined how Claude Mythos 5 could accelerate attacks against exposed keys, weak signing flows and misconfigured systems without creating a universal ability to defeat cryptography.

U.S. scrutiny now shifts toward testing controls

The dispute arrives as U.S. officials review how advanced AI systems should undergo cybersecurity testing. President Donald Trump directed advisers in June to develop a voluntary testing framework for leading models, according to Reuters. Anthropic’s disclosure may add pressure for clearer containment standards and incident reporting across AI laboratories and external evaluators.

Anthropic stopped its cyber evaluations on July 23, identified all three incidents the following day and notified affected organizations on July 27. It said METR would conduct an independent review and that it planned to release a redacted transcript of the malicious-package incident within one week. Irregular is conducting its own investigation.

Those disclosures, rather than movement from Belshe’s wallet, are the next formal checkpoints. Any transaction from the address would be visible onchain. However, investigators would still need to establish who authorized it, how the signing requirements were satisfied and whether a Claude model played any role.

Binance will remove Across Protocol, Hashflow, PIVX, Vulcan Forged PYR, Vanar and Viction from spot trading on Aug. 17, 2026, at 03:00 UTC after completing its latest asset review.

The exchange said every spot pair tied to ACX, HFT, PIVX, PYR, VANRY and VIC will close. Outstanding spot orders will be canceled. Binance did not identify a separate reason for each asset. Instead, it cited its broader review framework, which covers liquidity, development activity, network safety, team conduct, transparency, tokenomics and regulatory changes.

Binance EN: Binance Will Delist ACX, HFT, PIVX, PYR, VANRY, VIC on 2026-08-17

币安重要公告: 币安将于2026年8月17日下架 ACX、HFT、PIVX、PYR、VANRY、VIC

Advertisement

PYR MarketCap: 3.5M
PIVX MarketCap: 3.4M
HFT MarketCap: 7.7M
VANRY MarketCap: 8.8M
ACX MarketCap: 38.4M
VIC MarketCap: 5M…

— 方程式新闻 BWEnews 🏎️ (@bwenews) August 3, 2026

Binance delisting begins with an Aug. 7 futures cutoff

The first major deadline arrives before the spot removal. The company Futures will prevent users from opening new positions at 08:30 UTC on Aug. 7. It will close and automatically settle remaining contracts at 09:00 UTC. The exchange may also change leverage, margin tiers, funding rates or index components before settlement if markets become unusually volatile.

Loans and several payment services will also close that day. Binance Pool and Binance Pay will stop supporting the assets at 03:00 UTC. VIP Loan and Flexible Loan positions will close at 07:00 UTC, while cross and isolated margin positions will be settled at 10:00 UTC. Margin borrowing will already be suspended from 06:00 UTC on Aug. 4.

Spot Copy Trading will remove the affected pairs on Aug. 10. Remaining assets may be sold at market prices or transferred to users’ spot accounts when they cannot be sold. Simple Earn will redeem flexible and locked positions after 07:00 UTC on the same day and transfer the assets and accrued rewards to spot accounts.

Four Binance delistings followed earlier risk warnings

The decision was preceded by Monitoring Tags on four of the six tokens. The exchange added PIVX to the tag list on June 18, followed by PYR and VANRY on July 3. ACX received the tag on July 24. The exchange states that tagged assets carry greater volatility and risk and may be removed if they no longer satisfy its listing standards.

As crypto.news previously reported, the ACX warning came days before the latest removal decision. Monitoring Tags do not guarantee delisting, but they require users to pass a risk quiz every 90 days and notify holders that the exchange is conducting closer reviews.

In related coverage, Binance removed 20 tokens from its Alpha platform in May while preparing five other assets for spot delisting. The Alpha removals and full spot delistings were separate processes, but both followed reviews against the exchange’s platform standards.

ACX also entered Binance’s delisting process after Coinbase suspended its trading on July 28. Coinbase said the project team was winding down the token and directed holders to Across documentation.

Across previously proposed replacing its token-based DAO with a U.S. C-corporation. The published plan set out an equity exchange and a USDC buyout at $0.04375. However, legal restrictions apply to the equity option, and the proposal said its estimated timetable could change.

VANRY holders must complete the Base migration themselves

VANRY presents an extra operational issue. The exchange said it will not support Vanar’s contract swap. Holders seeking the replacement token must use the project’s migration portal rather than expecting Binance to complete the conversion automatically.

Vanar announced a 1:1 migration to Base and said the new token would have a supply of 10 billion. The project previously told users that participating centralized exchanges would handle the swap automatically. Binance’s new notice confirms that it is not one of those supporting venues.

The exchange will keep current VANRY withdrawals open through Ethereum and Polygon PoS. That gives Binance users a route to remove their tokens before completing the migration through Vanar’s official portal.

Vanar has warned holders to rely only on links distributed through its verified channels. The project advised users to ignore unsolicited messages and never share wallet seed phrases while completing the migration.

Withdrawals remain open until Oct. 17

The exchange Convert will remove the six assets at 02:00 UTC on Aug. 17, one hour before spot trading ends. Its low-value asset conversion feature will stop supporting them on Aug. 14. Deposits made after 03:00 UTC on Aug. 18 will not be credited.

Withdrawals will remain available until 03:00 UTC on Oct. 17. The exchange may convert balances left on the platform into stablecoins after Oct. 18, but the exchange said that conversion is “not guaranteed.” It will issue another notice where conversion is possible.

When conversion is not feasible, Binance said withdrawals may remain open, subject to network availability. Users should not rely on that possibility because the exchange has not committed to providing an extended withdrawal window.

The removal covers six tokens facing different project conditions rather than one shared event. Binance’s announcement gives users a common timetable but no token-by-token findings. The next confirmed developments will come from project responses, settlement notices and any changes to the withdrawal or migration arrangements.

Coldcard’s five-year seed-generation issue has turned into a wider debate over how hardware wallets are independently verified, according to Kraken’s chief security officer Nick Percoco. In an X post on Sunday, Percoco said the incident should prompt makers of self-custody devices to require end-to-end checks that confirm the randomness source reviewed in testing is the same one actually executed by production firmware.

The comments arrive amid an ongoing exploit believed to target vulnerable Coldcard devices by abusing weak seed phrases. By Sunday, more than 4,500 addresses had reportedly been affected, with losses estimated at nearly $90 million in Bitcoin, according to Cointelegraph’s reporting linked in the original article.

Key takeaways

• Kraken’s Nick Percoco argues hardware wallet certification should include verification that the approved entropy path is what production firmware uses in practice.
• Coldcard’s RNG-related flaw allegedly persisted for years after a seed-generation change introduced an unintended reliance on a weaker generator.
• Percoco cited existing standards used in the broader security and payments industries—such as NIST SP 800-90B and BSI AIS-31—as models for what should be standard for crypto self-custody.
• Coinkite says affected firmware has been halted in shipments and that remaining units containing the vulnerable code were destroyed, while it advised users not to discard certain devices.

A hardware wallet “wake-up call” for entropy verification

Percoco’s central point is about trust boundaries. Hardware wallet users are asked to rely on a manufacturer’s implementation of the randomness function that ultimately underpins seed phrase generation—yet, he said, there is often no independent method to confirm that the verified randomness source is the one the device will actually call in production.

“Consumers are asked to trust a manufacturer’s implementation of the single most critical function in the system, with no independent verification that the approved entropy path is the one actually executing,” Percoco wrote in his Sunday post.

He described this gap as an industry-wide weakness rather than a one-off mistake, noting that while some certifications exist for hardware components and secure elements, they do not “systematically force end-to-end verification” of the entropy source through to production code execution.

Percoco contrasted the crypto self-custody space with practices in other sectors. He pointed to the payments industry’s use of independent lab testing for devices that collect sensitive inputs, and to government expectations in the US around cryptographic module validation and entropy source testing.

How the Coldcard flaw allegedly slipped through

According to the original reporting, the vulnerability traces back to a software change disclosed by Coinkite. The company said the relevant issue has existed since March 2021, when Coldcard altered its seed-generation process as it integrated a new cryptographic library.

The update, per Coinkite’s postmortem referenced in the article, unintentionally routed wallet creation to a weaker MicroPython generator already present in the codebase. Coinkite’s explanation indicated that Coldcard’s intended true random number generator (TRNG) code existed and could be present and functioning, but was not reliably the one used for the core randomness needed for seeds.

In other words, reviewers could verify that the TRNG code was present and worked—but, without a mechanism to ensure the device actually called that TRNG during seed generation, the system could still produce outcomes derived from a different generator than intended.

The practical consequence is that seed phrases generated under the affected conditions may become more predictable than they should be. That predictive weakness is widely viewed by the security community as especially dangerous in wallet designs because compromised seeds can enable theft without needing to break keys directly.

Attack fallout and what’s changing for users

The ongoing exploit believed to target weak seed phrases generated by affected Coldcard devices has already resulted in extensive on-chain activity. As of Sunday, Cointelegraph’s figures cited in the original article reported over 4,500 impacted addresses and losses approaching $90 million in Bitcoin.

Coldcard (Coinkite) said it has halted all device shipments since confirming the vulnerability on Thursday. It also stated it destroyed remaining units at its facilities that contained the affected firmware.

At the same time, the company advised users with affected devices not to dispose of them immediately, saying doing so might become “essential if funds are recovered.” The company also indicated its legal team would coordinate with law enforcement across multiple jurisdictions to support efforts identifying those responsible.

For affected owners, the new information underscores a key operational point: device handling decisions may need to be aligned with recovery processes rather than treated as purely disposal or cleanup tasks. While that doesn’t eliminate the security risk of continuing exposure, it suggests an active incident-response posture where retaining evidence or workable hardware could matter.

Standards exist—what’s missing is enforcement

Percoco’s critique points to a tension that many investors and builders may recognize: crypto security often emphasizes reviewing code paths and cryptographic primitives, but not always the end-to-end behavior under production conditions—especially the specific entropy source used at runtime.

He referenced NIST SP 800-90B, which sets requirements for designing, testing and validating physical true random number generators for cryptographic security, and BSI AIS-31, a similar standard from Germany’s Federal Office for Information Security. In his view, such frameworks make it more difficult for systems to “pass review” without proving that the approved randomness pathway is actually used for critical operations.

Whether regulators and certifiers will adapt those expectations to consumer self-custody products remains uncertain. However, the Coldcard case demonstrates why the distinction matters: even when a correct TRNG implementation exists in the codebase, the seed-generation workflow can still be compromised if production firmware routes randomness differently than what independent review assumes.

Next, investors and users should watch for two things: clarification from Coinkite on exactly how to identify which devices/firmware are affected and what remediation steps best reduce future risk, and whether independent testers or certifiers move toward stronger “entropy source at runtime” validation—an area Percoco argues should not remain optional in digital asset custody.

South Korea saw 560.3 billion won ($367 million) in stablecoin outflows to overseas exchanges in June, extending the country’s streak of monthly net stablecoin outflows to 18 consecutive months. 

The figure comes from Financial Supervisory Service (FSS) data obtained by Yonhap News Agency through People Power Party lawmaker Lee Jong-wook. South Korea’s five major crypto exchanges — Upbit, Bithumb, Coinone, Korbit and Gopax — transferred 2.7 trillion won ($1.81 billion) in stablecoins offshore in June and received 2.2 trillion won ($1.44 billion) from foreign platforms.

Market participants cited by Yonhap attributed the transfers to demand for products restricted or unavailable on domestic exchanges, such as overseas derivatives, tokenized real-world assets (RWAs), decentralized finance and staking products. 

Lee has called on the government to reassess how it protects investors and supervises cross-border crypto activity as stablecoin outflows continue. “The government must comprehensively examine its investor protection and supervisory frameworks again and move swiftly to improve regulations,” he said, according to The Korea Times.

South Korea weighs tighter rules for offshore activity

The outflows come as South Korea works to complete a broader legal framework for digital assets. On Thursday, a policy report recommended that authorities introduce an interim licensing guidance and phase in stablecoin regulations before the Digital Asset Basic Act is finalized. 

The proposed act would create the country’s first comprehensive digital asset framework, including rules for stablecoin issuance, disclosures and market activity. However, lawmakers have yet to reconcile multiple proposals, with disagreements over which institutions should be allowed to issue won-pegged stablecoins contributing to delays. 

Related: South Korea plans stablecoin rules as opposition pushes crypto tax repeal

South Korean regulators have also sought to expand reporting requirements for crypto transfers. On June 22, South Korea’s Financial Intelligence Unit (FIU) proposed extending Travel Rule reporting requirements to transactions below 1 million won (about $650).

The FIU also called for stronger action against unregistered overseas exchanges serving South Koreans. The agency said uneven licensing and supervision across jurisdictions created opportunities for regulatory arbitrage, a concern underscored by the country’s continued stablecoin outflows.

Magazine: The real reason DeFi projects that survived 2022 crash are shutting down now

Cointelegraph is committed to independent, transparent journalism. This news article is produced in accordance with Cointelegraph’s Editorial Policy and aims to provide accurate and timely information. Readers are encouraged to verify information independently.