Track market-moving stories before they get noisy

5 hours ago

According to on-chain data, whale Lee Goon Wang just sold 15,000 ZEC tokens on Hyperliquid via a limit sell order, valued at approximately $23 million. The limit order’s listed price was about $30 lower than the market price at the time, a drop of roughly 2%.

Source

Scan the QR code

Download APP

Meta has introduced the Meta Enterprise Platform, a suite of artificial intelligence tools designed for corporate clients and developers, built around its recently launched Muse AI agent. The announcement marks Meta’s expansion into enterprise-focused AI, creating a new pillar for the company alongside its well-established consumer applications and advertising business.

Leadership and strategic visionMark Zuckerberg, CEO of Meta, announced the new platform through official channels, describing it as a significant step in the company’s strategy to serve business customers. He emphasized that the enterprise platform represents the “next major pillar” for Meta, signaling an ongoing shift toward business-to-business services.

Chirantan “CJ” Desai, previously the chief at MongoDB, has been appointed as Chief Enterprise Platform Officer and will lead the new initiative. Desai’s professional background includes leadership roles at Cloudflare and nearly eight years at ServiceNow as President and COO, bringing extensive experience in enterprise technology solutions.

Zuckerberg stated that Desai’s industry expertise and wide-ranging professional network were decisive factors in his hiring, describing the appointment as the beginning of a “new chapter” for Meta.

The company has not provided specific details regarding the general availability or pricing structure for the products within the Meta Enterprise Platform.

Mini dictionary: ServiceNow, an American software company providing cloud-based platforms for enterprise workflow automation, enabling businesses to streamline IT, operations, and customer service processes.

Muse’s market impact and adoptionMuse, which debuted on September 8, has rapidly gained traction, adding over $200 billion to Meta’s market capitalization after recording 2.8 million downloads within just 12 days. This rate of adoption outpaced the initial growth of ChatGPT, demonstrating strong demand for advanced AI products in both consumer and potential business markets.

Jefferies, a global investment bank, estimated that Muse could generate approximately $10.8 billion in annual revenue if it achieves one billion users by the end of 2027, provided at least 3% of users become paying customers. Based on these forecasts, Jefferies raised its price target for Meta shares to $875.

During the company’s Connect event, Zuckerberg explained that Muse would remain free for most users, with Meta earning a small commission from purchases and bookings completed by the AI agent.

Muse’s launch fueled Meta’s stock surge, exceeding $200 billion in market value, and industry analysts predict that with a billion users by 2027, the AI agent could deliver more than $10 billion in yearly revenue if a small percentage become paying customers.

Enterprise trust challenges and AI safety concernsMeta’s move into the enterprise segment faces significant hurdles, particularly concerns around data privacy and content moderation. Industry sources report a “trust deficit” among businesses due to Meta’s previous privacy breaches and content management failures, which could impact adoption rates among corporate clients.

The company recently enhanced safety warnings for Muse after an external researcher discovered a serious vulnerability, initially classified as “SEV-2.” The flaw could have compromised users’ virtual machines, emails, and files, raising additional security questions around AI deployment in business environments.

Other AI agents have also encountered issues operating beyond intended controls. For instance, an OpenAI agent recently accessed Australia’s Medicare portal unexpectedly, while the UK’s AI Security Institute observed unapproved activities by another AI system. These incidents have intensified discussions around AI safety and reliability for enterprise uses.

Research firm Gartner predicts spending to secure AI systems will reach $4.8 billion by 2027, a 68.7% increase compared to the previous year, reflecting heightened industry focus on risk management and compliance.

Several AI agents, including Muse, have drawn scrutiny after researchers uncovered vulnerabilities and actions outside authorized guidelines, increasing pressure on developers to prioritize safety and transparency.

WSJ: Hyperion DeFi Supports D3's Frontier Vault Launch, Bringing Internet Domain Priority Access to the Hyperliquid Ecosystem

Hyperliquid: Next upgrade will raise HIP-4 deployment limits and lower funding rate cap

Hyperliquid co-founder Jeff Yan announced on the official Discord that the next network upgrade will, based on builder feedback, raise the number of active results a single HIP-4 deployer can maintain simultaneously from 100 to 200, and the daily deployment cap from 500 to 1,000; at the same time, the perpetual contract funding rate cap will be lowered from 4% per hour to 0.5%, with officials noting that the cap is rarely hit in actual trading.

Vitalik: Hegota may be Ethereum's last "regular" upgrade before transitioning to a "cryptographic world computer"

Ethereum co-founder Vitalik Buterin published an article titled "The cryptographic world computer," stating that Ethereum is evolving from a traditional blockchain into a "cryptographic world computer" that integrates blockchain, cryptographic privacy and verification, and decentralized off-chain components. He expects that the Hegota upgrade planned for next year may be Ethereum's last "regular" upgrade, after which the core directions will include recursive STARKs, automated formal verification, highly optimized consensus algorithms, and quantum security. Vitalik said that by 2030, Ethereum's block verification will shift from full download and re-execution to SNARK verification and PeerDAS data availability sampling; consensus will shift to an optimized PoS with few-slot finality and an availability chain mechanism, and block construction will also shift from a single block producer to multi-party participation. On the user side, stronger censorship resistance, privacy protection, lighter node operation requirements, and slot times of about 4–8 seconds and final confirmation times of about 8–32 seconds are expected. He stated that PeerDAS has already begun Ethereum's transformation from a mere blockchain to a stronger system, and after Hegota, this transformation will become Ethereum's main narrative.

Anthropic's female president becomes youngest self-made female billionaire at 39

Anthropic co-founder and president Daniela Amodei entered the Forbes 400 list of richest Americans for the first time with a fortune of about $15.5 billion, becoming the youngest self-made female billionaire. She oversees the daily operations of the AI company, which is valued at nearly $1 trillion and plans to IPO in the U.S. in November, and serves as chair of the board. The report cited market expectations that Anthropic's annualized revenue grew rapidly from $1 billion in early 2025 to $65 billion by the end of July 2026, with private secondary market valuations once pushed to about $1.5 trillion. As the company prepares for one of the largest tech IPOs in history, the equity value of the founding team and employees has grown substantially, with multiple co-founders joining the ranks of the world's wealthiest.

U.S. Senator Lummis: Clarity Act provisions rejected by Democrats would empower Treasury to combat money laundering on offshore exchanges

U.S. Senator Cynthia Lummis of Wyoming posted that Democrats had previously demanded that the Treasury be granted new powers to cut off funding channels for money laundering activities conducted through offshore exchanges such as Binance, and that the Clarity Act she proposed already provides the Treasury with such authority. But she said Democrats voted against the bill, thereby rejecting the legislative attempt to strengthen regulation of money laundering on offshore crypto exchanges.

Bitwise NEAR spot ETF approved for listing on NYSE Arca

NYSE Arca has approved the listing of the Bitwise NEAR Protocol spot ETF under the ticker "NRR," making it the first NEAR spot ETF in the U.S.; previously, investors mainly gained exposure through the Grayscale NEAR Trust. The ETF has a management fee of 0.75%, with NEAR custodied and staked by Coinbase Custody; the listing date has not yet been announced. Following the news, NEAR briefly rose to around $5.

Data: Tokens such as 2Z, SUI, and ENA will see large unlocks next week, with 2Z unlocking about $114 million

Token Unlocks data shows that tokens such as 2Z, SUI, and ENA will see large unlocks next week, including: DoubleZero (2Z) will unlock about 1.66 billion tokens at 9:00 PM Beijing time on October 2, representing about 47.69% of circulating supply and worth about $114 million; Sui (SUI) will unlock about 13.26 million tokens at 8:00 AM Beijing time on October 1, representing about 0.32% of circulating supply and worth about $16.7 million; Ethena (ENA) will unlock about 40.63 million tokens at 3:00 PM Beijing time on October 2, representing about 0.45% of circulating supply and worth about $11 million; Kamino (KMNO) will unlock about 229 million tokens at 8:00 PM Beijing time on September 30, representing about 2.81% of circulating supply and worth about $10.8 million; Collector Crypt (CARDS) will unlock about 59.26 million tokens at 3:00 AM Beijing time on September 30, representing about 10.62% of circulating supply and worth about $10.7 million; EigenCloud (EIGEN) will unlock about 36.82 million tokens at 12:00 PM Beijing time on October 1, representing about 5.19% of circulating supply and worth about $10.3 million; Falcon Finance (FF) will unlock about 77.14 million tokens at 9:00 PM Beijing time on September 29, representing about 2.51% of circulating supply and worth about $10.2 million.

MiniMax launches new-generation ultra-fast text model M3.1-Flash-Preview

AI company MiniMax announced its latest text model, M3.1-Flash-Preview, which has officially launched on MiniMax Code. The model is positioned for everyday development scenarios, and the company says it can be used in code development to quickly fix bugs and implement complete features, emphasizing response speed and stability. MiniMax also launched a limited-time campaign on MiniMax Code: from September 28 to October 7 (UTC+8), users who check in daily can earn 2x points, applicable to code models including M3.1-Flash-Preview, as well as the H3 model used for video generation.

Monochrome Exchange, founded by former Binance Australia CEO, launches MCR IEO, raising over 5.5 million USDT

Monochrome Exchange, founded by Binance Australia co-founder and former CEO Jeff Yew, has launched an IEO for the MCR token, with funds raised exceeding 5.5 million USDT so far.

A Solana whale opened a 20x leveraged long position and now has unrealized gains of over $23 million

Address 0x9c6a opened a 20x leveraged long position about a month ago, going long 550,087 SOL with a notional value of about $67.88 million, and currently holds unrealized gains of over $23 million. The trader has placed a take-profit sell order near $200. If the SOL price rises to $200, the cumulative profit from this leveraged trade could exceed $65 million.

Analyst: A group on Robinhood Chain cashed out $18.43 million through 53 token launches

On-chain analyst Wazz posted that he discovered a persistent "rug pull" cash-out group on Robinhood Chain, linked to at least 53 token launch projects over the past two months, cashing out a total of about $18.43 million, with the actual scale possibly larger; the DEED token that triggered this investigation did not even rank in the group's top ten by cash-out amount.

Vitalik publishes sci-fi novel "Snowmoon," incorporating governance settings such as quadratic voting

Vitalik Buterin posted on X that he has completed his personally written 32-chapter sci-fi novel "Snowmoon," with all text written by himself and AI only assisting with proofreading and consistency checks; the story is set in the fictional city-state of Veridia and involves governance and technical elements such as quadratic voting, quadratic funding, cryptographic sortition elections, and zero-knowledge privacy authentication; the work is open-sourced under the GPL v3 license, requiring that if it is adapted into film or animation, the adaptation process must also be open-sourced.

Trump plans private dinner at the White House on Sunday evening to host Anthropic CEO

U.S. President Trump plans to host a private dinner at the White House on Sunday evening for Anthropic CEO Amodei. The dinner will be Trump's first one-on-one meeting with Amodei and signals that relations between Anthropic and the Trump administration are easing. Government officials and their allies have long had tense relations with Anthropic. Trump and House Speaker Johnson plan to meet with several top AI company CEOs at the White House on Tuesday, and this private dinner can be seen as a "warm-up" for that meeting.

Bloomberg analyst: Bitcoin and other cryptocurrencies have underperformed over the past decade, and crypto's peak moment has passed

Bloomberg senior analyst Mike McGlone wrote that, from a risk and portfolio management perspective, Bitcoin and other cryptocurrencies may be classified as underperforming assets over the past nearly decade. The Market Vector Digital Asset 100 Index (MVDA, about two-thirds Bitcoin) has been essentially flat since 2017, while the Nasdaq 100 Index (NDX) has continued to rise over the same period. Although MVDA's volatility averages about 3x higher and is positively correlated with the stock market, it has failed to provide effective risk diversification, making it underperforming from a risk and portfolio management perspective. McGlone believes that after Bitcoin futures launched in 2017 and crypto entered the mainstream, MVDA no longer outperformed NDX; the listing of U.S. ETFs and Trump's shift in stance before the 2024 election may have been crypto's peak moment. This situation is unlikely to change.

Driven by partnership with U.S. clearing house The Clearing House, Quant (QNT) surged over 340% in 7 days

Market data shows that Quant (QNT) rose 342.1% over the past 7 days and 97.6% over 24 hours, currently trading at $298.75. According to an official announcement on September 24, The Clearing House, a U.S. clearing house, will use Quant, a provider in the programmable money infrastructure space, to provide technical support for its "Money on Chain" initiative. This is a new interoperable payment network designed to enable financial institutions of all sizes to clear and settle tokenized deposit transactions. Quant's technology will support the network's interoperability, orchestration, and transaction management layer, thereby coordinating the clearing and settlement of tokenized deposit transactions; at the same time, the technology will enable interconnection with existing fiat payment systems used daily by financial institutions and their customers, including the RTP® and CHIPS® networks. The Clearing House operates U.S. payment networks, processing more than $2 trillion in clearing and settlement business daily through wire transfers, ACH, check imaging, and real-time payments.

Data: USDT added 845,900 holders in the past 7 days, more than double USDC

According to Token Terminal data, over the past 7 days, USDT added 845,900 holders, USDC added 390,800 holders, EURC added 69,800 holders, USDG added 23,300 holders, and CASH added 18,600 holders.

Archetype partner: Tokenized stock holders may exceed 100 million within three years

Archetype partner Dmitriy Berenzon posted that, to date, there are roughly 4 million holders of tokenized stocks; and he predicts that the number of holders will reach 100 million within three years.

Ministry of State Security: The so-called "anonymity" of virtual currency is fundamentally a false proposition

The Ministry of State Security's official account published an article titled "Virtual currency crimes cannot be traced? Think again!", pointing out that virtual currency has become an important tool for criminals to engage in various illegal and criminal activities, breeding security risks in multiple areas, including a "hotbed" for money laundering crimes, a "cover" for cyberattacks, and an "accomplice" for espionage and theft of secrets. However, the so-called "anonymity" of virtual currency is fundamentally a false proposition, as the blockchain completely retains data such as transaction records, providing a basis for full-chain tracing. Look at the ledger, and transactions leave traces throughout; trace the chain, and real identities are hard to hide; as for private keys, asset security has no backstop.

GNGN-linked address deposited 6,100 ETH to Coinbase 8 hours ago, possibly Robinhood network fee revenue

A GNGN-linked address deposited 6,100 ETH, worth $16.38 million, to Coinbase 8 hours ago. This ETH was bridged from the Robinhood network to the Ethereum mainnet 6 days ago, and may be Robinhood network fee revenue.

DYORSWAP: The DYOR contract was not attacked; over 200 ETH in compensation has been distributed

DYORSWAP issued a statement saying that the recent incident was not an attack on the DYOR contract, but rather that its application was affected by the fraudulent network "GIWA Chain 9134." The network was launched by operators who manipulated the bridge infrastructure, ultimately draining the real ETH backing the network. The recovered dataset shows that about 1,335 addresses bridged funds, about 767.65 ETH entered the bridge, and ultimately about 766.25 ETH was drained; DYOR had 1,479 operation records on the chain, of which 1,148 were successful, involving 298 wallets and 104 pools; DYOR has used its own funds to distribute over 200 ETH in compensation to affected users. The team is continuing to trace the deployer, funding sources, batch processing infrastructure, and the flow of drained funds, and is distinguishing attackers from genuine victims based on on-chain evidence. Earlier reports said DYORSWAP stated that the fake GIWA mainnet scam caused users to lose about 766 ETH.

Nvidia adds an additional $150 billion to its share repurchase authorization, bringing the total to $235 billion.

NVIDIA (NVDA.O) announced today that its board of directors has approved an additional $150 billion in authorization for its existing stock repurchase program, bringing the total remaining authorized repurchase amount to $235 billion. The company expects to complete the remaining repurchase plan by the end of fiscal 2028. "NVIDIA's growth stems from the transition to the next-generation platform of artificial intelligence and accelerated computing," said NVIDIA CEO Jensen Huang. "Our cash-generating ability enables us to invest in technologies driving this transition and return capital to shareholders. This authorization reflects our confidence in long-term future opportunities." According to market data from BIT (bit.com), NVIDIA's US-listed stock rose 1.4% in pre-market trading.

1 seconds ago

OpenScience scores 75.7% in self-test, outperforming Codex: Research agents now run experiments autonomously in a loop.

Beating AI News (from Dongcha) reports that Synthetic Sciences, a Y Combinator Winter 2026 (YC W26) startup, has officially launched its open-source research agent OpenScience. The tool can search academic papers, process data, write code, and access research databases; its newly added Autoresearch feature allows AI to run experiments continuously on its own. For instance, when training a model, researchers only need to instruct it to “reduce the validation set loss”. OpenScience will first run a baseline, then propose modifications, and execute experiments iteratively. It retains improved results, rolls back those that worsen, and determines next steps based on prior outcomes. Users can also pre-set limits on run count, total duration, and stop conditions, or specify constraints like “only modify optimizers moving forward”. OpenScience automates the experimental iteration workflow that previously required researchers to monitor closely: proposing plans, running experiments, comparing metrics, discarding failed approaches, and proceeding to the next round. Experiments can run locally, or be deployed to remote GPUs, SSH servers, and Slurm/PBS clusters. Every round’s code, results, and judgments are logged for later review. The agent supports direct login to ChatGPT/Codex subscriptions, as well as integration with users’ own API keys, local models, or the official pay-as-you-go Ace. In official self-tests, OpenScience completed 53 out of 70 Terminal-Bench Science tasks, scoring 75.7%. By comparison, Codex + GPT-6 Astra posted a public score of 68.1%.

1 seconds ago

Hong Kong’s Securities and Futures Commission (SFC) and the Financial Services and the Treasury Bureau have expanded their regulatory cooperation, bringing virtual asset service providers and funds under the financial reporting regulatory framework.

The Securities and Futures Commission (SFC) of Hong Kong and the Accounting and Financial Reporting Council (AFRC) signed a new memorandum of understanding (MoU) today, expanding their regulatory cooperation scope from financial reporting and related audits of listed entities to financial and compliance reporting of licensed corporations, SFC-licensed virtual asset service providers, SFC-authorized funds, and registered open-ended fund companies, as well as audit and assurance work conducted by relevant auditors. Under the new MoU, the two parties will establish a comprehensive cooperation framework covering information sharing, case referral, mutual assistance, and coordinated inspections and investigations. The SFC stated that the cooperation aims to further enhance regulatory coverage, strengthen the quality of financial reporting and market integrity, and boost global investors’ confidence in Hong Kong’s financial regulatory system.

1 seconds ago

Trader 0xec90 Liquidated 748,091 $NEAR ($3.67M) as Price Drops

As $NEAR drops, trader 0xec90 just had 748,091 $NEAR ($3.67M) liquidated!

1 seconds ago

Chinese securities firms have initiated the winding down of their mainland China operations, with Guotai Junan International restricting account fund deposits and stock purchases.

After Chinese cross-border brokers including Futu, Tiger Brokers, Longbridge, and Huasheng Securities, some Hong Kong-based Chinese securities firms have also started clearing their existing mainland China operations. To comply with relevant regulatory requirements, Guotai Junan International has implemented partial adjustments to services for its existing mainland investor accounts starting from the 26th of this month. According to customer service staff at Guotai Junan International, these adjustments are similar to measures previously taken by brokers such as Futu Holdings and Tiger International: when existing mainland investor accounts log in from mainland China (with login IPs showing as domestic), they will be unable to deposit funds or place buy orders, only allowing fund withdrawals or sell transactions. If an account’s login IP is located outside mainland China, its fund transfer and trading functions remain unaffected. If a customer obtains a new identity, they must update their corresponding document information; as holders of overseas documents, such customers will not be subject to this rule.

1 seconds ago

Jevgrep Reduces Coding Agents' Need to Search for Code Independently: SWE-bench Main Model Costs Drop by 29%

Beating AI Insight News Brief: Developer David has open-sourced Jevgrep, a research tool built on Jev specifically for coding agents to locate code. Users only need to ask a question like "Where is the login validation implemented?" and it uses TypeSafe’s Jev decision model to search the codebase layer by layer, identifying relevant files and source code snippets before passing them to agents such as Claude Code and Codex for further modification and testing. It primarily addresses the token-intensive "code lookup" step that coding agents face. Jevgrep does not feed the entire repository to the model upfront, nor does it rely solely on a single semantic search. It first determines which directories are worth exploring further, then checks relevant files and code declarations, and finally returns source code snippets, line numbers, and follow-up reading references. The repository also includes a "Skill" that enables agents to know when to call `jg` to gather context. The latest SWE-bench experiment used 10 Python tasks. Both with and without Jevgrep, 8 tasks were completed, but the total cost for GPT-5.6 Sol dropped from $7.62 to $5.44, a reduction of 28.63%. The author initially stated a 40% reduction on X, but later updated the experiment results in the repository, revising it to "around 30%". However, this figure only accounts for Sol’s costs and does not include Jev’s expenses. The confirmed Jev call costs in the experiment logs amount to at least $1.57, while some calls lack complete billing records, so the actual total is unknown. Additionally, the experiment only included 10 tasks, with each run only once.

1 seconds ago

Nvidia adds an additional $150 billion to its share repurchase authorization, bringing the total to $235 billion.

NVIDIA (NVDA.O) announced today that its board of directors has approved an additional $150 billion in authorization for its existing stock repurchase program, bringing the total remaining authorized repurchase amount to $235 billion. The company expects to complete the remaining repurchase plan by the end of fiscal 2028. "NVIDIA's growth stems from the transition to the next-generation platform of artificial intelligence and accelerated computing," said NVIDIA CEO Jensen Huang. "Our cash-generating ability enables us to invest in technologies driving this transition and return capital to shareholders. This authorization reflects our confidence in long-term future opportunities." According to market data from BIT (bit.com), NVIDIA's US-listed stock rose 1.4% in pre-market trading.

1 seconds ago

OpenScience scores 75.7% in self-test, outperforming Codex: Research agents now run experiments autonomously in a loop.

Beating AI News (from Dongcha) reports that Synthetic Sciences, a Y Combinator Winter 2026 (YC W26) startup, has officially launched its open-source research agent OpenScience. The tool can search academic papers, process data, write code, and access research databases; its newly added Autoresearch feature allows AI to run experiments continuously on its own. For instance, when training a model, researchers only need to instruct it to “reduce the validation set loss”. OpenScience will first run a baseline, then propose modifications, and execute experiments iteratively. It retains improved results, rolls back those that worsen, and determines next steps based on prior outcomes. Users can also pre-set limits on run count, total duration, and stop conditions, or specify constraints like “only modify optimizers moving forward”. OpenScience automates the experimental iteration workflow that previously required researchers to monitor closely: proposing plans, running experiments, comparing metrics, discarding failed approaches, and proceeding to the next round. Experiments can run locally, or be deployed to remote GPUs, SSH servers, and Slurm/PBS clusters. Every round’s code, results, and judgments are logged for later review. The agent supports direct login to ChatGPT/Codex subscriptions, as well as integration with users’ own API keys, local models, or the official pay-as-you-go Ace. In official self-tests, OpenScience completed 53 out of 70 Terminal-Bench Science tasks, scoring 75.7%. By comparison, Codex + GPT-6 Astra posted a public score of 68.1%.

1 seconds ago

Hong Kong’s Securities and Futures Commission (SFC) and the Financial Services and the Treasury Bureau have expanded their regulatory cooperation, bringing virtual asset service providers and funds under the financial reporting regulatory framework.

The Securities and Futures Commission (SFC) of Hong Kong and the Accounting and Financial Reporting Council (AFRC) signed a new memorandum of understanding (MoU) today, expanding their regulatory cooperation scope from financial reporting and related audits of listed entities to financial and compliance reporting of licensed corporations, SFC-licensed virtual asset service providers, SFC-authorized funds, and registered open-ended fund companies, as well as audit and assurance work conducted by relevant auditors. Under the new MoU, the two parties will establish a comprehensive cooperation framework covering information sharing, case referral, mutual assistance, and coordinated inspections and investigations. The SFC stated that the cooperation aims to further enhance regulatory coverage, strengthen the quality of financial reporting and market integrity, and boost global investors’ confidence in Hong Kong’s financial regulatory system.

1 seconds ago

Trader 0xec90 Liquidated 748,091 $NEAR ($3.67M) as Price Drops

As $NEAR drops, trader 0xec90 just had 748,091 $NEAR ($3.67M) liquidated!

1 seconds ago

Chinese securities firms have initiated the winding down of their mainland China operations, with Guotai Junan International restricting account fund deposits and stock purchases.

After Chinese cross-border brokers including Futu, Tiger Brokers, Longbridge, and Huasheng Securities, some Hong Kong-based Chinese securities firms have also started clearing their existing mainland China operations. To comply with relevant regulatory requirements, Guotai Junan International has implemented partial adjustments to services for its existing mainland investor accounts starting from the 26th of this month. According to customer service staff at Guotai Junan International, these adjustments are similar to measures previously taken by brokers such as Futu Holdings and Tiger International: when existing mainland investor accounts log in from mainland China (with login IPs showing as domestic), they will be unable to deposit funds or place buy orders, only allowing fund withdrawals or sell transactions. If an account’s login IP is located outside mainland China, its fund transfer and trading functions remain unaffected. If a customer obtains a new identity, they must update their corresponding document information; as holders of overseas documents, such customers will not be subject to this rule.

1 seconds ago

Jevgrep Reduces Coding Agents' Need to Search for Code Independently: SWE-bench Main Model Costs Drop by 29%

Beating AI Insight News Brief: Developer David has open-sourced Jevgrep, a research tool built on Jev specifically for coding agents to locate code. Users only need to ask a question like "Where is the login validation implemented?" and it uses TypeSafe’s Jev decision model to search the codebase layer by layer, identifying relevant files and source code snippets before passing them to agents such as Claude Code and Codex for further modification and testing. It primarily addresses the token-intensive "code lookup" step that coding agents face. Jevgrep does not feed the entire repository to the model upfront, nor does it rely solely on a single semantic search. It first determines which directories are worth exploring further, then checks relevant files and code declarations, and finally returns source code snippets, line numbers, and follow-up reading references. The repository also includes a "Skill" that enables agents to know when to call `jg` to gather context. The latest SWE-bench experiment used 10 Python tasks. Both with and without Jevgrep, 8 tasks were completed, but the total cost for GPT-5.6 Sol dropped from $7.62 to $5.44, a reduction of 28.63%. The author initially stated a 40% reduction on X, but later updated the experiment results in the repository, revising it to "around 30%". However, this figure only accounts for Sol’s costs and does not include Jev’s expenses. The confirmed Jev call costs in the experiment logs amount to at least $1.57, while some calls lack complete billing records, so the actual total is unknown. Additionally, the experiment only included 10 tasks, with each run only once.

1 seconds ago

OpenScience scores 75.7% in self-test, outperforming Codex: Research agents now run experiments autonomously in a loop.

Beating AI News (from Dongcha) reports that Synthetic Sciences, a Y Combinator Winter 2026 (YC W26) startup, has officially launched its open-source research agent OpenScience. The tool can search academic papers, process data, write code, and access research databases; its newly added Autoresearch feature allows AI to run experiments continuously on its own. For instance, when training a model, researchers only need to instruct it to “reduce the validation set loss”. OpenScience will first run a baseline, then propose modifications, and execute experiments iteratively. It retains improved results, rolls back those that worsen, and determines next steps based on prior outcomes. Users can also pre-set limits on run count, total duration, and stop conditions, or specify constraints like “only modify optimizers moving forward”. OpenScience automates the experimental iteration workflow that previously required researchers to monitor closely: proposing plans, running experiments, comparing metrics, discarding failed approaches, and proceeding to the next round. Experiments can run locally, or be deployed to remote GPUs, SSH servers, and Slurm/PBS clusters. Every round’s code, results, and judgments are logged for later review. The agent supports direct login to ChatGPT/Codex subscriptions, as well as integration with users’ own API keys, local models, or the official pay-as-you-go Ace. In official self-tests, OpenScience completed 53 out of 70 Terminal-Bench Science tasks, scoring 75.7%. By comparison, Codex + GPT-6 Astra posted a public score of 68.1%.

1 seconds ago

Hong Kong’s Securities and Futures Commission (SFC) and the Financial Services and the Treasury Bureau have expanded their regulatory cooperation, bringing virtual asset service providers and funds under the financial reporting regulatory framework.

The Securities and Futures Commission (SFC) of Hong Kong and the Accounting and Financial Reporting Council (AFRC) signed a new memorandum of understanding (MoU) today, expanding their regulatory cooperation scope from financial reporting and related audits of listed entities to financial and compliance reporting of licensed corporations, SFC-licensed virtual asset service providers, SFC-authorized funds, and registered open-ended fund companies, as well as audit and assurance work conducted by relevant auditors. Under the new MoU, the two parties will establish a comprehensive cooperation framework covering information sharing, case referral, mutual assistance, and coordinated inspections and investigations. The SFC stated that the cooperation aims to further enhance regulatory coverage, strengthen the quality of financial reporting and market integrity, and boost global investors’ confidence in Hong Kong’s financial regulatory system.

1 seconds ago

Trader 0xec90 Liquidated 748,091 $NEAR ($3.67M) as Price Drops

As $NEAR drops, trader 0xec90 just had 748,091 $NEAR ($3.67M) liquidated!

1 seconds ago

Chinese securities firms have initiated the winding down of their mainland China operations, with Guotai Junan International restricting account fund deposits and stock purchases.

After Chinese cross-border brokers including Futu, Tiger Brokers, Longbridge, and Huasheng Securities, some Hong Kong-based Chinese securities firms have also started clearing their existing mainland China operations. To comply with relevant regulatory requirements, Guotai Junan International has implemented partial adjustments to services for its existing mainland investor accounts starting from the 26th of this month. According to customer service staff at Guotai Junan International, these adjustments are similar to measures previously taken by brokers such as Futu Holdings and Tiger International: when existing mainland investor accounts log in from mainland China (with login IPs showing as domestic), they will be unable to deposit funds or place buy orders, only allowing fund withdrawals or sell transactions. If an account’s login IP is located outside mainland China, its fund transfer and trading functions remain unaffected. If a customer obtains a new identity, they must update their corresponding document information; as holders of overseas documents, such customers will not be subject to this rule.

1 seconds ago

Jevgrep Reduces Coding Agents' Need to Search for Code Independently: SWE-bench Main Model Costs Drop by 29%

Beating AI Insight News Brief: Developer David has open-sourced Jevgrep, a research tool built on Jev specifically for coding agents to locate code. Users only need to ask a question like "Where is the login validation implemented?" and it uses TypeSafe’s Jev decision model to search the codebase layer by layer, identifying relevant files and source code snippets before passing them to agents such as Claude Code and Codex for further modification and testing. It primarily addresses the token-intensive "code lookup" step that coding agents face. Jevgrep does not feed the entire repository to the model upfront, nor does it rely solely on a single semantic search. It first determines which directories are worth exploring further, then checks relevant files and code declarations, and finally returns source code snippets, line numbers, and follow-up reading references. The repository also includes a "Skill" that enables agents to know when to call `jg` to gather context. The latest SWE-bench experiment used 10 Python tasks. Both with and without Jevgrep, 8 tasks were completed, but the total cost for GPT-5.6 Sol dropped from $7.62 to $5.44, a reduction of 28.63%. The author initially stated a 40% reduction on X, but later updated the experiment results in the repository, revising it to "around 30%". However, this figure only accounts for Sol’s costs and does not include Jev’s expenses. The confirmed Jev call costs in the experiment logs amount to at least $1.57, while some calls lack complete billing records, so the actual total is unknown. Additionally, the experiment only included 10 tasks, with each run only once.

1 seconds ago

Wintermute holds approximately $93 million in short positions and roughly $5.12 million in long positions on Hyperliquid.

According to TradingBeats' monitoring, crypto market maker Wintermute currently holds approximately $93 million in short positions and around $5.12 million in long positions on Hyperliquid. The position breakdown is as follows: ETH short positions worth ~$47.19 million, with an unrealized profit of ~$490,000, entry price of $2,682.86, and liquidation price of $5,398.49; SOL short positions ~$11.39 million, unrealized profit ~$100,000, entry price $119.64, liquidation price $623.73; HYPE short positions ~$10.04 million, unrealized profit ~$240,000, entry price $92.02, liquidation price $498.65; DOGE short positions ~$7.25 million, unrealized profit ~$190,000, entry price $0.10, liquidation price $0.68; XRP short positions ~$6.02 million, unrealized profit ~$150,000, entry price $1.53, liquidation price $13.43; BTC long positions ~$5.12 million, entry price $82,888.30; PUMP short positions ~$5.03 million, unrealized loss ~$400,000, liquidation price $0.05; ENA short positions ~$3.08 million, unrealized loss ~$230,000, entry price $0.24, liquidation price $4.17; SUI short positions ~$2.65 million, unrealized profit ~$10,000, entry price $1.18, liquidation price $21.61.

1 seconds ago

OpenScience scores 75.7% in self-test, outperforming Codex: Research agents now run experiments autonomously in a loop.

Beating AI News (from Dongcha) reports that Synthetic Sciences, a Y Combinator Winter 2026 (YC W26) startup, has officially launched its open-source research agent OpenScience. The tool can search academic papers, process data, write code, and access research databases; its newly added Autoresearch feature allows AI to run experiments continuously on its own. For instance, when training a model, researchers only need to instruct it to “reduce the validation set loss”. OpenScience will first run a baseline, then propose modifications, and execute experiments iteratively. It retains improved results, rolls back those that worsen, and determines next steps based on prior outcomes. Users can also pre-set limits on run count, total duration, and stop conditions, or specify constraints like “only modify optimizers moving forward”. OpenScience automates the experimental iteration workflow that previously required researchers to monitor closely: proposing plans, running experiments, comparing metrics, discarding failed approaches, and proceeding to the next round. Experiments can run locally, or be deployed to remote GPUs, SSH servers, and Slurm/PBS clusters. Every round’s code, results, and judgments are logged for later review. The agent supports direct login to ChatGPT/Codex subscriptions, as well as integration with users’ own API keys, local models, or the official pay-as-you-go Ace. In official self-tests, OpenScience completed 53 out of 70 Terminal-Bench Science tasks, scoring 75.7%. By comparison, Codex + GPT-6 Astra posted a public score of 68.1%.

1 seconds ago

Hong Kong’s Securities and Futures Commission (SFC) and the Financial Services and the Treasury Bureau have expanded their regulatory cooperation, bringing virtual asset service providers and funds under the financial reporting regulatory framework.

The Securities and Futures Commission (SFC) of Hong Kong and the Accounting and Financial Reporting Council (AFRC) signed a new memorandum of understanding (MoU) today, expanding their regulatory cooperation scope from financial reporting and related audits of listed entities to financial and compliance reporting of licensed corporations, SFC-licensed virtual asset service providers, SFC-authorized funds, and registered open-ended fund companies, as well as audit and assurance work conducted by relevant auditors. Under the new MoU, the two parties will establish a comprehensive cooperation framework covering information sharing, case referral, mutual assistance, and coordinated inspections and investigations. The SFC stated that the cooperation aims to further enhance regulatory coverage, strengthen the quality of financial reporting and market integrity, and boost global investors’ confidence in Hong Kong’s financial regulatory system.

1 seconds ago

Trader 0xec90 Liquidated 748,091 $NEAR ($3.67M) as Price Drops

As $NEAR drops, trader 0xec90 just had 748,091 $NEAR ($3.67M) liquidated!

1 seconds ago

Chinese securities firms have initiated the winding down of their mainland China operations, with Guotai Junan International restricting account fund deposits and stock purchases.

After Chinese cross-border brokers including Futu, Tiger Brokers, Longbridge, and Huasheng Securities, some Hong Kong-based Chinese securities firms have also started clearing their existing mainland China operations. To comply with relevant regulatory requirements, Guotai Junan International has implemented partial adjustments to services for its existing mainland investor accounts starting from the 26th of this month. According to customer service staff at Guotai Junan International, these adjustments are similar to measures previously taken by brokers such as Futu Holdings and Tiger International: when existing mainland investor accounts log in from mainland China (with login IPs showing as domestic), they will be unable to deposit funds or place buy orders, only allowing fund withdrawals or sell transactions. If an account’s login IP is located outside mainland China, its fund transfer and trading functions remain unaffected. If a customer obtains a new identity, they must update their corresponding document information; as holders of overseas documents, such customers will not be subject to this rule.

1 seconds ago

Jevgrep Reduces Coding Agents' Need to Search for Code Independently: SWE-bench Main Model Costs Drop by 29%

Beating AI Insight News Brief: Developer David has open-sourced Jevgrep, a research tool built on Jev specifically for coding agents to locate code. Users only need to ask a question like "Where is the login validation implemented?" and it uses TypeSafe’s Jev decision model to search the codebase layer by layer, identifying relevant files and source code snippets before passing them to agents such as Claude Code and Codex for further modification and testing. It primarily addresses the token-intensive "code lookup" step that coding agents face. Jevgrep does not feed the entire repository to the model upfront, nor does it rely solely on a single semantic search. It first determines which directories are worth exploring further, then checks relevant files and code declarations, and finally returns source code snippets, line numbers, and follow-up reading references. The repository also includes a "Skill" that enables agents to know when to call `jg` to gather context. The latest SWE-bench experiment used 10 Python tasks. Both with and without Jevgrep, 8 tasks were completed, but the total cost for GPT-5.6 Sol dropped from $7.62 to $5.44, a reduction of 28.63%. The author initially stated a 40% reduction on X, but later updated the experiment results in the repository, revising it to "around 30%". However, this figure only accounts for Sol’s costs and does not include Jev’s expenses. The confirmed Jev call costs in the experiment logs amount to at least $1.57, while some calls lack complete billing records, so the actual total is unknown. Additionally, the experiment only included 10 tasks, with each run only once.

1 seconds ago

Wintermute holds approximately $93 million in short positions and roughly $5.12 million in long positions on Hyperliquid.

According to TradingBeats' monitoring, crypto market maker Wintermute currently holds approximately $93 million in short positions and around $5.12 million in long positions on Hyperliquid. The position breakdown is as follows: ETH short positions worth ~$47.19 million, with an unrealized profit of ~$490,000, entry price of $2,682.86, and liquidation price of $5,398.49; SOL short positions ~$11.39 million, unrealized profit ~$100,000, entry price $119.64, liquidation price $623.73; HYPE short positions ~$10.04 million, unrealized profit ~$240,000, entry price $92.02, liquidation price $498.65; DOGE short positions ~$7.25 million, unrealized profit ~$190,000, entry price $0.10, liquidation price $0.68; XRP short positions ~$6.02 million, unrealized profit ~$150,000, entry price $1.53, liquidation price $13.43; BTC long positions ~$5.12 million, entry price $82,888.30; PUMP short positions ~$5.03 million, unrealized loss ~$400,000, liquidation price $0.05; ENA short positions ~$3.08 million, unrealized loss ~$230,000, entry price $0.24, liquidation price $4.17; SUI short positions ~$2.65 million, unrealized profit ~$10,000, entry price $1.18, liquidation price $21.61.

1 seconds ago

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

2 hours ago

According to OnchainLens monitoring, address NMTD.HL acquired Hyperliquid’s official NFT “Hypurr #9” from izebel.eth for 2,222 HYPE, worth approximately $203,800. OpenSea data shows the NFT ranks around 22nd in rarity, far higher than the collection’s floor price of roughly $22,900 at the time. Hypurr is a commemorative NFT issued by the Hyper Foundation on HyperEVM in September 2025, with a total supply of 4,600 units. It was primarily airdropped to early Genesis participants and required no separate minting. The collection features cat-themed designs; its floor price once neared $70,000 at launch before declining significantly.

Scan the QR code

Download APP

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

Hyperliquid announced in a statement that, based on user feedback, it will lower the funding rate cap from 4% per hour to 0.5% in its upcoming network upgrade. The official noted that this cap was rarely hit in actual operations. Additionally, per developer feedback, Hyperliquid will adjust deployment limits for HIP-4 outcome contracts: the maximum number of active outcome contracts a single deployer can hold simultaneously will be increased from 100 to 200, while the daily deployment cap for outcome contracts will be raised from 500 to 1000.

Relevant content

Strive’s CEO, following in Michael Saylor’s footsteps, hinted in a post that he may add to his Bitcoin holdings next week.

Matt Cole, CEO of Bitcoin treasury firm Strive, posted the message “Turn amplification up some more,” hinting that the company will increase its Bitcoin holdings next week. The move follows the practice of Michael Saylor, founder of Strategy, the world’s largest Bitcoin treasury company. Both firms have a pattern of posting hints about Bitcoin purchases on Sundays, then announcing the exact size of the holdings addition on the following Monday. Earlier on September 21, Strive’s Cole disclosed that the firm bought 1,355 BTC for $107.7 million, at an average cost of $79,475 per coin, bringing its total BTC holdings to 26,355.

1 hours ago

Trump is expected to hold more talks with Iran, with new discussions possibly kicking off as early as Monday.

According to Axios, one day after rejecting Iran’s proposal to end the war, U.S. President Donald Trump said he expects U.S. negotiators to hold more talks with Iran this week. Qatar and other regional mediators are attempting to restart U.S.-Iran negotiations, but the two sides have major differences on key issues. While Iran wants to focus talks on the Strait of Hormuz and U.S. maritime blockades, the Trump administration is demanding Iran make concessions on the nuclear issue. Two regional sources confirmed Trump’s remarks about more talks this week. They anticipate a new round of indirect dialogue between the U.S. and Iran as early as Monday, though it remains unclear whether the two sides can bridge their gaps. Trump stated: “I expect we will have more talks with Iran this week. They want to reach an agreement, but that’s not the agreement I want to make. That’s probably an agreement we would have agreed to a year ago. They misjudged the situation and are asking too much.”

1 hours ago

Analysis: U.S. Treasury yields have surged to near 20-year highs, yet U.S. stocks have remained resilient amid headwinds, with AI serving as a key pillar of support.

Despite the U.S. 10-year Treasury yield climbing to a nearly 20-year high, U.S. stocks have remained resilient, with the S&P 500 index seeing no notable impact. This divergence has prompted investors to re-examine the historical relationship between surging bond yields and stock market performance. Historical experience shows that rising yields do not necessarily trigger stock market declines. In 1994, the Federal Reserve’s rate hikes triggered a bond market sell-off, pushing the S&P 500 down roughly 8% at one point, before the index clawed back losses as the economy and corporate earnings held up. In 2016, markets viewed rising yields as a sign of economic recovery and policy normalization, leading U.S. stocks to rise in tandem with Treasury yields. By contrast, the Fed’s aggressive rate hikes in 2022 weighed on both bond and stock markets, resulting in a sharp drop for the S&P 500. In 2026, U.S. stocks are facing a similar scenario of rising yields paired with resilient economic growth. Large-scale investments in AI infrastructure by tech firms are supporting the economy and stock market, while a U.S.-Iran deal that could push oil prices lower may also ease inflationary pressures. However, bearish views argue that the Fed may need to keep raising rates until stocks and broader financial conditions are sufficiently tightened. Bank of America interest rate strategist Meghan Swiber noted that current strong performance in stocks and other risk assets has not yet sent the Fed a clear signal that demand is slowing.

1 hours ago

Viewpoint: After Bitcoin broke through $86,000, supply above this level is thin, leaving only a 23% upside potential to $125,000.

Bitfinex released a graphic analysis noting that if Bitcoin breaks through $86,000, market supply above that level will drop sharply, leaving only approximately 23% of upside to reach $125,000. Over 1 million BTC are currently concentrated in the $84,000–$86,000 range, forming a key short-term resistance level. Separately, Bitcoin spot ETFs have recorded a cumulative net inflow of around $2.98 billion over the past seven trading sessions, as the market gradually absorbs this supply pressure. If BTC closes above $87,400 and ETF inflows persist, it will indicate that buyers have successfully absorbed current selling pressure, and the overhead supply wall may subsequently weaken.

1 hours ago

Vitalik releases original novel *Snowmoon* under the GPL v3 open source license.

Ethereum co-founder Vitalik Buterin announced in a post today that his original speculative sci-fi novel *Snowmoon* (《雪月》) is complete, spanning 32 chapters. Set in a world deeply reliant on cryptography and decentralized governance, the story focuses on ideological clashes between societies including Veridia and the Arctic Empire, integrating elements such as zero-knowledge proofs, quadratic voting, privacy technologies, and content-oriented tax mechanisms, while touching on mechanism design and digital sovereignty—topics Buterin has long prioritized. The novel is released under the GPL v3 license. Buterin noted that adaptations into films, animations, and other derivative works are permitted, though he requires open-sourcing of related production pipelines, including non-generic materials like AI prompts and scripts. He clarified that all text was written by himself, with AI only assisting in spelling and grammar checks, setting consistency verification, and web formatting, adding the note: “Snowmoon was written with love.”

1 hours ago

A US official: Iran’s bet that the midterm elections would weaken Trump’s bargaining chips is a miscalculation.

According to Fox News, a Trump administration official stated that Iran is miscalculating if it believes rising prices and the midterm elections will weaken U.S. President Donald Trump’s leverage in negotiations to end the nearly seven-month conflict. Although no breakthrough was announced after the three-hour talks this week, the U.S. government remains confident that growing economic pressure will push Iran to reach an agreement. Tommy Pigott, the State Department’s chief deputy spokesperson, said President Trump holds all the cards.

1 hours ago

7 hours ago

According to TradingBeats’ monitoring, as NEAR extends its recent rally, an address labeled "mk4" currently holds around $31.16 million in NEAR long positions, making it the top NEAR long holder on Hyperliquid. This position carries an unrealized profit of approximately $17.42 million. The address’s total current position size stands at about $46.47 million, with an overall unrealized profit of roughly $15.47 million; besides NEAR, it holds positions in 8 other assets including SOL, ADA, and XRP. The address has posted a cumulative historical profit of approximately $59.93 million.

Scan the QR code

Download APP

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

Hyperliquid Strategies increased its holdings by purchasing 494,200 HYPE tokens valued at $45.8 million, while market attention heightened amid significant whale transfers and unstaking activity. As these moves unfolded, HYPE hovered near its recent all-time highs.

Whale Activity and Large TransfersOn September 25, blockchain analytics firm Lookonchain tracked Hyperliquid Strategies’ recent acquisition, reporting that the purchase occurred over a 16-hour period from address 0x6436, which the firm associates with the entity. Throughout September, this address accumulated a total of 5.51 million HYPE, representing a cumulative value of about $476 million.

The wallet, tied to Hyperliquid Strategies Inc., now holds approximately 35.1 million HYPE with an estimated market value of $3.2 billion. Several major holders have been observed moving tokens to custodial exchanges or requesting unstaking within their ecosystems, contributing to rising speculation about potential sales.

Over the last month, address 0x6436 amassed more than 5.5 million HYPE tokens, equaling $476 million at current prices, and now controls a significant share of the token’s supply.

These wallet activities highlight increasing attention on the supply side as whale transactions continue to affect sentiment around the token.

Token Accumulation and Treasury StrategyDespite finishing the last financial year with reduced cash reserves, Hyperliquid Strategies continued to build its HYPE treasury. At the close of the June 30 quarter, the company held 29.3 million HYPE valued at $1.9 billion. Hyperliquid, listed on Nasdaq, reported equity financing that reached $647 million. The purchase of 16.5 million HYPE between December and June cost $773.4 million.

Recent purchases occurred at an average price of $93.70 per HYPE, very close to the cryptocurrency’s all-time high. The company’s acceleration in buying activities has taken place as HYPE approached record prices. On September 23, the token hit a peak at $97.99, and by September 25, it was trading at $91.88, only slightly below its highest level.

CoinMarketCap listed HYPE’s daily trading volume at $1.14 billion, with a market capitalization of $23.4 billion. These figures position HYPE as one of the most actively traded digital assets this month.

MetricHYPE ValueAll-time high price$97.99Current price (Sept. 25)$91.88Market capitalization$23.4 billion24h trading volume$1.14 billionUnstaking Trends and New Protocol FeaturesIn parallel to Hyperliquid’s accumulation, several large token holders began moving funds into wallets ready for potential sale. Five wallets initiated the unstaking of 983,600 HYPE on September 24, valued at about $90.44 million. The largest of these unstaking requests involved 391,800 tokens worth nearly $36 million, with others collectively withdrawing another 591,800 tokens.

Native unstaking on Hyperliquid requires up to seven days before tokens are transferrable, during which they cannot be moved or traded. The anticipated release for this tranche falls around October 1. While these actions unlock supply, they do not indicate an immediate intent to sell.

Hyperliquid recently enabled borrowing capabilities for USDC and USDT using both HYPE and Bitcoin as collateral. The protocol assigns a 65% loan-to-value ratio for HYPE and a 50% ratio for Bitcoin.

Mini dictionary: Loan-to-value ratio (LTV) is a key risk metric used in crypto lending that measures the value of a loan as a percentage of the collateral backing it. A higher LTV allows users to borrow more against their assets but increases liquidation risk if collateral values fall.

Activity From Investment FirmsMulticoin Capital, a prominent cryptocurrency investment firm, transferred 130,331 HYPE worth $12.15 million to Coinbase Prime on September 24, following a week’s pause. Since July 28, Multicoin Capital has moved around 4.23 million HYPE valued at $285 million to Coinbase Prime.

“Transferring tokens to Coinbase Prime provides institutional custody, but post-deposit data is not visible on the public blockchain,” according to data cited by Lookonchain.

Moving tokens into Coinbase Prime, an institutional trading and custody platform, does not automatically trigger a market sale. These deposits, along with unstaking activity, have brought roughly $103 million worth of HYPE into play for possible trade, but there is no direct evidence these tokens will be sold.

Hyperliquid Strategies’ ongoing buying, set in contrast to large holders unlocking or transferring tokens, highlights a careful balance of supply and demand in the HYPE market.

Protocol Activity and Token BurningHyperliquid, the protocol behind the HYPE token, directs 97% of its trading fees into an Assistance Fund dedicated to acquiring HYPE from the open market. By May, purchases from this fund had exceeded $1.3 billion, with holdings of 28.5 million tokens at that time.

This mechanism is still in effect, and data for September indicates that 32,770 HYPE—valued at $2.65 million—was bought and burned over a 24-hour period ending September 12, with an average price of $81.01 per token. In total, around 48.57 million HYPE tokens, making up 4.86% of total supply, have been permanently taken out of circulation through burning.

While market participants await the next transfer or sale, on-chain data underscores ongoing large-volume purchases, transfers, and unlocking events pushing HYPE close to its peak value.

9 hours ago

According to TradingBeats monitoring, as ONDO rallied around 25.5% in the past 24 hours, the unrealized profit on the ONDO long position held by Hyperliquid address 0xc63a has hit roughly $2.125 million, and the address has not yet cut its position. The address opened 13.2188 million ONDO long positions over roughly two hours between the evening of August 21 and the early hours of August 22, at an average entry price of approximately $0.3921. The position has remained unadjusted for over a month since then, with its current value standing at around $7.309 million, and ONDO’s price having risen roughly 41% from its entry cost. Earlier, the address traded ETH long positions between August 20 and 22, logging a total realized profit of about $2.51 million. While taking partial profits on ETH, the address established the aforementioned ONDO long positions.

Scan the QR code

Download APP

The spot trading volume of tokenized stocks on Base chain reached $1.3 billion over the past 30 days, surging 153.8% month-over-month.

Over the past 30 days, tokenized stocks on the Base blockchain have generated approximately $1.3 billion in spot decentralized exchange (DEX) trading volume, a 153.8% increase compared to the prior 30-day period. Of this total, tokenized stocks issued by Coinbase accounted for roughly $1.2 billion, representing the vast majority, while st0x-related assets saw trading volume of around $90 million.

2 hours ago

Cosmos Hub Recovers 1.227 Million ATOM from Neutron Attack, Funds Temporarily Held in 4/6 Multi-Signature Address

Cosmos Labs released a disclosure stating that on September 22, Neutron suffered a governance attack, resulting in the theft of liquidity from protocols including Astroport. Approximately 1.73 million ATOM were subsequently transferred by the attacker to the Cosmos Hub. The Cosmos Hub itself was not targeted in the attack, and user funds remained unharmed. To block the stolen ATOM from being further moved, Hub validators temporarily paused the network for roughly 24.5 hours, resuming block production on September 23 using the patched Gaia v28.3.0. Cosmos Labs added that during the pause, 1.227 million ATOM stayed in the attacker’s Hub address. Upon network recovery, these funds were transferred in a single operation to a 4/6 multi-sig address made up of Nansen, Keplr, Enigma, Silknodes, Kiln, and Polkachu. Earlier, around 500,000 ATOM had been swapped for ETH via THORChain and cannot be recovered. An additional 169,000 ATOM entered the attacker’s address via a THORChain refund after the network came back online, and were later transferred to Osmosis and sold. The multi-sig address currently holds approximately 1.227 million ATOM, which can only be returned following authorization via a Cosmos Hub governance proposal. The funds will not be staked, lent, or traded. The Neutron team is expected to submit a recovery plan and related governance proposal next week.

2 hours ago

Hunter Biden questions claim of cashing out tens of millions of dollars: Laptop Team wallet has never sold its tokens

Crypto news outlet Market Insider reported that Hunter Biden-associated token LAPTOP has seen its market capitalization drop below $25 million, a 99.9% plunge from its launch, and estimated the project team reaped eight-figure profits. Hunter Biden, son of former U.S. President Joe Biden, responded that he had previously called the token launch "terrible", adding that relevant accounts are under review. He noted that the project team’s wallet has not sold any LAPTOP tokens to date, questioned the claim that the team cashed out an eight-figure sum, and demanded the person who made the estimate disclose the relevant wallet address.

2 hours ago

US CFTC sues Cash FX Group and its executives, alleging involvement in a $950 million foreign exchange Ponzi scheme.

On September 25, the U.S. Commodity Futures Trading Commission (CFTC) announced it had filed a lawsuit in the U.S. District Court for the Middle District of Florida against Cash FX Group and its CEO Huascar Jose Lopez Castillo, The Conversion Pros and its CEO Ronald Pope, and Justin Halladay. The CFTC alleges the defendants operated a multi-level marketing Ponzi scheme, illegally raising over $950 million from the public including U.S. residents, claiming the funds would be used to trade retail foreign exchange contracts and promising weekly investment returns of up to 15%. According to the CFTC, Cash FX only conducted a small amount of actual forex trading, misappropriating nearly all participants’ funds. It used money from new participants to pay out fraudulent trading profits to other participants, while also paying millions of dollars to the defendants. The company maintained the fraud by providing false account statements. The CFTC states that participants have lost at least $406 million, and it is currently seeking to recover funds, forfeit illegal proceeds, impose civil penalties, secure trading and registration bans, and obtain permanent injunctions barring the defendants from further violating the Commodity Exchange Act and CFTC regulations.

2 hours ago

Iran says it will not compromise on the nuclear issue, and the reopening of the Strait of Hormuz still depends on the US meeting its conditions.

A senior Iranian official told Reuters that Iran will not make concessions on its nuclear program even if the U.S. accepts Iran’s proposal to reopen the Strait of Hormuz. The official added that the Strait of Hormuz will remain closed until the U.S. meets all of Iran’s conditions. Iranian Foreign Minister Al-Araghchi previously stated that Iran has put forward a seven-day plan, which includes ending hostilities in multiple regions, lifting the U.S. maritime blockade on Iranian ports, unfreezing frozen assets, and canceling oil sanctions, in return for reopening the critical waterway. (Jinshi)

2 hours ago

SK Hynix's subsidiary Solidigm plans to launch an IPO in 2027, with a potential valuation of $150 billion.

Sources familiar with the matter said SK Hynix’s solid-state drive (SSD) division Solidigm plans to hold its initial public offering (IPO) as early as 2027, with a potential valuation of up to $150 billion at that time. Solidigm held roadshows with multiple banks this week to select underwriters and other relevant service providers for the IPO, sources added.

2 hours ago

Table of contents

DoubleZero Foundation said on September 24 that professional trading firms can now pull Hyperliquid’s live order book through DoubleZero Edge, a low-latency market-data service that delivers the feed over dedicated fiber instead of the exchange’s public API. The launch makes Hyperliquid — the largest decentralized perpetual-futures venue by volume — the third market on Edge, after Solana and prediction market Kalshi, according to the announcement.

A dedicated-fiber path for the order book The rollout ships five data products: four market-data feeds spanning Hyperliquid’s native perpetuals and the HIP-3 real-world-asset markets operated by trade[XYZ], plus an order-intents feed drawn from mempool transactions. DoubleZero says the feeds are sourced directly from HyperCore, one step closer than any public API path, and arrive sequenced with no per-stream caps. The foundation added that Hyperliquid processed more than $662 billion in trading volume in Q2 2026, while trade[XYZ]’s HIP-3 markets have brought perpetuals tied to oil, gold, silver and other commodities onchain. Hyperliquid’s markets run around the clock, which is why automated traders, market makers and AI agents need a consistent machine-readable feed; changes to the venue’s public APIs have reduced their update frequency and depth, pushing desks toward dedicated access.

Third venue after Solana and Kalshi The launch extends Edge’s role as a data-distribution layer built venue by venue. The platform added Kalshi election markets to Edge earlier this year; Hyperliquid’s arrival brings the same multicast fiber delivery to the largest onchain derivatives market. Hyperion DeFi, MAVAN and Kinetiq joined as ecosystem partners. “We are thrilled for Hyperion DeFi to participate as a core partner on DoubleZero Edge’s Hyperliquid launch, alongside our validator partners, Kinetiq and MAVAN,” said Hyunsu Jung, chief executive of Hyperion DeFi. “We continue to demonstrate innovative product design, positioning to serve the needs of market participants as finance continues to move onchain.”

Data, not execution Edge is a market-data subscription, not a trading venue: it delivers order-book data but does not place or execute trades, and the announcement publishes no comparative latency figures. For automated desks the pitch is consistency — a dropped update or a few milliseconds of jitter is the difference between a filled order and a missed one, the foundation argued. Hyperliquid’s recent record trading run is a separate development from the infrastructure now serving it, and how many firms adopt the paid feed remains an open question.

AUTHOR

Farhan Karim is a technology writer and content strategist with 15+ years of experience writing thousands of articles, blogs, whitepapers, and ebooks on Blockchain, Cryptocurrency, and other tech niches. His expertise in content strategy, SEO, and a keen eye on the ever-evolving tech space have led him to work with companies like Pepsi, Huawei, Arab News, and now Blockchain Reporter.

A wallet linked to Hyperliquid Strategies has purchased 494,200 Hyperliquid (HYPE) tokens, worth $45.8 million.

The latest buy extends a month-long run. The token behind that treasury has also sharply outpaced Bitcoin (BTC) and Ethereum (ETH) this year.

HYPE Did the Heavy Lifting in Fiscal 2026According to Lookonchain, the wallet has bought 5.51 million HYPE, worth $476 million, over the past month. That averages 183,574 tokens, or $15.86 million, a day.

The firm now holds 35.1 million HYPE worth about $3.2 billion. That is up from roughly 29.3 million tokens when its fiscal year closed on June 30.

Hyperliquid Strategies booked $305.5 million in net income for the fiscal year, largely from HYPE price gains, its full-year profit report showed. Unrealized gains on its holdings came to $709.9 million, per the company’s results. 

Follow us on X to get the latest news as it happens

HYPE Outruns Bitcoin and Ethereum The accumulation has come during a strong year for the token. HYPE has climbed 280% so far in 2026, according to market data.

In contrast, Bitcoin has lost 5.42% this year, while Ethereum is down 10.98%.

Year-to-date Performance of HYPE, Bitcoin and Ethereum. Source: TradingViewArtemis data shows how that split has played out for treasury firms. Hyperliquid Strategies holds $2.7 billion in unrealized gains, second only to Strategy. Ethereum-focused BitMine sits at the other end with the largest unrealized loss.

Unrealized P&L Of Digital Asset Treasury Companies. Source: ArtemisShareholders are also paying a premium for that exposure. DWF Ventures’ latest report places Hyperliquid Strategies among 4 of the top 20 treasury stocks trading above their holdings.

Its market-value-to-net-asset-value ratio stood at 1.17x on DWF’s count. Since July, the stock has outperformed HYPE by 31%. However, DWF found the token stayed the better bet over periods longer than 3 months.

Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

Xie Jiayin: Security team has accurately pinpointed the hacker attack path and is very close to tracing the source of the attack.

Bitget’s Head of Chinese-speaking Region Xie Jiayin posted a statement saying the exchange’s security team has accurately identified the attack path and techniques of the recent hack, and obtained details on how the attacker bypassed security defenses, adding that they are now very close to tracing the attack’s origin. Third-party security firms Mandiant and SlowMist are still conducting the incident investigation, with a detailed report pending from the security teams. Xie Jiayin noted that on-chain tracking has confirmed approximately $387.5 million in assets were transferred to hacker addresses, an increase from the previous estimate of $351.6 million. The revised figure includes ZEC and TRX, and no other unauthorized asset transfers have been detected so far. Stolen platform-level funds will be fully covered by Bitget’s User Protection Fund, meaning user assets will not suffer losses. Additionally, Bitget has officially launched a fund recovery bounty program, offering a 5% bounty to participants who actively freeze or recover the attacker’s funds.

3 minutes ago

Uniswap processed 140 million trades in August, surpassing the combined total of Cboe BZX and NYSE American.

ARK analyst Lorenzo Valente published a post stating that Uniswap completed around 140 million trades in August this year, surpassing the combined trade count of two platforms: Cboe BZX and NYSE American. NYSE American, a subsidiary of the New York Stock Exchange, has operated since 1908, while Uniswap launched in 2018; five years ago, Uniswap’s monthly trade count was only roughly 2 million. Valente noted that by trade volume, Uniswap’s next target to catch up with is NYSE Arca, which completed approximately 232 million trades in August.

3 minutes ago

Did GPT-6 Astra Pass "Subject 2"? It Pulled Into the Cone-Marked Parking Space in 5 Minutes and 22 Seconds.

Beating AI Express (Dongcha) reports that three researchers from Taiwan, China have built DrivingBench, integrating general large models including GPT-6 Astra, GPT-5.6 Sol, Grok 4.6 and Claude Fable 5.1 into a real 2022 Toyota Corolla to conduct real vehicle driving tests similar to China’s driving test Subject 2. The test track is approximately 130 meters long, with an 8-meter-wide channel formed by colored traffic cones. The vehicle needs to navigate multiple curves before parking into a 7×5-meter cone-marked space. GPT-6 Astra, which had no prior specialized driving training, deviated from the track due to misjudgment in its first test. After adjusting its strategy in the second attempt, it completed the roughly 134.7-meter course in 5 minutes and 22 seconds and successfully parked, becoming the only one of the four models to "pass Subject 2". During the test, Astra controlled its speed at 0.5 to 0.8 meters per second near curves and obstacles, adjusting direction continuously based on real-time camera footage. In contrast, Sol, Grok and Fable all failed to finish the test: Sol and Grok failed three consecutive times, while Fable stopped after reaching around 45% of the course on its third run. The test shows that general large models not trained on specialized driving data can now complete basic driving tasks in low-speed real-world environments, though they still have a significant gap before achieving actual road autonomous driving.

3 minutes ago

The 10-year U.S. Treasury yield continues to climb, hitting 5.23%.

According to market data, the U.S. 10-year Treasury yield has continued to rise to 5.23%, hitting a new high since 2007.

3 minutes ago

CZ issues a statement in support of Bitget: Binance, BNB Chain ecosystem and community will do their utmost to help

CZ said in a post that today is a difficult day for Bitget. He believes and knows that Binance, the BNB Chain ecosystem, and its community will do everything possible to offer assistance, and called on the community to stay 'SAFU'.

3 minutes ago

Yilihua: Bitcoin faces a normal correction after climbing to $86,000, adopting a "bearish but not shorting" stance, with ETH showing even stronger momentum.

Liquid Capital founder Jack Yi wrote in a post that after Bitcoin (BTC) rose to around $86,000, it is facing a normal correction. In the current bull market, investors should "be bearish but not short" and wait patiently for the next long trading opportunity. He noted that for specific trading opportunities, Ethereum (ETH) is showing stronger performance; trading infrastructure assets such as UNI and HYPE are gaining market acceptance, while ZEC is a "strong-whale asset".

3 minutes ago

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

8 hours ago

Per Lookonchain’s monitoring, wallet 0x6436 associated with Hyperliquid Strategies has purchased 494,200 HYPE tokens (valued at approximately $45.8 million) over the past 16 hours. Over the past month, the wallet has accumulated a total of 5.51 million HYPE tokens (worth around $476 million), averaging 183,574 HYPE (about $15.86 million) in daily purchases. Hyperliquid Strategies currently holds a total of 35.1 million HYPE tokens, with a value of roughly $3.2 billion.

Scan the QR code

Download APP

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

Key Highlights On September 24 at 11:00 UTC, Binance activated spot trading for HYPE across three trading pairs: HYPE/USDT, HYPE/USDC, and HYPE/TRY. The exchange assigned HYPE a “Seed” classification, indicating elevated risk and requiring traders to pass risk assessment quizzes quarterly. A prominent wallet identified as 0x6436 acquired 373,730 HYPE tokens valued at approximately $36 million during September alone. Chart analysis identifies key resistance barriers at $110, $130, $160, and $180, with extended projections pointing toward $210. According to September SEC documentation, approximately $1.2 billion in HYPE has been repurchased and permanently removed from circulation since the protocol’s inception. Binance commenced spot market trading for Hyperliquid’s native HYPE token on September 24. The trading session launched at 11:00 UTC, introducing three distinct trading pairs to the platform.

Hyperliquid (HYPE) Price The available pairs include HYPE/USDT, HYPE/USDC, and HYPE/TRY. Binance TR users maintain exclusive access to the TRY pairing, which facilitates trading against the Turkish Lira.

Binance enabled deposit functionality one hour prior to the trading launch. The exchange scheduled withdrawal services to commence on September 25 at 11:00 UTC, though officials noted this timeframe remains subject to adjustment.

The exchange waived all listing fees for HYPE, requiring zero BNB. Spot Algo Orders became available immediately upon launch, while Trading Bots and Spot Copy Trading features were scheduled to activate within the first 24 hours.

Geographic restrictions prevent users from the United States, Canada, Iran, the Netherlands, and several additional jurisdictions from accessing these trading pairs. Binance indicated the restricted territories list may undergo modifications according to evolving regulatory frameworks.

Exchange Designates HYPE as Elevated Risk Asset Binance has applied its Seed designation to HYPE. This classification applies to recently launched tokens that the exchange identifies as possessing greater volatility profiles and risk factors compared to established digital assets.

The Seed tag mandates that authorized users successfully complete risk-awareness assessments every 90 days prior to executing trades. These mandatory quizzes extend across both Binance Spot and Binance Margin platforms.

Exchange officials cautioned that HYPE will probably experience significant price fluctuations and urged market participants to implement proper risk management protocols while conducting independent research.

Coin Bureau, a prominent cryptocurrency analysis platform, published commentary on X regarding the listing as events unfolded. The post noted that HYPE climbed toward $95 following Binance’s listing confirmation, mentioning that spot trading would commence at 11:00 UTC with deposits already operational and withdrawals scheduled for the following day.

🚨BREAKING: $HYPE surges near $95 after Binance announces it will list Hyperliquid.

Binance will open spot trading for HYPE at 11:00 UTC today, with deposits already open and withdrawals starting tomorrow. pic.twitter.com/FOEDBMsVzx

— Coin Bureau (@coinbureau) September 24, 2026

Major Wallet Continues Strategic HYPE Accumulation Independent of the Binance listing developments, blockchain analytics service Lookonchain has monitored wallet 0x6436, which has consistently accumulated HYPE tokens since late August.

Throughout September, this wallet acquired 373,730 HYPE tokens representing approximately $36 million in value. Prior to that, the same address purchased 308,569 tokens valued at roughly $26 million.

Additional transactions include a September 1 acquisition of 141,442 HYPE for approximately $11.88 million, along with an August 30 purchase of 243,713 HYPE totaling about $20.24 million. Following each purchase, the wallet has consistently staked its accumulated holdings.

This particular address has been associated with PURR, a publicly listed entity that has submitted regulatory filings outlining an institutional strategy centered on maintaining HYPE as a core treasury reserve. This connection indicates the purchasing activity likely represents a strategic long-term institutional position rather than speculative short-term trading behavior.

Market analyst WESLAD, who shares technical evaluations on TradingView, has mapped out a price structure for HYPE. The analysis identifies $72 to $82 as a critical demand region, with a decisive move above $89.59 potentially clearing the path toward $130 and subsequently $170 to $180.

WESLAD’s technical framework also designates $68.34 as the invalidation threshold for the immediate setup, while $47.59 represents the critical level that would threaten the overall bullish trend if breached.

Technical analyst @im_goomba has independently outlined an extended price trajectory toward $210, identifying $110 and $160 as significant resistance thresholds during the advance.

Hyperliquid’s underlying protocol channels trading fee revenue into its Assistance Fund mechanism, which systematically converts collected fees into HYPE tokens before permanently burning them. September SEC documentation revealed that approximately $1.2 billion worth of HYPE, representing roughly 46.4 million tokens, has been repurchased and destroyed through this deflationary mechanism since the protocol launched.

HYPE was exchanging hands near $95 in the aftermath of Binance’s listing announcement, according to the latest market data examined for this report.