7*24

Hot topic

Market

Activities

According to Beating monitoring, the Intellectual Property and Commercial Court in Taiwan has issued a first-instance ruling on the TSMC 2nm process trade secret case. Chen Liming, who previously worked for TSMC and Tokyo Electron's Taiwan subsidiary, has been sentenced to 10 years in prison, while the Tokyo Electron Taiwan subsidiary has been fined NT$150 million. The ruling found that after leaving TSMC, Chen Liming joined the marketing department of Tokyo Electron's Taiwan subsidiary and solicited advanced process information from engineers still employed at TSMC to help Tokyo Electron secure more equipment orders from TSMC. The involved information included trade secrets related to etching equipment used in 2nm production, with some data obtained through photography and reproduction. Among other defendants, three TSMC employees, Wu Bingjun, Ge Yiping, and Chen Weijie, were sentenced to 3 years, 2 years, and 6 years, respectively; Tokyo Electron Taiwan subsidiary employee Lu Yiyin was sentenced to 10 months, with a 3-year probation. The NT$150 million fine for Tokyo Electron's Taiwan subsidiary can be suspended on the condition that they compensate TSMC NT$100 million and pay NT$50 million to the public treasury. After being additionally prosecuted last year, Tokyo Electron stated that the prosecution did not accuse the Tokyo Electron Group or its Taiwan subsidiary of organizationally instructing employees to improperly obtain information, and the company's internal investigation found no evidence of relevant confidential information leaking to third parties. TSMC has previously stated that it adopts a zero-tolerance policy towards trade secret leaks and has strengthened internal controls.

[TSMC 2nm Leak Case: Chen Liming Sentenced to 10 Years, Tokyo Electron Subsidiary Fined NT$150 Million]According to Beating monitoring, the Intellectual Property and Commercial Court in Taiwan has issued a first-instance ruling on the TSMC 2nm process trade secret case. Chen Liming, who previously worked for TSMC and Tokyo Electron's Taiwan subsidiary, has been sentenced to 10 years in prison, while the Tokyo Electron Taiwan subsidiary has been fined NT$150 million. The ruling found that after leaving TSMC, Chen Liming joined the marketing department of Tokyo Electron's Taiwan subsidiary and solicited advanced process information from engineers still employed at TSMC to help Tokyo Electron secure more equipment orders from TSMC. The involved information included trade secrets related to etching equipment used in 2nm production, with some data obtained through photography and reproduction. Among other defendants, three TSMC employees, Wu Bingjun, Ge Yiping, and Chen Weijie, were sentenced to 3 years, 2 years, and 6 years, respectively; Tokyo Electron Taiwan subsidiary employee Lu Yiyin was sentenced to 10 months, with a 3-year probation. The NT$150 million fine for Tokyo Electron's Taiwan subsidiary can be suspended on the condition that they compensate TSMC NT$100 million and pay NT$50 million to the public treasury. After being additionally prosecuted last year, Tokyo Electron stated that the prosecution did not accuse the Tokyo Electron Group or its Taiwan subsidiary of organizationally instructing employees to improperly obtain information, and the company's internal investigation found no evidence of relevant confidential information leaking to third parties. TSMC has previously stated that it adopts a zero-tolerance policy towards trade secret leaks and has strengthened internal controls.

TSMC 2nm Leak Case: Chen Liming Sentenced to 10 Years, Tokyo Electron Subsidiary Fined NT$150 Million

On April 27, news emerged from Arkham monitoring that MachiBigBrother (Huang Licheng) currently holds long positions of $44.2 million in BTC and $41.8 million in ETH, totaling $86 million. Over the past six months, he has incurred a loss of $73.44 million in cryptocurrency trading.

[Huang Licheng Currently Holds $86 Million in Long Positions on BTC and ETH, Suffering $73.44 Million Loss in Last 6 Months]On April 27, news emerged from Arkham monitoring that MachiBigBrother (Huang Licheng) currently holds long positions of $44.2 million in BTC and $41.8 million in ETH, totaling $86 million. Over the past six months, he has incurred a loss of $73.44 million in cryptocurrency trading.

Huang Licheng Currently Holds $86 Million in Long Positions on BTC and ETH, Suffering $73.44 Million Loss in Last 6 Months

On April 27, news emerged that multiple users reported issues with the cryptocurrency exchange Websea, which previously invited actress Takizawa Laura to the Hong Kong Web3 conference. Users are currently unable to withdraw funds, and the C2C channel has been closed, leading to suspicions of an exit scam. Users reported that the platform initially restricted withdrawals and then directly closed the C2C channel, causing panic selling in the secondary market, with the price of USDT dropping to half. There has been no official response yet, and the situation is still developing.

[Websea Cryptocurrency Exchange Suspected of Exit Scam, Withdrawal Channel Closed]On April 27, news emerged that multiple users reported issues with the cryptocurrency exchange Websea, which previously invited actress Takizawa Laura to the Hong Kong Web3 conference. Users are currently unable to withdraw funds, and the C2C channel has been closed, leading to suspicions of an exit scam. Users reported that the platform initially restricted withdrawals and then directly closed the C2C channel, causing panic selling in the secondary market, with the price of USDT dropping to half. There has been no official response yet, and the situation is still developing.

Websea Cryptocurrency Exchange Suspected of Exit Scam, Withdrawal Channel Closed

In 2003, Elon Musk delivered a speech at Stanford University, discussing his motivation for founding SpaceX: it began with a reflection on the stagnation in space exploration—why has human progress slowed down after the moon landing? To address this, he proposed the "Mars Oasis" plan, aiming to send life to Mars at a low cost and reignite public enthusiasm for space. However, during the process, he discovered that the real bottleneck was not technology, but the cost of launches. After multiple failed attempts to seek low-cost rockets in Russia, he decided to take matters into his own hands—by building affordable rockets himself.

[Musk's 20-Year-Old Stanford Speech on the Founding of SpaceX]In 2003, Elon Musk delivered a speech at Stanford University, discussing his motivation for founding SpaceX: it began with a reflection on the stagnation in space exploration—why has human progress slowed down after the moon landing? To address this, he proposed the "Mars Oasis" plan, aiming to send life to Mars at a low cost and reignite public enthusiasm for space. However, during the process, he discovered that the real bottleneck was not technology, but the cost of launches. After multiple failed attempts to seek low-cost rockets in Russia, he decided to take matters into his own hands—by building affordable rockets himself.

Musk's 20-Year-Old Stanford Speech on the Founding of SpaceX

According to monitoring by Dongcha Beating, the Alibaba PAI team has released and open-sourced a small intelligent language model designed for industrial-grade tool invocation, named AgenticQwen (which includes 8B and 30B-A3B versions). This series of models is trained using an innovative 'dual data flywheel' reinforcement learning framework, significantly reducing inference costs while achieving agent capabilities comparable to large models with nearly 100 billion parameters. The core mechanism lies in its 'dual data flywheel' training method. Traditional synthetic data often leads to homogenization, capping model performance; in response, AgenticQwen introduces two flywheels: the inference flywheel automatically generates harder variants from the model's mistakes, while the agent flywheel expands simple linear workflows (such as a single ticket booking process) into multi-branch behavior trees that include constraints, refusals, and adversarial conditions, simulating real complex decision-making scenarios. Evaluations show that AgenticQwen-8B scores an average of 47.4 in real tool environment benchmarks (such as TAU-2 and BFCL-V4), far exceeding the baseline Qwen3-8B (23.8) and approaching Qwen3-235B (52.0). AgenticQwen-30B-A3B (activating only 3B parameters) achieved a score of 50.2. The model has already been deployed in internal production systems similar to Manus, significantly narrowing the gap with the 235B large model (with shorter end-to-end inference times); however, the paper also acknowledges that, limited by a native context length of 40K, the small model still has limitations in deep search tasks.

[Alibaba PAI Releases Open Source AgenticQwen Small Model: Achieving 8B Performance Close to 235B with 'Dual Data Flywheel']According to monitoring by Dongcha Beating, the Alibaba PAI team has released and open-sourced a small intelligent language model designed for industrial-grade tool invocation, named AgenticQwen (which includes 8B and 30B-A3B versions). This series of models is trained using an innovative 'dual data flywheel' reinforcement learning framework, significantly reducing inference costs while achieving agent capabilities comparable to large models with nearly 100 billion parameters. The core mechanism lies in its 'dual data flywheel' training method. Traditional synthetic data often leads to homogenization, capping model performance; in response, AgenticQwen introduces two flywheels: the inference flywheel automatically generates harder variants from the model's mistakes, while the agent flywheel expands simple linear workflows (such as a single ticket booking process) into multi-branch behavior trees that include constraints, refusals, and adversarial conditions, simulating real complex decision-making scenarios. Evaluations show that AgenticQwen-8B scores an average of 47.4 in real tool environment benchmarks (such as TAU-2 and BFCL-V4), far exceeding the baseline Qwen3-8B (23.8) and approaching Qwen3-235B (52.0). AgenticQwen-30B-A3B (activating only 3B parameters) achieved a score of 50.2. The model has already been deployed in internal production systems similar to Manus, significantly narrowing the gap with the 235B large model (with shorter end-to-end inference times); however, the paper also acknowledges that, limited by a native context length of 40K, the small model still has limitations in deep search tasks.

Alibaba PAI Releases Open Source AgenticQwen Small Model: Achieving 8B Performance Close to 235B with 'Dual Data Flywheel'

On April 27, Cointelegraph reported that Vanessa Perrée, the national prosecutor for organized crime in France, stated that French law enforcement has charged at least 88 individuals in connection with 12 'ransom' cases targeting cryptocurrency holders, including 10 minors, with 75 individuals currently in custody awaiting trial. These cases typically involve violent methods such as home invasions and kidnappings to coerce victims into surrendering their crypto wallet assets. Vanessa Perrée noted that some suspects are linked to multiple cases, and investigations reveal the existence of a structured criminal network behind these incidents. Records from French authorities indicate that there were 18 such cases in 2024, which increased to 67 in 2025, and there have been 47 cases so far in 2026.

[French Authorities Charge 88 in 12 Crypto 'Ransom' Cases, Including 10 Minors]On April 27, Cointelegraph reported that Vanessa Perrée, the national prosecutor for organized crime in France, stated that French law enforcement has charged at least 88 individuals in connection with 12 'ransom' cases targeting cryptocurrency holders, including 10 minors, with 75 individuals currently in custody awaiting trial. These cases typically involve violent methods such as home invasions and kidnappings to coerce victims into surrendering their crypto wallet assets. Vanessa Perrée noted that some suspects are linked to multiple cases, and investigations reveal the existence of a structured criminal network behind these incidents. Records from French authorities indicate that there were 18 such cases in 2024, which increased to 67 in 2025, and there have been 47 cases so far in 2026.

French Authorities Charge 88 in 12 Crypto 'Ransom' Cases, Including 10 Minors

On April 27, as Bitcoin briefly dropped below $78,000 and Ethereum fell over 2.4% in nearly one hour, the crypto market experienced a wave of rapid liquidations. According to Coinglass data, over $70 million was liquidated across the network in nearly one hour, with $68.32 million coming from long positions.

[Crypto Market Plunges, Over $70 Million Liquidated in One Hour]On April 27, as Bitcoin briefly dropped below $78,000 and Ethereum fell over 2.4% in nearly one hour, the crypto market experienced a wave of rapid liquidations. According to Coinglass data, over $70 million was liquidated across the network in nearly one hour, with $68.32 million coming from long positions.

Crypto Market Plunges, Over $70 Million Liquidated in One Hour

On April 27, Morpho CEO Paul Frambot posted on X, stating that in the past week he has communicated with several large institutions to understand their views on the current DeFi situation, with the following core conclusions. 
First, institutional interest will not disappear. The reason is simple: distribution channels will not vanish — massive assets under management (AUM), payment, and lending operations are moving on-chain. Almost all fintech companies aim for full on-chain integration. For institutions, this is not a question of 'whether to do it,' but rather 'it must be done.' 
Second, they have lost trust in the pool/hub model. Institutions and distributors want to control everything, including control over the code, control over risks, and control over compliance. At the same time, they seek flexibility, meaning they want to be able to isolate their own business and connect to a global liquidity network that is compatible with their operations.

[Morpho CEO: Institutional Entry into DeFi is Inevitable, Control and Compliance are Core Demands]On April 27, Morpho CEO Paul Frambot posted on X, stating that in the past week he has communicated with several large institutions to understand their views on the current DeFi situation, with the following core conclusions. 
First, institutional interest will not disappear. The reason is simple: distribution channels will not vanish — massive assets under management (AUM), payment, and lending operations are moving on-chain. Almost all fintech companies aim for full on-chain integration. For institutions, this is not a question of 'whether to do it,' but rather 'it must be done.' 
Second, they have lost trust in the pool/hub model. Institutions and distributors want to control everything, including control over the code, control over risks, and control over compliance. At the same time, they seek flexibility, meaning they want to be able to isolate their own business and connect to a global liquidity network that is compatible with their operations.

Morpho CEO: Institutional Entry into DeFi is Inevitable, Control and Compliance are Core Demands

On April 27, according to Southern Daily, the Guangdong Province Artificial Intelligence Application Matching Conference was held. It is reported that the conference adopted a '1+5' model, featuring one main conference and five specialized matching sessions. The main conference in the morning announced a series of policy measures to support the development of artificial intelligence in Guangdong Province, showcasing 23 typical application cases in fields such as 'Artificial Intelligence +' technology, manufacturing, healthcare, education, government affairs, agriculture, elderly care, and public services. It also held promotional activities for important products, research and supply-demand platforms, application scenarios, and industrial ecosystems. In the afternoon, five specialized matching sessions focused on 'Artificial Intelligence +' technology innovation, manufacturing, trade circulation, public services, and computing power collaboration, organizing various parties to conduct face-to-face precise supply-demand matching, promoting the deep integration of artificial intelligence with the real economy. (Dongxin News Agency)

[Guangdong Province Hosts Artificial Intelligence Application Matching Conference]On April 27, according to Southern Daily, the Guangdong Province Artificial Intelligence Application Matching Conference was held. It is reported that the conference adopted a '1+5' model, featuring one main conference and five specialized matching sessions. The main conference in the morning announced a series of policy measures to support the development of artificial intelligence in Guangdong Province, showcasing 23 typical application cases in fields such as 'Artificial Intelligence +' technology, manufacturing, healthcare, education, government affairs, agriculture, elderly care, and public services. It also held promotional activities for important products, research and supply-demand platforms, application scenarios, and industrial ecosystems. In the afternoon, five specialized matching sessions focused on 'Artificial Intelligence +' technology innovation, manufacturing, trade circulation, public services, and computing power collaboration, organizing various parties to conduct face-to-face precise supply-demand matching, promoting the deep integration of artificial intelligence with the real economy. (Dongxin News Agency)

Guangdong Province Hosts Artificial Intelligence Application Matching Conference

On April 26, data from Dune shows that the total value bridged on zkSync reached 3,890,899 ETH, while Starknet's total bridged storage value (TVB) stands at 1,019,545 ETH, with a total of 1,229,232 bridging user addresses. The total bridged storage value for Arbitrum is 6,609,250 ETH, Optimism is 1,101,274 ETH, and Base is 3,331,912 ETH.

[Starknet Bridge Storage Value Exceeds 1 Million ETH]On April 26, data from Dune shows that the total value bridged on zkSync reached 3,890,899 ETH, while Starknet's total bridged storage value (TVB) stands at 1,019,545 ETH, with a total of 1,229,232 bridging user addresses. The total bridged storage value for Arbitrum is 6,609,250 ETH, Optimism is 1,101,274 ETH, and Base is 3,331,912 ETH.

Starknet Bridge Storage Value Exceeds 1 Million ETH

Strategy's 'Money Printer': Is STRC Bitcoin's Savior or Destroyer?

Bitcoin's Bull-Bear Range Battle Continues, HYPE Faces Critical Test of Wave V Support | Exclusive Analysis

When Wallets Start Embedding AI Agent: The New Interaction Paradigm of ERC-8211, Why Is It Worth Attention?

Prediction Markets Under Bias

Bitcoin Price Could See Another Crash, But What Is The Long-Term Prognosis?

BTC Market Pulse: Week 17

Crypto Whale Wallets Accumulate Ozak AI Aggressively as Presale Funding Surpasses $6.8 Million

Shiba Inu Crosses 20,000 Burn Transactions Milestone, Dogecoin Eyes X Money, But Why Are Prices Down?

Strategy Surpasses 800K BTC as $2.5B Bitcoin Buying Spree Continues

Bitcoin Must Do This To Continue The Rally, Or It Will Be Over

Article

If Bitcoin Surpasses $80,000, CEX Cumulative Short Liquidation Intensity Will Reach $619 Million

Starknet Bridge Storage Value Exceeds 1 Million ETH

Disclaimer：Il contenuto di cui sopra non rappresenta le posizioni di HTX.，HTX non fornisce alcuna raccomandazione di trading.。

Tutti i commenti0RecentePopolare

Temi di tendenza