Lookonchain APP

App Store

Hyperliquid is scheduled to undergo a network upgrade at 16:15 today, with an estimated downtime of approximately 10 minutes.

47 minutes ago

Hyperliquid plans to carry out a network upgrade around 16:15 Beijing Time today, with an estimated downtime of approximately 10 minutes.

Relevant content

2 New Wallets Withdraw 280.75M $龙虾 ($12.21M) From Exchanges in 3 Days

2 newly created wallets (0x4d28 and 0x5a42) withdrew 280.75M $龙虾 ($12.21M) from exchanges over the past 3 days, 28.08% of the total supply.

11 minutes ago

Supermicro contractor pleads guilty; $2.5 billion server order tied to smuggling case.

Beating AI Express News: Ting-Wei Sun, a former contractor of Supermicro, has pleaded guilty in a New York federal court in the U.S. He was involved in illegally diverting servers equipped with NVIDIA high-end AI chips to China, admitting to four charges: conspiracy to violate export controls, conspiracy to smuggle, conspiracy to defraud the U.S. government, and obstruction of justice. Previously, U.S. prosecutors indicted three people total. Besides Sun, the defendants include Supermicro co-founder Yih-Shyan Liaw and Ruei-Tsang Chang, general manager of the company’s Taiwan office. The U.S. Department of Justice (DOJ) disclosed earlier that between 2024 and 2025, a Southeast Asian middleman purchased approximately $2.5 billion worth of equipment from the involved U.S. server manufacturer. Prosecutors allege that at least $510 million worth of these servers were illegally diverted to China in just a few weeks in 2025. The involved servers were typically first shipped via Taiwan to Southeast Asia, then repackaged and sent to China. To evade inspections, the suspects placed thousands of non-functional dummy servers in warehouses and used hair dryers to remove labels and serial numbers. Prosecutors said Sun participated in arranging the dummy servers and made false statements during inspections by the U.S. Department of Commerce. Supermicro was not named as a defendant in the case and has cut ties with the three involved individuals. An independent investigation released by the company in August found no evidence that current executives were aware of the smuggling scheme.

11 minutes ago

Two addresses went long on 7.2 million SYN on Aster DEX, posting an unrealized profit of $450,000.

According to Lookonchain’s monitoring, two wallet addresses have opened long positions for SYN on Aster DEX, holding a total of roughly 7.2 million SYN tokens worth approximately $1.34 million. The combined unrealized profit of the two addresses stands at around $450,000, with the highest return on investment reaching 116%.

11 minutes ago

Two wallets opened longs on 7.2M $SYN ($1.34M) on @Aster_DEX and now sit on $450K in unrealized profit, with returns up to 116%.

Two wallets opened longs on 7.2M $SYN ($1.34M) on @Aster_DEX and now sit on $450K in unrealized profit, with returns up to 116%.

11 minutes ago

Two wallets turn $1.34M $SYN long into $450K unrealized profit on Aster DEX, 116% returns

Two wallets opened longs on 7.2M $SYN ($1.34M) on @Aster_DEX and now sit on $450K in unrealized profit, with returns up to 116%.

11 minutes ago

Prime Agent has assembled 2,200 Agents to rewrite its own system, delivering a 13x faster startup speed.

Beating AI News (Dongcha) reports that Prime Intellect has released the Rust version of its open-source AI coding tool Prime Agent. Over two weeks, the project orchestrated more than 2,200 agents to rewrite its own codebase from TypeScript to Rust, leveraging over 10,000 isolated sandboxes and consuming more than 200 billion tokens via the GLM-5.3 inference service. The workflow uses a lead orchestrator agent to break down tasks and prioritize them, with other agents handling planning, coding, review, and validation respectively. New code is cross-checked against the old version to verify UI display, session logs, and model requests; changes that fail review or testing are sent back for revision. The lead orchestrator agent does not write product code itself. Tests released by Prime Intellect show the new version’s cold start time (from launch to ready for input) dropped from ~738ms to 56ms, a roughly 13x speedup. Startup memory usage fell from ~607MB to 106MB, an 83% reduction. The new version also adds Windows beta support, with the code now open-sourced on GitHub. However, automated testing did not catch all issues. In the coming weeks, the team will continue to have AI fix bugs, complete features, and optimize performance, while human engineers will identify problems, guide revisions, and review results.

11 minutes ago

Popular tokens

BitcoinEthereumHyperliquidSolanaTRONBNBTetherAaveXRPPepeFartcoinOndoJupiterUniswapBonkPendleEthenaArbitrumAvalancheLidoChainlinkPolygonDogecoinCardano