Samsung plans to invest $1 billion in Helix, an AI infrastructure company backed by KKR.

According to Bitget market data, SHEIN-W (Hong Kong stock code: 00625), the global online fast fashion e-commerce platform, saw its shares fall more than 13% intraday, trading at HK$30.6. On September 28, after market close, SHEIN-W released its first unaudited interim results since listing: its first-half net revenue hit US$20.1 billion, up just 1% year-on-year; total order volume reached 549 million units, up 6.4% year-on-year. While its book net profit rose sharply due to non-operating items such as changes in the fair value of convertible preferred shares, adjusted net profit—a metric that better reflects core business performance—stood at US$499 million, down 55.6% year-on-year, accounting for only 2.5% of net revenue. Operating profit dropped 52.9% year-on-year. Adjusted net profit in the second quarter alone plummeted 66.6% year-on-year to US$228 million, with its margin narrowing from 6.2% in the year-ago period to 2.1%.

According to on-chain analyst Ai Yi (handle @ai_9684xtpa), an address that built a QNT position seven years ago realized a profit after three years of holding. The 0x94e-starting address withdrew 53,632.95 QNT tokens (valued at $1.21 million) from major crypto exchanges between May 2019 and November 2022. An hour ago, it deposited 9,000 of these QNT tokens into Coinbase and Kraken, worth $1.998 million. The deposit price stood at $222.09, while the address’s average cost basis was as low as $22.57, translating to an 884% return on investment.

Cross-chain transaction protocol Relay has disclosed that its API exposed pending transaction information, enabling MEV searchers to front-run and sandwich user orders. The incident occurred between September 12 and 26, with activity concentrated from September 23 to 26. Approximately 5,600 users were affected, suffering a median loss of around $11.88, while searchers collectively profited roughly $136,000. Relay has awarded a $50,000 bug bounty to @Outputlayer, the entity that discovered and reported the vulnerability. The protocol will also automatically compensate affected user wallets a total of approximately $312,000, with no separate application required. Relay noted that it will continue to enhance execution quality and privacy protection across its transaction paths.

According to TradingBeats' monitoring, ZEC is currently trading at $1,388.4, down approximately 12.1% in 24 hours, with a low of $1,379. Although the decline has exceeded 10%, large nearby long positions have not been liquidated by this pullback. Among million-dollar-level holdings, only one long position of around $1 million has been liquidated by the market so far. The nearest long liquidation line remains at $1,359.45, about 2.1% below the current price. Corresponding to this, approximately $17.455 million in long positions are still held, with an average entry price of $1,558.41, resulting in an unrealized loss of about $2.137 million (0xcbab…). Further down, there is an estimated liquidation line for approximately $2.631 million in long positions at $1,357.59. The two large positions, totaling around $20.086 million, have concentrated risk near the $1,358–$1,359 level.

According to HTX market data, as Bitcoin pulled back below $83,000, altcoins saw sharp declines. Specifics include: Q trades at $0.022159, down 23.07% in 24 hours; INX at $0.0057, down 21.52%; QNT at $207.12, down 20.39%; FARTCOIN at $0.1623, down 15.60%; USELESS at $0.22742, down 14.86%; SEI at $0.07364, down 13.90%; RAY at $1.8417, down 13.83%; ONDO at $0.4964, down 13.72%; ARB at $0.19367, down 13.33%; and NEAR at $4.604, down 13.07% over the same period.

Beating AI News Flash: Anthropic has released Claude Sonnet 5.5, continuing to focus on daily work and programming use cases. Its API unit prices remain unchanged from Sonnet 5: $2 per million input tokens and $10 per million output tokens, exactly half the price of Opus 5.5. Anthropic states the new model is over 30% faster in output, typically requires fewer tokens to complete the same task, cutting per-task costs by up to 30%. The most notable improvement is in agent programming. In Terminal-Bench 4.0, which tests models on completing multi-step tasks via the command line, Sonnet 5.5 jumps from Sonnet 5’s 10.3% to 70.6%, surpassing Opus 5.5’s 66.4%. However, among the 8 official benchmarks released, Sonnet 5.5 only leads Opus in this single category; the other 7 still favor Opus 5.5, though most gaps are narrow. Anthropic clarifies that complex, open-ended tasks requiring ongoing judgment remain better suited for Opus 5.5. But actual per-task savings may not be half. Artificial Analysis measured an Intelligence Index of 56 for Sonnet 5.5 at its highest Max reasoning level, versus 58 for Opus 5.5 Max. However, Sonnet 5.5 used an average of ~193,000 output tokens per task, resulting in a per-task cost of $7.60 in this test—higher than Opus 5.5 Max’s $5.98. Whether it saves money depends heavily on the task and reasoning level selected. Claude Sonnet 5.5 is now available on Claude, Claude API, AWS, Google Cloud, and Microsoft Foundry. Haiku 5.5 will launch in the coming weeks.