Ten European financial institutions have jointly established RL1 to provide infrastructure for tokenized assets.

HSBC and Ant Digital Technology have announced the successful completion of an AI Agent micropayment technology verification test. The test integrates HSBC’s tokenized deposit service, Ant Digital’s Anvita Flow network, and the Jovay testnet, showcasing the full process where AI Agents automatically discover digital services, select them, and complete payments, with each micropayment typically valued at under $2. HSBC is responsible for payment settlement and real-time risk checks, Anvita Flow coordinates service access and payment execution, and Jovay provides the Layer 2 blockchain test environment. All transactions were settled via the blockchain testnet. Both parties emphasized that the test is purely for technical verification purposes and does not signify the formal commercialization or customer rollout of related products. As AI Agents gradually gain the capability to independently invoke services and execute transactions, tokenized deposits, stablecoins, and blockchain payment infrastructure have emerged as key focus areas for financial institutions exploring machine-to-machine payments.

According to a Wall Street Journal (WSJ) report, prediction market platform Kalshi is investigating a series of suspicious trades tied to the appointment of a new White House press secretary by U.S. President Donald Trump. Trump confirmed Friday that conservative commentator Katie Zacharia would take on the role—a pick not widely viewed as a frontrunner, whose win probability in Kalshi’s related markets once stood at just around 1%. An analysis of Kalshi’s public trading data by the WSJ found at least three trades that correctly bet on Zacharia’s appointment ahead of the announcement. One trade, placed Thursday evening for $19, is projected to pay out $1,896; the other two, for roughly $74 and $80 respectively, are expected to yield $3,689 and $4,023. A Kalshi spokesperson confirmed the platform is investigating the matter but did not disclose additional details. As prediction markets expand in scale, the risk of trading with non-public information has come under growing scrutiny. Earlier, Gabriel Perez, a teleprompter operator for the Trump White House, was fired for earning over $100,000 from bets on the president’s speeches, and reached a settlement with the U.S. Commodity Futures Trading Commission (CFTC) in August this year. The White House also previously sent a memo to staff warning against using non-public information for personal financial gain.

Trader web3vc withdrew 1.79M $UNI ($13.18M) from Venus Protocol and deposited it into #Binance 1 hour ago. He started buying $UNI a year ago at an average price of ~$9. He finally turned a profit a few days ago but didn't sell. Now he seems ready to give up. If sold, he would

A survey of 14,250 consumers in the Asia-Pacific region conducted by Visa shows that 46% of respondents said they would likely use stablecoins over the next five years, higher than the 16% who have actually used stablecoins in the past 12 months. Additionally, 49% of respondents expect stablecoins to become a common method for cross-border money transfers in the coming five years. The survey indicates that consumer interest in stablecoins is growing, but a significant knowledge gap remains: only 6% of respondents correctly understand how stablecoins operate, while roughly half believe stablecoins can only be used to trade other cryptocurrencies. For those who are aware of stablecoins but have not used them, fraud and scam risks are the primary concern. Nischint Sanghavi, Visa’s Head of Digital Currencies for Asia-Pacific, stated that consumers are increasingly recognizing stablecoins’ potential in daily consumption, travel, and cross-border transfers. Visa believes the next critical step is to translate market interest into a trusted, familiar, and scalable payment experience.

SkyAI, which holds over 2M $SOL($220M), appears to be dumping $SOL at a loss! Over the past 3 days, they have transferred 568,000 $SOL ($63.87M) to exchanges. Their average purchase price was ~$194, and they would lose ~$46M if sold.

Dongcha Beating AI News Flash: Alibaba’s Qwen has launched Qwen-Image-2.1-Turbo, an accelerated version of Qwen-Image-2.1 that retains the original’s 7-billion-parameter image generation architecture. The official original model uses 40 denoising steps, while the Turbo version cuts this to 8 steps – an 80% reduction. Denoising refers to the process where a model gradually generates an image from random noise. However, an 80% step reduction does not translate to an equivalent 80% time cut. The new model still produces 2K images, supporting text-guided editing, multi-image reference, and transparent background output. Developers can run it locally via Hugging Face’s Diffusers library, which automatically applies the recommended 8-step sampling setting. APIs for both Turbo and Pro versions are now live. Based on Alibaba Cloud’s Beijing Bailian region original pricing, Turbo costs 0.1 yuan per image, while Pro is priced at 0.25 yuan – making Turbo 60% cheaper. Some users have reported significantly faster generation speeds, but quality reviews are mixed. One user testing on an RTX 5090 found Turbo over-sharpens skin textures in generated characters. There is currently a lack of sufficient independent evaluations to confirm its image quality matches the original. Notably, Turbo is licensed under the Qwen Research License: while model weights are downloadable, free use is limited to non-commercial research and evaluation. Commercial development using the weights requires a separate authorization application.