India has launched a pilot for tokenized corporate bonds, with the first three issuers raising a combined 10.25 billion rupees, or about $107 million, through a new regulated market setup. The Securities and Exchange Board of India said the Demat 2.0 framework allows corporate bonds to be issued and held as digital tokens on a distributed ledger operated by the country’s statutory depositories, while settlement is handled through the Reserve Bank of India’s wholesale central bank digital currency via the Unified Market Interface. The initial issuers were public-sector lender REC, engineering conglomerate Larsen & Toubro, and non-bank lender IIFL. SEBI said the system lets issuers receive funds on the day of bidding, with atomic settlement removing the time gap between money and bond transfers and smart contracts automating interest and redemption payments. The regulator also said later phases of the pilot will add secondary trading through existing request-for-quote platforms and open access to retail investors, while keeping tokenized bonds within the country’s current legal and investor-protection framework.

Bitget said it has completed a full upgrade of the market data source used for its U.S. equities products and has formally integrated Nasdaq’s official data feed, making it the first crypto trading platform in the industry to adopt that standard. According to the company, the switch from its previous data model is designed to deliver more accurate and more timely market information, while also improving price stability and continuity during periods of sharp market volatility. To support data quality after the integration, Bitget said it built multiple safeguards on top of the official metadata layer, including data processing, real-time monitoring, quality comparison, simulation testing, and abnormal data filtering. The company added that Nasdaq has long served as a core data service provider in global financial markets and that its standards have been tested in traditional finance over many years. Bitget said the upgrade reflects deeper work on its global asset trading infrastructure as it expands from digital assets into U.S. stocks and other global asset classes.

Crypto markets saw an unusually dense flow of developments between Sept. 9 and Sept. 10, with LAPTOP drawing the most attention after its launch, airdrop claims, market-maker allocations, liquidity incentives and extreme price moves all unfolded in quick succession. Security concerns emerged at the same time, including phishing warnings tied to LAPTOP and a separate email phishing alert from hardware wallet maker Trezor. Beyond LAPTOP, on-chain activity showed large ZEC buying through three wallets, a profitable six-month run in VVV and LIT by whale address 0x7378, and fresh positioning by trader Based16z. Binance listed Niulai and expanded related services, while Binance Alpha added 4Stock and A Meme Coin. Institutional and infrastructure news also arrived in size. Aave launched an official MCP server for AI agents, Consensys spun out MetaMask into an independent entity, U.S. Bank tested cross-border payments using USBDC on Stellar, PayPal formally rolled out its PYUSDx platform, and Tether teamed up with Fasanara Capital on a $400 million private credit fund. Regulators and policymakers were active as well, with U.S. enforcement action against Xinbi Guarantee, fresh pressure in Washington around the CLARITY Act, and an EU move to extend local contact-point rules to crypto service providers.

LoopX, an open-source project, has released version 1.0, which sits on top of AI agents like Codex, Claude Code, and Cursor, unifying long-running tasks into a single workspace. Users can view all tasks across agents, including statuses such as executing, pending confirmation, monitoring, and scheduled. The workspace is accessible via browser or desktop apps (macOS, Windows), sharing the same local service. LoopX does not automatically read all agents; it only shows registered ones. Multiple agents can collaborate on the same goal by claiming tasks, working in parallel, or handing off tasks. This collaboration is based on shared tasks and state, not free-form chat. The release aims to streamline the management of multi-agent workflows in development environments.

UC Berkeley researchers have released CUA-Lite, an open-source platform for computer use agents (CUA) that integrates agents, environments, trajectories, evaluation, and training frameworks under a unified workspace, data schema, and command interface. The platform covers desktop, browser, and mobile environments. Its lightweight sandbox runs on any Docker host without requiring /dev/kvm support. The core component, Lite.OSWorld, replicates the OSWorld task suite and evaluator in standard Docker containers, reducing memory usage from 4.1GB to 0.9GB and increasing parallel instances by approximately 4.6 times while maintaining evaluation scores consistent with the original virtual machine version. The platform also includes LiteSample, a unified data schema, and has preprocessed over a dozen existing CUA datasets, published on Hugging Face. It integrates more than 10 agents based on models such as GPT, Claude, and Gemini, along with over 15 benchmarks spanning desktop, browser, and mobile. The framework supports supervised fine-tuning and reinforcement learning via a single command. For example, fine-tuning the Qwen3-VL-2B-Instruct model on Lite.ScaleCUA improved the average episode return from 0.138 to 0.237 on a 332-task evaluation set, an increase of about 71%.

Hyperliquid Research Collective said trading on Hyperliquid’s HIP-4 prediction market layer accelerated after the platform opened third-party permissionless deployments on Aug. 29. According to the report, HIP-4 averaged about $545,000 in daily volume during the first 28 days of August, then climbed to $1.97 million on Aug. 31 and reached $2.75 million over the past 24 hours. Active traders increased from 1,256 to 1,841. The report said Outcome has been the main beneficiary of the change, now accounting for nearly 85% of HIP-4 trading volume. Its $1 million trading incentive program was also cited as a driver of deeper liquidity. HRC argued that Hyperliquid’s main edge is its unified account structure, which lets prediction markets share the same account environment as perpetual futures and HIP-3 assets, allowing users to hedge perpetual positions with prediction market contracts. HRC added that sports prediction markets could represent HIP-4’s largest growth area. It also said the current bottleneck is not on-chain deployment, but regulatory access. In the U.S., the market may fall under frameworks involving the CFTC and SEC, while sports contracts may face gambling-related scrutiny.

Gold and silver markets experienced sharp volatility on September 3, with prices rebounding after a swift pullback. XAU price rose 3.01% to $4,442.91, while XAG gained 3.78% to $66.24, according to Gate platform data. The price swings fueled demand for derivatives trading, with CoinGlass data showing Gate XAU contracts generated $381 million in 24-hour trading volume and $431 million in open interest, ranking second in the industry. Gate XAG open interest reached $94.53 million, also ranking second. Gate offers metal perpetual contracts and CFDs, covering gold, silver, platinum and other metals, with 7x24 trading and USDT settlement. The platform's unified account system enables users to allocate assets across stocks, metals, forex, indices and commodities without switching platforms, providing a flexible and efficient one-stop global asset allocation solution.

Fudan University's Neolab team has proposed the 'Life Operator', a modeling framework intended to unify life phenomena from molecules to ecosystems. The team has published six consecutive papers in Nature-family journals. Positioned as a significant advance in computational life science following AIVC, the Life Operator aims to connect life science data and knowledge across levels through mathematical and computational models, enabling more systematic inference and understanding of life. Reported by Techub News, citing Chinese tech outlet QbitAI.